On 30/05/2012 22:55, Ole wrote:
> This patch adds info for the required Atari OS.
BTW, how many actual requirements are there for MiNT itself?
ie. how much is missing of MiNT-specific APIs to have it running on, say
MagiC or plain TOS for ex?
A friend has a Falcon running MagiC (cause MiNT is "slower")...
François.
Wednesday, 30 May 2012
Patch: specify required Atari OS
Index: netsurfweb/downloads/atari/index.en
===================================================================
--- netsurfweb/downloads/atari/index.en (Revision 13933)
+++ netsurfweb/downloads/atari/index.en (Arbeitskopie)
@@ -65,7 +65,7 @@
<!--<p class="downloadinstructions"><a href="/documentation/roinfo#GettingStartedInstallation">Installation instructions</a></p>-->
<p class="preul">Requires:</p>
<ul>
-<li>Atari OS version?</li>
+<li>FreeMiNT 1.17</li>
</ul>
</div>
This patch adds info for the required Atari OS.
Greets,
Ole
===================================================================
--- netsurfweb/downloads/atari/index.en (Revision 13933)
+++ netsurfweb/downloads/atari/index.en (Arbeitskopie)
@@ -65,7 +65,7 @@
<!--<p class="downloadinstructions"><a href="/documentation/roinfo#GettingStartedInstallation">Installation instructions</a></p>-->
<p class="preul">Requires:</p>
<ul>
-<li>Atari OS version?</li>
+<li>FreeMiNT 1.17</li>
</ul>
</div>
This patch adds info for the required Atari OS.
Greets,
Ole
Tuesday, 29 May 2012
Re: Current version?
On Tue, 2012-05-29 at 22:15 +0200, John Williams wrote:
> In article <5285034c43JohnRW@ukgateway.net>,
> John Williams <JohnRW@ukgateway.net> wrote:
>
> > > We will enable the autobuilder again once trunk NetSurf reaches a
> > > satisfactory state.
>
> > Could this be announced here, please, when 'normal service resumes'?
>
> Is there any news on this?
No.
John-Mark.
> In article <5285034c43JohnRW@ukgateway.net>,
> John Williams <JohnRW@ukgateway.net> wrote:
>
> > > We will enable the autobuilder again once trunk NetSurf reaches a
> > > satisfactory state.
>
> > Could this be announced here, please, when 'normal service resumes'?
>
> Is there any news on this?
No.
John-Mark.
Re: Current version?
On 29 May, JohnRW@ukgateway.net wrote:
> It's been a long, long time!
Is that a quote from 'Rocket Man'?
John
> It's been a long, long time!
Is that a quote from 'Rocket Man'?
John
Re: Current version?
In article <5285034c43JohnRW@ukgateway.net>,
John Williams <JohnRW@ukgateway.net> wrote:
> > We will enable the autobuilder again once trunk NetSurf reaches a
> > satisfactory state.
> Could this be announced here, please, when 'normal service resumes'?
Is there any news on this?
It's been a long, long time!
John
John Williams <JohnRW@ukgateway.net> wrote:
> > We will enable the autobuilder again once trunk NetSurf reaches a
> > satisfactory state.
> Could this be announced here, please, when 'normal service resumes'?
Is there any news on this?
It's been a long, long time!
John
Monday, 28 May 2012
Re: [GeSHi-devel] new language file GNU/Octave
On 25 May 2012 17:37, Carnë Draug <carandraug+dev@gmail.com> wrote:
> Hi everyone
>
> me and Juan Carbajal have wrote a language file to highlight
> GNU/octave source code that is attached.
>
> There is only one problem that we could not fix with adding URLS. We
> have 10 different groups on KEYWORDS, 5 of them with rules for URLs.
> The rule is the same for all of them. However, 1 of them (the very
> first) doesn't work.
>
> The following code:
> uint32
> cell
>
> creates
> function/uint32.html">uint32
> function/cell.html">cell
>
> with links for
> http://octave.sourceforge.net/octave/%3Cspan%20class=
> http://octave.sourceforge.net/octave/%3Cspan%20class=
>
> The really weird thing is that keywords from the other lists, even
> though having the same style and URL rule, work normally.
Hi everyone
I managed to fix the problem on the file but I think it is also a bug
with GeSHi. The problem is that one of the keywords (function) is also
a word on the URL. The keyword function was on the element #4. The
only words that were giving me a problem were on element #1. When I
swapped those two, the problem went away. Still, this sounds like a
bug to me.
Another problem I noticed, shows up when one of the keywords is DOT,
whcich will also break the URLs (since the actual dots are replaced by
<DOT> and then highlighted. Again this will only happen for the
elements that have smaller key # than the one with the keyword DOT). I
noticed this since octave actually has a function `dot', and before I
made the match case sensitive, this also broke the highlighting.
Should I report this bug somewhere? I do not know how to code PHP,
fixing this is beyond my abilities.
Anyway, please see attached the final version of the octave.php
language file and I hope it's acceptable for inclusion on GeSHi. It
passes on langcheck.php file amd we tried to leave useful comments on
the file which should hopefully make clear what each option we made.
Carnë
> Hi everyone
>
> me and Juan Carbajal have wrote a language file to highlight
> GNU/octave source code that is attached.
>
> There is only one problem that we could not fix with adding URLS. We
> have 10 different groups on KEYWORDS, 5 of them with rules for URLs.
> The rule is the same for all of them. However, 1 of them (the very
> first) doesn't work.
>
> The following code:
> uint32
> cell
>
> creates
> function/uint32.html">uint32
> function/cell.html">cell
>
> with links for
> http://octave.sourceforge.net/octave/%3Cspan%20class=
> http://octave.sourceforge.net/octave/%3Cspan%20class=
>
> The really weird thing is that keywords from the other lists, even
> though having the same style and URL rule, work normally.
Hi everyone
I managed to fix the problem on the file but I think it is also a bug
with GeSHi. The problem is that one of the keywords (function) is also
a word on the URL. The keyword function was on the element #4. The
only words that were giving me a problem were on element #1. When I
swapped those two, the problem went away. Still, this sounds like a
bug to me.
Another problem I noticed, shows up when one of the keywords is DOT,
whcich will also break the URLs (since the actual dots are replaced by
<DOT> and then highlighted. Again this will only happen for the
elements that have smaller key # than the one with the keyword DOT). I
noticed this since octave actually has a function `dot', and before I
made the match case sensitive, this also broke the highlighting.
Should I report this bug somewhere? I do not know how to code PHP,
fixing this is beyond my abilities.
Anyway, please see attached the final version of the octave.php
language file and I hope it's acceptable for inclusion on GeSHi. It
passes on langcheck.php file amd we tried to leave useful comments on
the file which should hopefully make clear what each option we made.
Carnë
Sunday, 27 May 2012
Netsurf 2.9 for Mac OS X? (intel)
Hi, I am running Netsurf 2.8 on Mac OS X Lion and I am very happy with it. I wonder if a release of version 2.9 is planned. Or if anyone could compile a version for Snow Leopard or Lion, it would be great if you could send it to me.
Thanks and congrats for this great browser!
Best,
Martin
Thanks and congrats for this great browser!
Best,
Martin
Friday, 25 May 2012
Re: NetSurf crashing on cpc site
In article <1337939988.16298.7.camel@duiker>,
John-Mark Bell <jmb@netsurf-browser.org> wrote:
> 1. The version of NetSurf you are using
2.9
> 2. What hardware and operating system version you are using
Risc PC RO 4.39
> 3. Whether this issue is reproducible
Several times at the time but not 5mins ago.
> 4. The full URL of the page that causes problems
Difficult as it crashes as soon as it tries to go to it but I guess
http://cpc.farnell.com/mercury/651-662uk/soft-start-inverter-12v-300w/dp/PW02762?Ntt=PW02762
as reported by Brian Howlett
> 1. We document our bug reporting guidelines here:
> http://www.netsurf-browser.org/documentation/info#Bugs
> Please endeavour to follow them as it saves everybody time.
Yes, I have endeavoured to create a bug report:
http://sourceforge.net/tracker/?func=detail&aid=3529714&group_id=51719&atid=464312
but didn't have the intelligence to try zipping the 807,548 bytes log
file, as I mentioned in my earlier post.
--
Stuart Winsor
Only plain text for emails
http://www.asciiribbon.org
John-Mark Bell <jmb@netsurf-browser.org> wrote:
> 1. The version of NetSurf you are using
2.9
> 2. What hardware and operating system version you are using
Risc PC RO 4.39
> 3. Whether this issue is reproducible
Several times at the time but not 5mins ago.
> 4. The full URL of the page that causes problems
Difficult as it crashes as soon as it tries to go to it but I guess
http://cpc.farnell.com/mercury/651-662uk/soft-start-inverter-12v-300w/dp/PW02762?Ntt=PW02762
as reported by Brian Howlett
> 1. We document our bug reporting guidelines here:
> http://www.netsurf-browser.org/documentation/info#Bugs
> Please endeavour to follow them as it saves everybody time.
Yes, I have endeavoured to create a bug report:
http://sourceforge.net/tracker/?func=detail&aid=3529714&group_id=51719&atid=464312
but didn't have the intelligence to try zipping the 807,548 bytes log
file, as I mentioned in my earlier post.
--
Stuart Winsor
Only plain text for emails
http://www.asciiribbon.org
Re: NetSurf crashing on cpc site
In article <5b18549552.old_coaster@old_coaster.yahoo.co.uk>,
Tony Moore <old_coaster@yahoo.co.uk> wrote:
> Here, using NetSurf 2.9 on RO 6.20, a search for PW02762 leads to
> http://cpc.farnell.com/mercury/651-662uk/soft-start-inverter-12v-300w/dp/PW02762?Ntt=PW02762
> without any crash.
Well, I admit I did see it work once but it did crash on me several times
--
Stuart Winsor
Only plain text for emails
http://www.asciiribbon.org
Tony Moore <old_coaster@yahoo.co.uk> wrote:
> Here, using NetSurf 2.9 on RO 6.20, a search for PW02762 leads to
> http://cpc.farnell.com/mercury/651-662uk/soft-start-inverter-12v-300w/dp/PW02762?Ntt=PW02762
> without any crash.
Well, I admit I did see it work once but it did crash on me several times
--
Stuart Winsor
Only plain text for emails
http://www.asciiribbon.org
Re: NetSurf crashing on cpc site
In article <1337939902.4fbf57be100eb@netmail.pipex.net>,
dave higton <davehigton@dsl.pipex.com> wrote:
> > P.S. I can't upload the log file to the "Bug Tracker" because it is
> > 807,548 bytes.
> Zip it up. It will compress to a small fraction of that size.
Now why didn't I think of that. Doh!
So how can I add this to the bug report I have already created?
http://sourceforge.net/tracker/?func=detail&aid=3529714&group_id=51719&atid=464312
--
Stuart Winsor
Only plain text for emails
http://www.asciiribbon.org
dave higton <davehigton@dsl.pipex.com> wrote:
> > P.S. I can't upload the log file to the "Bug Tracker" because it is
> > 807,548 bytes.
> Zip it up. It will compress to a small fraction of that size.
Now why didn't I think of that. Doh!
So how can I add this to the bug report I have already created?
http://sourceforge.net/tracker/?func=detail&aid=3529714&group_id=51719&atid=464312
--
Stuart Winsor
Only plain text for emails
http://www.asciiribbon.org
[GeSHi-devel] new language file GNU/Octave
Hi everyone
me and Juan Carbajal have wrote a language file to highlight
GNU/octave source code that is attached.
There is only one problem that we could not fix with adding URLS. We
have 10 different groups on KEYWORDS, 5 of them with rules for URLs.
The rule is the same for all of them. However, 1 of them (the very
first) doesn't work.
The following code:
uint32
cell
creates
function/uint32.html">uint32
function/cell.html">cell
with links for
http://octave.sourceforge.net/octave/%3Cspan%20class=
http://octave.sourceforge.net/octave/%3Cspan%20class=
The really weird thing is that keywords from the other lists, even
though having the same style and URL rule, work normally.
Thanks in advance,
Carnë
me and Juan Carbajal have wrote a language file to highlight
GNU/octave source code that is attached.
There is only one problem that we could not fix with adding URLS. We
have 10 different groups on KEYWORDS, 5 of them with rules for URLs.
The rule is the same for all of them. However, 1 of them (the very
first) doesn't work.
The following code:
uint32
cell
creates
function/uint32.html">uint32
function/cell.html">cell
with links for
http://octave.sourceforge.net/octave/%3Cspan%20class=
http://octave.sourceforge.net/octave/%3Cspan%20class=
The really weird thing is that keywords from the other lists, even
though having the same style and URL rule, work normally.
Thanks in advance,
Carnë
Re: NetSurf crashing on cpc site
On Fri, 2012-05-25 at 09:35 +0100, lists wrote:
> The home page:
>
> http://cpc.farnell.com/
>
> works fine so does a search on something like "inverter" but trying to
> select or search for an individual item such as PW02762 results in an
> immediate crash
Please provide the following information[1]:
1. The version of NetSurf you are using
2. What hardware and operating system version you are using
3. Whether this issue is reproducible
4. The full URL of the page that causes problems
John-Mark.
1. We document our bug reporting guidelines here:
http://www.netsurf-browser.org/documentation/info#Bugs
Please endeavour to follow them as it saves everybody time.
> The home page:
>
> http://cpc.farnell.com/
>
> works fine so does a search on something like "inverter" but trying to
> select or search for an individual item such as PW02762 results in an
> immediate crash
Please provide the following information[1]:
1. The version of NetSurf you are using
2. What hardware and operating system version you are using
3. Whether this issue is reproducible
4. The full URL of the page that causes problems
John-Mark.
1. We document our bug reporting guidelines here:
http://www.netsurf-browser.org/documentation/info#Bugs
Please endeavour to follow them as it saves everybody time.
Re: NetSurf crashing on cpc site
Quoting lists <Stuartlists@orpheusinternet.co.uk>:
> In article <5295500e83Stuartlists@orpheusinternet.co.uk>,
> lists <Stuartlists@orpheusinternet.co.uk> wrote:
> > The home page:
>
> > http://cpc.farnell.com/
>
> > works fine so does a search on something like "inverter" but trying to
> > select or search for an individual item such as PW02762 results in an
> > immediate crash
>
> P.S. I can't upload the log file to the "Bug Tracker" because it is
> 807,548 bytes.
Zip it up. It will compress to a small fraction of that size.
Dave
> In article <5295500e83Stuartlists@orpheusinternet.co.uk>,
> lists <Stuartlists@orpheusinternet.co.uk> wrote:
> > The home page:
>
> > http://cpc.farnell.com/
>
> > works fine so does a search on something like "inverter" but trying to
> > select or search for an individual item such as PW02762 results in an
> > immediate crash
>
> P.S. I can't upload the log file to the "Bug Tracker" because it is
> 807,548 bytes.
Zip it up. It will compress to a small fraction of that size.
Dave
Re: NetSurf crashing on cpc site
On 25 May, Tony Moore wrote:
> On 25 May 2012, lists <Stuartlists@orpheusinternet.co.uk> wrote:
>> The home page:
>>
>> http://cpc.farnell.com/
>>
>> works fine so does a search on something like "inverter" but trying to
>> select or search for an individual item such as PW02762 results in an
>> immediate crash
> Here, using NetSurf 2.9 on RO 6.20, a search for PW02762 leads to
> http://cpc.farnell.com/mercury/651-662uk/soft-start-inverter-12v-300w/
> dp/PW02762?Ntt=PW02762
> without any crash.
Same here, with r13571 on 5.18.
--
Brian Howlett
-------------------------------------------------------------
You possess a mind not only twisted, but actually sprained...
> On 25 May 2012, lists <Stuartlists@orpheusinternet.co.uk> wrote:
>> The home page:
>>
>> http://cpc.farnell.com/
>>
>> works fine so does a search on something like "inverter" but trying to
>> select or search for an individual item such as PW02762 results in an
>> immediate crash
> Here, using NetSurf 2.9 on RO 6.20, a search for PW02762 leads to
> http://cpc.farnell.com/mercury/651-662uk/soft-start-inverter-12v-300w/
> dp/PW02762?Ntt=PW02762
> without any crash.
Same here, with r13571 on 5.18.
--
Brian Howlett
-------------------------------------------------------------
You possess a mind not only twisted, but actually sprained...
Re: NetSurf crashing on cpc site
In article <5295500e83Stuartlists@orpheusinternet.co.uk>,
lists <Stuartlists@orpheusinternet.co.uk> wrote:
> The home page:
> http://cpc.farnell.com/
> works fine so does a search on something like "inverter" but trying to
> select or search for an individual item such as PW02762 results in an
> immediate crash
P.S. I can't upload the log file to the "Bug Tracker" because it is
807,548 bytes.
--
Stuart Winsor
Only plain text for emails
http://www.asciiribbon.org
lists <Stuartlists@orpheusinternet.co.uk> wrote:
> The home page:
> http://cpc.farnell.com/
> works fine so does a search on something like "inverter" but trying to
> select or search for an individual item such as PW02762 results in an
> immediate crash
P.S. I can't upload the log file to the "Bug Tracker" because it is
807,548 bytes.
--
Stuart Winsor
Only plain text for emails
http://www.asciiribbon.org
Re: NetSurf crashing on cpc site
On 25 May 2012, lists <Stuartlists@orpheusinternet.co.uk> wrote:
> The home page:
>
> http://cpc.farnell.com/
>
> works fine so does a search on something like "inverter" but trying to
> select or search for an individual item such as PW02762 results in an
> immediate crash
Here, using NetSurf 2.9 on RO 6.20, a search for PW02762 leads to
http://cpc.farnell.com/mercury/651-662uk/soft-start-inverter-12v-300w/dp/PW02762?Ntt=PW02762
without any crash.
Tony
> The home page:
>
> http://cpc.farnell.com/
>
> works fine so does a search on something like "inverter" but trying to
> select or search for an individual item such as PW02762 results in an
> immediate crash
Here, using NetSurf 2.9 on RO 6.20, a search for PW02762 leads to
http://cpc.farnell.com/mercury/651-662uk/soft-start-inverter-12v-300w/dp/PW02762?Ntt=PW02762
without any crash.
Tony
NetSurf crashing on cpc site
The home page:
http://cpc.farnell.com/
works fine so does a search on something like "inverter" but trying to
select or search for an individual item such as PW02762 results in an
immediate crash
--
Stuart Winsor
Only plain text for emails
http://www.asciiribbon.org
http://cpc.farnell.com/
works fine so does a search on something like "inverter" but trying to
select or search for an individual item such as PW02762 results in an
immediate crash
--
Stuart Winsor
Only plain text for emails
http://www.asciiribbon.org
Thursday, 24 May 2012
[GeSHi-devel] support for asymptote
Dear GeSHi developers,
First of all, I would like to thank you for that useful piece of
software, I use it all the time in dokuwiki's.
For the asymptote language (http://asymptote.sourceforge.net) I needed
syntax highlighting so I adapted your cpp language plugin to
asymptote.
You can find it attached.
I hope that it will be usefull for other GeSHi users.
- sincerely,
Manuel
First of all, I would like to thank you for that useful piece of
software, I use it all the time in dokuwiki's.
For the asymptote language (http://asymptote.sourceforge.net) I needed
syntax highlighting so I adapted your cpp language plugin to
asymptote.
You can find it attached.
I hope that it will be usefull for other GeSHi users.
- sincerely,
Manuel
Wednesday, 23 May 2012
Re: [gccsdk] !PDF
In article <52936e89dfchrisg@care4free.net>,
Chris Gransden <chrisg@care4free.net> wrote:
> It's been a long time but I've at last switched the back end of !PDF to the
> Poppler (v.0.20.0) PDF rendering library. See
> http://poppler.freedesktop.org. This brings many immediate benefits plus
> future ones too.
I've uploaded another test version of !PDF (3.03.1.25alpha 5) here
http://www.cgransden.co.uk/pdftest.zip.
For this version 'Bitmap mode' is now integrated into !PDF. There's a new
menu option 'Bitmap Mode' on the 'Display' menu. Once selected click on
'Save Choices' to make it the default.
By default !PDF starts up in 'Vector' mode as there are still a few things
that don't work properly. Being scaling and rotation. Possibly others.
Chris.
--
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Chris Gransden <chrisg@care4free.net> wrote:
> It's been a long time but I've at last switched the back end of !PDF to the
> Poppler (v.0.20.0) PDF rendering library. See
> http://poppler.freedesktop.org. This brings many immediate benefits plus
> future ones too.
I've uploaded another test version of !PDF (3.03.1.25alpha 5) here
http://www.cgransden.co.uk/pdftest.zip.
For this version 'Bitmap mode' is now integrated into !PDF. There's a new
menu option 'Bitmap Mode' on the 'Display' menu. Once selected click on
'Save Choices' to make it the default.
By default !PDF starts up in 'Vector' mode as there are still a few things
that don't work properly. Being scaling and rotation. Possibly others.
Chris.
--
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
[dsilvers@netsurf-browser.org: Re: small bug in nsurl.c]
Oops, forgot to CC the list
D.
--
Daniel Silverstone http://www.digital-scurf.org/
PGP mail accepted and encouraged. Key Id: 3CCE BABE 206C 3B69
D.
--
Daniel Silverstone http://www.digital-scurf.org/
PGP mail accepted and encouraged. Key Id: 3CCE BABE 206C 3B69
small bug in nsurl.c
in the nsurl.c file, in nsurl__get_string_markers, there an error in the 'ftp:' detection
original code:
correct code:
I send this mail here, because bugs@netsurf-browser.org is not responding
original code:
} else if (off == SLEN("https") &&
(((*(pos - off + 0) == 'f') ||
(*(pos - off + 0) == 'F')) &&
((*(pos - off + 1) == 't') ||
(*(pos - off + 1) == 'T')) &&
((*(pos - off + 2) == 'p') ||
(*(pos - off + 2) == 'P')))) {
marker.scheme_type = NSURL_SCHEME_FTP;
correct code:
} else if (off == SLEN("ftp") &&
(((*(pos - off + 0) == 'f') ||
(*(pos - off + 0) == 'F')) &&
((*(pos - off + 1) == 't') ||
(*(pos - off + 1) == 'T')) &&
((*(pos - off + 2) == 'p') ||
(*(pos - off + 2) == 'P')))) {
marker.scheme_type = NSURL_SCHEME_FTP;
I send this mail here, because bugs@netsurf-browser.org is not responding
Tuesday, 22 May 2012
Re: Conversion from SVN to Git
On 04/30/12 09:00, Daniel Silverstone wrote:
>
> Yes, the move is much more "away from Subversion" than towards any particular
> other VCS. Specifically we're fed up with the abysmal management of merges
> which Subversion just can't do well. We're also wanting to make it possible
> for anyone to have useful branches of our codebase without having to have
> commit rights to any part of our repositories.
That would explain why I've not had much trouble with Subversion. I have
generally been the sole committer so branching and merging have never
been much of an issue.
> Mercurial: I've not spent a lot of time with this. Horror stories of data
> loss and confusing errors from those I know who still love it have
> prevented me.
I'm using this at the moment but only as a "satellite" to Subversion. I
have Mercurial on my laptop and sync it with the central Subversion
repository when I'm in the office. In that mode it works pretty nicely.
> DARCS: I just don't trust it to be predictably performant. I've seen people
> wait hours for a commit to complete sometimes. It seems to be a nice
> idea in theory, but I don't approve of its reality.
I tried it many years ago but gave up after it ate my source, in that
way SourceSafe is famous for, twice in 2 days. I imagine it is more
reliable now if not faster.
> I hope this was helpful/interesting for you.
Yes, thank you.
--
Matthew Hambley
>
> Yes, the move is much more "away from Subversion" than towards any particular
> other VCS. Specifically we're fed up with the abysmal management of merges
> which Subversion just can't do well. We're also wanting to make it possible
> for anyone to have useful branches of our codebase without having to have
> commit rights to any part of our repositories.
That would explain why I've not had much trouble with Subversion. I have
generally been the sole committer so branching and merging have never
been much of an issue.
> Mercurial: I've not spent a lot of time with this. Horror stories of data
> loss and confusing errors from those I know who still love it have
> prevented me.
I'm using this at the moment but only as a "satellite" to Subversion. I
have Mercurial on my laptop and sync it with the central Subversion
repository when I'm in the office. In that mode it works pretty nicely.
> DARCS: I just don't trust it to be predictably performant. I've seen people
> wait hours for a commit to complete sometimes. It seems to be a nice
> idea in theory, but I don't approve of its reality.
I tried it many years ago but gave up after it ate my source, in that
way SourceSafe is famous for, twice in 2 days. I imagine it is more
reliable now if not faster.
> I hope this was helpful/interesting for you.
Yes, thank you.
--
Matthew Hambley
Monday, 21 May 2012
[gccsdk] !PDF
It's been a long time but I've at last switched the back end of !PDF to the
Poppler (v.0.20.0) PDF rendering library. See
http://poppler.freedesktop.org. This brings many immediate benefits plus
future ones too.
I've uploaded a test version of !PDF (3.03.1.25alpha 4) here
http://www.cgransden.co.uk/pdftest.zip. The application folder is named
!PDFtest so it can be run alongside the existing version for comparison.
It's a bit rough around the edges but should be ok.
The save menu has expanded to include,
Save 'Page as png', 'Page as jpeg', 'Page as pdf' and 'Page as html'.
If you combine !Netsurf with the 'Cyberbit' font it's possible to create
'Draw' files with non latin characters from the exported the html page.
Chinese seems to work ok.
The majority of the changes are behind the scenes. One visible benefit is
'Bitmap' mode should render pages more accurately.
Chris.
--
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Poppler (v.0.20.0) PDF rendering library. See
http://poppler.freedesktop.org. This brings many immediate benefits plus
future ones too.
I've uploaded a test version of !PDF (3.03.1.25alpha 4) here
http://www.cgransden.co.uk/pdftest.zip. The application folder is named
!PDFtest so it can be run alongside the existing version for comparison.
It's a bit rough around the edges but should be ok.
The save menu has expanded to include,
Save 'Page as png', 'Page as jpeg', 'Page as pdf' and 'Page as html'.
If you combine !Netsurf with the 'Cyberbit' font it's possible to create
'Draw' files with non latin characters from the exported the html page.
Chinese seems to work ok.
The majority of the changes are behind the scenes. One visible benefit is
'Bitmap' mode should render pages more accurately.
Chris.
--
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Sunday, 20 May 2012
Re: DoNotTrack support
On Sun, 2012-05-20 at 13:39 +0100, Chris Young wrote:
> Attached simple patch for NetSurf to send the DoNotTrack header.
> We are supposed to be setting a DOM document property too* but I'm not
> sure how to do this. I guess it isn't relevant until we get
> Javascript anyway.
Looks fine to me.
J.
> Attached simple patch for NetSurf to send the DoNotTrack header.
> We are supposed to be setting a DOM document property too* but I'm not
> sure how to do this. I guess it isn't relevant until we get
> Javascript anyway.
Looks fine to me.
J.
Saturday, 19 May 2012
Re: [Rpcemu] rpcemu on Ubuntu 12.04LTS ctrl-end not working
In message <4FB522BE.1050204@gmail.com>
RDS <skeggrd2@gmail.com> wrote:
> It all seems to work fine except... I can't bring the
> configuration menu up with ctrl-end.
I remember this being reported before and it was due to some
desktop option grabbing the ctrl key before RPCemu got to see it.
A search through old messages turned up this:
> > > There's an option "Show position of pointer when the Control
> > > key is pressed" in the "General" pane of the Mouse
> > > Accessibility options in GNOME that causes the Ctrl key to
> > > behave as you're describing... do animated circles appear
> > > around the mouse pointer when you press the Ctrl key by any
> > > chance? ;)
Does that help?
Bryan.
--
RISC OS User Group Of London - http://www.rougol.jellybaby.net/
RISC OS London Show - http://www.riscoslondonshow.co.uk/
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
RDS <skeggrd2@gmail.com> wrote:
> It all seems to work fine except... I can't bring the
> configuration menu up with ctrl-end.
I remember this being reported before and it was due to some
desktop option grabbing the ctrl key before RPCemu got to see it.
A search through old messages turned up this:
> > > There's an option "Show position of pointer when the Control
> > > key is pressed" in the "General" pane of the Mouse
> > > Accessibility options in GNOME that causes the Ctrl key to
> > > behave as you're describing... do animated circles appear
> > > around the mouse pointer when you press the Ctrl key by any
> > > chance? ;)
Does that help?
Bryan.
--
RISC OS User Group Of London - http://www.rougol.jellybaby.net/
RISC OS London Show - http://www.riscoslondonshow.co.uk/
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
Friday, 18 May 2012
Re: More examples using libcss [Xpost from netsurf-users]
On Fri, 2012-05-18 at 16:55 +0530, Syed Mushtaq wrote:
> Maybe I am getting the example wrong. So you begin by giving creating
> a selector context with the node to be search as the selector ( h1 to
> h7 ) in the example. At this point the parse tree would have already
> been created. The function css_select_style() stores the result of the
> search in a css_select_results * . Now I am unable to understand why
> we pass a handler struct which contains various function pointers.
> When will those functions get called ?
LibCSS knows nothing about your document structure, but it needs to be
able to move between nodes in the document when working out which style
rules to apply to the target element. To achieve this, it requires you
provide it with a handler struct containing pointers to functions it may
use when considering selectors.
J.
> Maybe I am getting the example wrong. So you begin by giving creating
> a selector context with the node to be search as the selector ( h1 to
> h7 ) in the example. At this point the parse tree would have already
> been created. The function css_select_style() stores the result of the
> search in a css_select_results * . Now I am unable to understand why
> we pass a handler struct which contains various function pointers.
> When will those functions get called ?
LibCSS knows nothing about your document structure, but it needs to be
able to move between nodes in the document when working out which style
rules to apply to the target element. To achieve this, it requires you
provide it with a handler struct containing pointers to functions it may
use when considering selectors.
J.
Re: More examples using libcss [Xpost from netsurf-users]
On Fri, May 18, 2012 at 4:46 PM, John-Mark Bell <jmb@netsurf-browser.org> wrote:
Maybe I am getting the example wrong. So you begin by giving creating a selector context with the node to be search as the selector ( h1 to h7 ) in the example. At this point the parse tree would have already been created. The function css_select_style() stores the result of the search in a css_select_results * . Now I am unable to understand why we pass a handler struct which contains various function pointers. When will those functions get called ?LibCSS does not provide this kind of API. Its parser is to be consideredOn Fri, 2012-05-18 at 16:39 +0530, Syed Mushtaq wrote:
>
> Hi,
>
> I wanted to write a css spriter using libcss to parse css files. I was
> looking at the example provided in the code, I find that the example
> is very limited and not very clear. Is there some place where I can
> find more examples which use libcss. What I am looking at is some kind
> of event based API where my functions get called when parsing a
> ruleset,rule, etc. Similar to libcroco.
a black box by the outside world. The existing example demonstrates
using LibCSS to parse a stylesheet and then apply the resultant style
information to a document tree. What is unclear about it?
J
Thanks
-Syed
Re: More examples using libcss [Xpost from netsurf-users]
On Fri, 2012-05-18 at 16:39 +0530, Syed Mushtaq wrote:
>
> Hi,
>
> I wanted to write a css spriter using libcss to parse css files. I was
> looking at the example provided in the code, I find that the example
> is very limited and not very clear. Is there some place where I can
> find more examples which use libcss. What I am looking at is some kind
> of event based API where my functions get called when parsing a
> ruleset,rule, etc. Similar to libcroco.
LibCSS does not provide this kind of API. Its parser is to be considered
a black box by the outside world. The existing example demonstrates
using LibCSS to parse a stylesheet and then apply the resultant style
information to a document tree. What is unclear about it?
J.
>
> Hi,
>
> I wanted to write a css spriter using libcss to parse css files. I was
> looking at the example provided in the code, I find that the example
> is very limited and not very clear. Is there some place where I can
> find more examples which use libcss. What I am looking at is some kind
> of event based API where my functions get called when parsing a
> ruleset,rule, etc. Similar to libcroco.
LibCSS does not provide this kind of API. Its parser is to be considered
a black box by the outside world. The existing example demonstrates
using LibCSS to parse a stylesheet and then apply the resultant style
information to a document tree. What is unclear about it?
J.
More examples using libcss [Xpost from netsurf-users]
Hi,
I wanted to write a css spriter using libcss to parse css files. I was looking at the example provided in the code, I find that the example is very limited and not very clear. Is there some place where I can find more examples which use libcss. What I am looking at is some kind of event based API where my functions get called when parsing a ruleset,rule, etc. Similar to libcroco.
Thanks
-Syed
I wanted to write a css spriter using libcss to parse css files. I was looking at the example provided in the code, I find that the example is very limited and not very clear. Is there some place where I can find more examples which use libcss. What I am looking at is some kind of event based API where my functions get called when parsing a ruleset,rule, etc. Similar to libcroco.
Thanks
-Syed
Re: More examples using libcss
Sure. My apologies
On Fri, May 18, 2012 at 4:36 PM, David J. Ruck <druck@druck.org.uk> wrote:
Please can the netsurf developers list be used for development issues, and the user list only for user feedback.
Re: More examples using libcss
On 18/05/2012 11:51, Syed Mushtaq wrote:
> I wanted to write a css spriter using libcss to parse css files. I was
> looking at the example provided in the code, I find that the example is
> very limited and not very clear. Is there some place where I can find
> more examples which use libcss. What I am looking at is some kind of
> event based API where my functions get called when parsing a
> ruleset,rule, etc. Similar to libcroco.
Please can the netsurf developers list be used for development issues,
and the user list only for user feedback.
--
David J. Ruck
email: druck@druck.org.uk
phone: +44(0)7974 108301
> I wanted to write a css spriter using libcss to parse css files. I was
> looking at the example provided in the code, I find that the example is
> very limited and not very clear. Is there some place where I can find
> more examples which use libcss. What I am looking at is some kind of
> event based API where my functions get called when parsing a
> ruleset,rule, etc. Similar to libcroco.
Please can the netsurf developers list be used for development issues,
and the user list only for user feedback.
--
David J. Ruck
email: druck@druck.org.uk
phone: +44(0)7974 108301
More examples using libcss
Hi,
I wanted to write a css spriter using libcss to parse css files. I was looking at the example provided in the code, I find that the example is very limited and not very clear. Is there some place where I can find more examples which use libcss. What I am looking at is some kind of event based API where my functions get called when parsing a ruleset,rule, etc. Similar to libcroco.
Thanks
-Syed
I wanted to write a css spriter using libcss to parse css files. I was looking at the example provided in the code, I find that the example is very limited and not very clear. Is there some place where I can find more examples which use libcss. What I am looking at is some kind of event based API where my functions get called when parsing a ruleset,rule, etc. Similar to libcroco.
Thanks
-Syed
Thursday, 17 May 2012
[Rpcemu] rpcemu on Ubuntu 12.04LTS ctrl-end not working
Hi there
I hope someone can help with problem :-)
I've successfully installed rpcemu 0.8.9 on the latest ubuntu release on
an Advent (dixon's own brand - I was given it!)
It all seems to work fine except... I can't bring the configuration menu
up with ctrl-end.
What I'm attempting to do is access the networking configuration in
particular. However, it's a problem not being to access the
configuration menu anyway. Short of that if anyone can tell me what
should go into the configuration file that would help.
Many thanks
Richard
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
I hope someone can help with problem :-)
I've successfully installed rpcemu 0.8.9 on the latest ubuntu release on
an Advent (dixon's own brand - I was given it!)
It all seems to work fine except... I can't bring the configuration menu
up with ctrl-end.
What I'm attempting to do is access the networking configuration in
particular. However, it's a problem not being to access the
configuration menu anyway. Short of that if anyone can tell me what
should go into the configuration file that would help.
Many thanks
Richard
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
Wednesday, 16 May 2012
Re: [Rpcemu] RPC Emu on Mac OS x
On 16 May 2012, at 21:02, John Edgley wrote:
> I have tried removing the application and all files from the Mac and
> re-installing from the original USB stick that I bought at the London
> show, but I still get the same.
Hi,
You might not have deleted everything - RPCEmu keeps some files in the standard OS X places, which aren't easy to find if you don't know where they are: ~/Library/Preferences/org.devrx.RPCEmu.plist and ~/Library/Application Support/RPCEmu (where ~ is your home directory).
What I'd recommend is:
First try this:
1a) open RPCEmu, choose RPCEmu -> Preferences from the menu at the top of the screen.
1b) in the preferences window, click 'Choose...' to the right of 'Data Directory', and choose your data directory (the one with roms and rpc.cfg etc in it)
1c) quit RPCEmu and open it again
If that doesn't work try this, to completely start from scratch:
2a) delete ~/Library/Preferences/org.devrx.RPCEmu.plist and ~/Library/Application Support/RPCEmu
If you can't see Library that could be because it is hidden by default (I think). You can open it by launching Terminal and typing:
open ~/Library
This should open the Library folder in Finder
2b) re-install RPCEmu
Hope this helps,
Francis
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
> I have tried removing the application and all files from the Mac and
> re-installing from the original USB stick that I bought at the London
> show, but I still get the same.
Hi,
You might not have deleted everything - RPCEmu keeps some files in the standard OS X places, which aren't easy to find if you don't know where they are: ~/Library/Preferences/org.devrx.RPCEmu.plist and ~/Library/Application Support/RPCEmu (where ~ is your home directory).
What I'd recommend is:
First try this:
1a) open RPCEmu, choose RPCEmu -> Preferences from the menu at the top of the screen.
1b) in the preferences window, click 'Choose...' to the right of 'Data Directory', and choose your data directory (the one with roms and rpc.cfg etc in it)
1c) quit RPCEmu and open it again
If that doesn't work try this, to completely start from scratch:
2a) delete ~/Library/Preferences/org.devrx.RPCEmu.plist and ~/Library/Application Support/RPCEmu
If you can't see Library that could be because it is hidden by default (I think). You can open it by launching Terminal and typing:
open ~/Library
This should open the Library folder in Finder
2b) re-install RPCEmu
Hope this helps,
Francis
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
Tuesday, 15 May 2012
[gccsdk] [Bug 238] Temporary filenames with RISC OS filename semantics
http://www.riscos.info/bugzilla3/show_bug.cgi?id=238
John Tytgat <John.Tytgat@aaug.net> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #1 from John Tytgat <John.Tytgat@aaug.net> 2012-05-15 17:55:47 PDT ---
Solved with r5822 (trunk/gcc4) and r5823 (branches/release_4_1_2).
--
Configure bugmail: http://www.riscos.info/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
John Tytgat <John.Tytgat@aaug.net> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
--- Comment #1 from John Tytgat <John.Tytgat@aaug.net> 2012-05-15 17:55:47 PDT ---
Solved with r5822 (trunk/gcc4) and r5823 (branches/release_4_1_2).
--
Configure bugmail: http://www.riscos.info/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Re: No comment...
Harriet Bazley <lists@orange.wingsandbeaks.org.uk> wrote:
>Yes, but how? Can you 'cascade' style sheets locally over a remote page?
I think so, in some browsers, dunno about NS.
--
Jeremy C B Nicoll - my opinions are my own.
>Yes, but how? Can you 'cascade' style sheets locally over a remote page?
I think so, in some browsers, dunno about NS.
--
Jeremy C B Nicoll - my opinions are my own.
Re: No comment...
On 15 May 2012 as I do recall,
Jeremy Nicoll - ml netsurf wrote:
> Harriet Bazley <lists@orange.wingsandbeaks.org.uk> wrote:
>
> > Edit: just found it. Right down the bottom:
> >
> ><style type='text/css'>
> > #comments {display:none;}
> > </style>
> >
> > *sigh* Now why would anyone do that?
>
> Maybe it's the equivalent of clearing a bit flag before then setting some of
> the bits... That is, it looks daft in isolation but mayb eisn't if you can
> see whatever affects it as well... and something must allow display because
> I can see comments (in Firefox).
Well, looking at it, the entire comment content appears to be duplicated
in JavaScript. So I suppose they must be switching 'off' the normal
display with the purpose of providing some 'enhanced' JS version.
Unfortunately they forgot to make this conditional within the script
section, presumably never having tested it without JS active....
>
> > And is there any way of overriding it?
>
> Overriding the CSS?
>
Yes, but how? Can you 'cascade' style sheets locally over a remote page?
--
Harriet Bazley == Loyaulte me lie ==
Some people cause happiness wherever they go; others, whenever they go.
Jeremy Nicoll - ml netsurf wrote:
> Harriet Bazley <lists@orange.wingsandbeaks.org.uk> wrote:
>
> > Edit: just found it. Right down the bottom:
> >
> ><style type='text/css'>
> > #comments {display:none;}
> > </style>
> >
> > *sigh* Now why would anyone do that?
>
> Maybe it's the equivalent of clearing a bit flag before then setting some of
> the bits... That is, it looks daft in isolation but mayb eisn't if you can
> see whatever affects it as well... and something must allow display because
> I can see comments (in Firefox).
Well, looking at it, the entire comment content appears to be duplicated
in JavaScript. So I suppose they must be switching 'off' the normal
display with the purpose of providing some 'enhanced' JS version.
Unfortunately they forgot to make this conditional within the script
section, presumably never having tested it without JS active....
>
> > And is there any way of overriding it?
>
> Overriding the CSS?
>
Yes, but how? Can you 'cascade' style sheets locally over a remote page?
--
Harriet Bazley == Loyaulte me lie ==
Some people cause happiness wherever they go; others, whenever they go.
Re: No comment...
Harriet Bazley <lists@orange.wingsandbeaks.org.uk> wrote:
> Edit: just found it. Right down the bottom:
>
><style type='text/css'>
> #comments {display:none;}
> </style>
>
> *sigh* Now why would anyone do that?
Maybe it's the equivalent of clearing a bit flag before then setting some of
the bits... That is, it looks daft in isolation but mayb eisn't if you can
see whatever affects it as well... and something must allow display because
I can see comments (in Firefox).
> And is there any way of overriding it?
Overriding the CSS?
--
Jeremy C B Nicoll - my opinions are my own.
> Edit: just found it. Right down the bottom:
>
><style type='text/css'>
> #comments {display:none;}
> </style>
>
> *sigh* Now why would anyone do that?
Maybe it's the equivalent of clearing a bit flag before then setting some of
the bits... That is, it looks daft in isolation but mayb eisn't if you can
see whatever affects it as well... and something must allow display because
I can see comments (in Firefox).
> And is there any way of overriding it?
Overriding the CSS?
--
Jeremy C B Nicoll - my opinions are my own.
No comment...
What is hiding the comments e.g. on this blog http://ofdreamsandseams.blogspot.co.uk/2012/05/9-shirts-off-his-back.html
from Netsurf?
JavaScript seems to be involved, but the actual body of the comment text
appears in the form
<li id='bc_0_8B' class='comment' kind='b'><div class='avatar-image-container'><img src='http://img1.blogblog.com/img/anon36.png'></img></div><div id='c3866699261828389934' class='comment-block'><div id='bc_0_8M' class='comment-header' kind='m'><cite class='user'><a rel='nofollow' href='http://alittleboutiquenearhome.com'>Laura</a></cite><span class='icon user'></span><span class='datetime secondary-text'><a rel='nofollow' href='http://ofdreamsandseams.blogspot.com/2012/05/9-shirts-off-his-back.html?showComment=1336690576841#c3866699261828389934'>May 11, 2012 12:56 AM</a></span></div><p id='bc_0_8MC' class='comment-content'>I absolutely adore this! Nice job!</p><span id='bc_0_8MN' class='comment-actions secondary-text' kind='m'><a kind='i' href='javascript:;' target='_self' o='r'>Reply</a><span class='item-control blog-admin pid-548093041'><a o='d' target='_self' href='http://www.blogger.com/delete-comment.g?blogID=1991608771892905541&postID=3866699261828389934'>Delete</a></span></span></div><div id='bc_0_8BR' class='comment-replies'></div><div id='bc_0_8B_box' class='comment-replybox-single'></div></li>
and indeed if I select and drag out the entire <div id='comment-holder'>
section from the source, then double-click on the selection, I get an
ordinary Ordered List. So what is being used to hide the text?
Edit: just found it. Right down the bottom:
<style type='text/css'>
#comments {display:none;}
</style>
*sigh* Now why would anyone do that? And is there any way of
overriding it?
--
Harriet Bazley == Loyaulte me lie ==
Cloning is the sincerest form of flattery.
from Netsurf?
JavaScript seems to be involved, but the actual body of the comment text
appears in the form
<li id='bc_0_8B' class='comment' kind='b'><div class='avatar-image-container'><img src='http://img1.blogblog.com/img/anon36.png'></img></div><div id='c3866699261828389934' class='comment-block'><div id='bc_0_8M' class='comment-header' kind='m'><cite class='user'><a rel='nofollow' href='http://alittleboutiquenearhome.com'>Laura</a></cite><span class='icon user'></span><span class='datetime secondary-text'><a rel='nofollow' href='http://ofdreamsandseams.blogspot.com/2012/05/9-shirts-off-his-back.html?showComment=1336690576841#c3866699261828389934'>May 11, 2012 12:56 AM</a></span></div><p id='bc_0_8MC' class='comment-content'>I absolutely adore this! Nice job!</p><span id='bc_0_8MN' class='comment-actions secondary-text' kind='m'><a kind='i' href='javascript:;' target='_self' o='r'>Reply</a><span class='item-control blog-admin pid-548093041'><a o='d' target='_self' href='http://www.blogger.com/delete-comment.g?blogID=1991608771892905541&postID=3866699261828389934'>Delete</a></span></span></div><div id='bc_0_8BR' class='comment-replies'></div><div id='bc_0_8B_box' class='comment-replybox-single'></div></li>
and indeed if I select and drag out the entire <div id='comment-holder'>
section from the source, then double-click on the selection, I get an
ordinary Ordered List. So what is being used to hide the text?
Edit: just found it. Right down the bottom:
<style type='text/css'>
#comments {display:none;}
</style>
*sigh* Now why would anyone do that? And is there any way of
overriding it?
--
Harriet Bazley == Loyaulte me lie ==
Cloning is the sincerest form of flattery.
Friday, 11 May 2012
Re: [gccsdk] fork/exec problems
In message <b102448d52.Jo@hobbes.bass-software.com>
John Tytgat <John.Tytgat@aaug.net> wrote:
> In message <e420538c52.beeb@ron1954.woosh.co.nz>
> Ron <beeb@woosh.co.nz> wrote:
>
> > I dont know if unixlib's version of fork() does this "copy on
> > write" or not, or any other details,
>
> As I mentioned in another mail, UnixLib's implementation is rather
> simple here. It just temporary moves parent out of the way (but in
> the same application workspace), runs the child and then you have
> the return to the parent.
>
> > just that fork()/exec has been
> > causing unstability in the past a (few months ago) and now with
> > latest svn upgrades stops program completely at the fork()/exec.
> > The earlier situation was erratic and I think would be difficult
> > to build a small test case to show it happening. It possibly is
> > only happening while running a largish program.
> >
> > I am getting a better picture of things now, for example I know
> > now that it is fork()/exec causing unstability, and there are
> > obvious advantages to using vfork()/exec in maybe all cases.
>
> IMHO you're drawing the wrong conclusion. fork+exec on its own is not
> causing any unstability by definition. It is just for your case hitting a
> currently unknown problem. That might be in your program, use, setup
> and/or in UnixLib. As long there isn't attempt made to drill down to the
> real issue, we're unable to draw any valid conclusion at all.
>
> John.
Taking into consideration your point that fork() is probably not buggy,
and the fact that vfork() is working OK, and that the two methods use
the memory allocation differently, I will try some variations in
this dept.
The second day I used vfork()/exec I run into sigsev error at the point,
similar to the latest fork()/exec, but it turned out I needed a larger
wimpslot. That was puzzling because I was performing the same operation
on the same files as the previous day.
I won't rule out the possibility of running the binary in a different
manner, for example the running of Tar from an Obey file that is run
from a Taskwindow compared to running it more directly, and I have run
it from a wimptask also. Checking for results on my RiscPc 3.7 should
do things differently too.
It is less frantic trying things out now that I have a bulletproof
version using vfork()/exec (save for the guessing game for a wimpslot).
One thing that was not very manageable by using fork()/exec was the
incrementing of memory used by about 740K every time it was used.
So a program that can run in say, 1280K using vfork() alternatively
would swallow up over 32MB with 50 or so calls to fork()/exec
(speculation based on what happened in the Task display)
It is possible that tar is not closing the script file each time it
is run from the fork(0)/exec and not releasing the memory.
Thanks for your tips.
Ron M.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
John Tytgat <John.Tytgat@aaug.net> wrote:
> In message <e420538c52.beeb@ron1954.woosh.co.nz>
> Ron <beeb@woosh.co.nz> wrote:
>
> > I dont know if unixlib's version of fork() does this "copy on
> > write" or not, or any other details,
>
> As I mentioned in another mail, UnixLib's implementation is rather
> simple here. It just temporary moves parent out of the way (but in
> the same application workspace), runs the child and then you have
> the return to the parent.
>
> > just that fork()/exec has been
> > causing unstability in the past a (few months ago) and now with
> > latest svn upgrades stops program completely at the fork()/exec.
> > The earlier situation was erratic and I think would be difficult
> > to build a small test case to show it happening. It possibly is
> > only happening while running a largish program.
> >
> > I am getting a better picture of things now, for example I know
> > now that it is fork()/exec causing unstability, and there are
> > obvious advantages to using vfork()/exec in maybe all cases.
>
> IMHO you're drawing the wrong conclusion. fork+exec on its own is not
> causing any unstability by definition. It is just for your case hitting a
> currently unknown problem. That might be in your program, use, setup
> and/or in UnixLib. As long there isn't attempt made to drill down to the
> real issue, we're unable to draw any valid conclusion at all.
>
> John.
Taking into consideration your point that fork() is probably not buggy,
and the fact that vfork() is working OK, and that the two methods use
the memory allocation differently, I will try some variations in
this dept.
The second day I used vfork()/exec I run into sigsev error at the point,
similar to the latest fork()/exec, but it turned out I needed a larger
wimpslot. That was puzzling because I was performing the same operation
on the same files as the previous day.
I won't rule out the possibility of running the binary in a different
manner, for example the running of Tar from an Obey file that is run
from a Taskwindow compared to running it more directly, and I have run
it from a wimptask also. Checking for results on my RiscPc 3.7 should
do things differently too.
It is less frantic trying things out now that I have a bulletproof
version using vfork()/exec (save for the guessing game for a wimpslot).
One thing that was not very manageable by using fork()/exec was the
incrementing of memory used by about 740K every time it was used.
So a program that can run in say, 1280K using vfork() alternatively
would swallow up over 32MB with 50 or so calls to fork()/exec
(speculation based on what happened in the Task display)
It is possible that tar is not closing the script file each time it
is run from the fork(0)/exec and not releasing the memory.
Thanks for your tips.
Ron M.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Wednesday, 9 May 2012
Re: [gccsdk] fork/exec problems
In message <e420538c52.beeb@ron1954.woosh.co.nz>
Ron <beeb@woosh.co.nz> wrote:
> I dont know if unixlib's version of fork() does this "copy on
> write" or not, or any other details,
As I mentioned in another mail, UnixLib's implementation is rather
simple here. It just temporary moves parent out of the way (but in
the same application workspace), runs the child and then you have
the return to the parent.
> just that fork()/exec has been
> causing unstability in the past a (few months ago) and now with
> latest svn upgrades stops program completely at the fork()/exec.
> The earlier situation was erratic and I think would be difficult
> to build a small test case to show it happening. It possibly is
> only happening while running a largish program.
>
> I am getting a better picture of things now, for example I know
> now that it is fork()/exec causing unstability, and there are
> obvious advantages to using vfork()/exec in maybe all cases.
IMHO you're drawing the wrong conclusion. fork+exec on its own is not
causing any unstability by definition. It is just for your case hitting a
currently unknown problem. That might be in your program, use, setup
and/or in UnixLib. As long there isn't attempt made to drill down to the
real issue, we're unable to draw any valid conclusion at all.
John.
--
John Tytgat, in his comfy chair at home BASS
John.Tytgat@aaug.net ARM powered, RISC OS driven
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Ron <beeb@woosh.co.nz> wrote:
> I dont know if unixlib's version of fork() does this "copy on
> write" or not, or any other details,
As I mentioned in another mail, UnixLib's implementation is rather
simple here. It just temporary moves parent out of the way (but in
the same application workspace), runs the child and then you have
the return to the parent.
> just that fork()/exec has been
> causing unstability in the past a (few months ago) and now with
> latest svn upgrades stops program completely at the fork()/exec.
> The earlier situation was erratic and I think would be difficult
> to build a small test case to show it happening. It possibly is
> only happening while running a largish program.
>
> I am getting a better picture of things now, for example I know
> now that it is fork()/exec causing unstability, and there are
> obvious advantages to using vfork()/exec in maybe all cases.
IMHO you're drawing the wrong conclusion. fork+exec on its own is not
causing any unstability by definition. It is just for your case hitting a
currently unknown problem. That might be in your program, use, setup
and/or in UnixLib. As long there isn't attempt made to drill down to the
real issue, we're unable to draw any valid conclusion at all.
John.
--
John Tytgat, in his comfy chair at home BASS
John.Tytgat@aaug.net ARM powered, RISC OS driven
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
[gccsdk] [Bug 239] opendir() with trailing directory separator and RISC OS filename semantics.
http://www.riscos.info/bugzilla3/show_bug.cgi?id=239
John Tytgat <John.Tytgat@aaug.net> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |WONTFIX
--- Comment #1 from John Tytgat <John.Tytgat@aaug.net> 2012-05-09 09:40:40 PDT ---
I disagree. When you specify __RISCOSIFY_NO_PROCESS you should use valid RISC
OS filename and directory names. "dir." is not a valid RISC OS directory name
(e.g. try "*ex dir" vs "*ex dir."). Moreover, a directory name + "." is a RISC
OS (search) path.
--
Configure bugmail: http://www.riscos.info/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
John Tytgat <John.Tytgat@aaug.net> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |WONTFIX
--- Comment #1 from John Tytgat <John.Tytgat@aaug.net> 2012-05-09 09:40:40 PDT ---
I disagree. When you specify __RISCOSIFY_NO_PROCESS you should use valid RISC
OS filename and directory names. "dir." is not a valid RISC OS directory name
(e.g. try "*ex dir" vs "*ex dir."). Moreover, a directory name + "." is a RISC
OS (search) path.
--
Configure bugmail: http://www.riscos.info/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
re: netsurf-sdl on Windows?
John-Mark Bell wrote:
>Please do the following:
>
> 1) make TARGET=windows clean
> 2) make TARGET=windows Q= >buildlog-win 2>&1
> 3) Send the verbatim buildlog-win to me
Before I try tackling the framebuffer again, I tried this with the
Windows native version. I deleted everything in
build-windows-windows. I ran make TARGET=windows clean. I ran make
TARGET=windows Q= and sent stdout and stderr to a log file. I don't
see any errors in that build. John-Mark, will send you a copy of the
file directly instead of taking up space on the mailing list.
> The output binary is placed in the top level of the source tree, not in
> the build-* directory. For the Windows target, it is called NetSurf.exe
I ran a search from c:\mingw down and could find no executables with
the name netsurf.exe or anything similar using n*.exe as the search
pattern.
Thanks to John-Mark and Ole for the assistance and clarifications.
>Please do the following:
>
> 1) make TARGET=windows clean
> 2) make TARGET=windows Q= >buildlog-win 2>&1
> 3) Send the verbatim buildlog-win to me
Before I try tackling the framebuffer again, I tried this with the
Windows native version. I deleted everything in
build-windows-windows. I ran make TARGET=windows clean. I ran make
TARGET=windows Q= and sent stdout and stderr to a log file. I don't
see any errors in that build. John-Mark, will send you a copy of the
file directly instead of taking up space on the mailing list.
> The output binary is placed in the top level of the source tree, not in
> the build-* directory. For the Windows target, it is called NetSurf.exe
I ran a search from c:\mingw down and could find no executables with
the name netsurf.exe or anything similar using n*.exe as the search
pattern.
Thanks to John-Mark and Ole for the assistance and clarifications.
Re: Fw: snag on installing Netsurf
On 8 May 2012 DJ BRADLEY wrote:
> I am in process of installing Netsurf on my Riscpc. I downloaded the
> system on to a CD on my PC computer, then used this to transfer to the
> Risc machine. This method seems to work excpt that I have not
> succeeded with the Boot Merge. I get a well known error message saying
> that an application to load the file has not been found, etc.
> I have not been able to get round this - can you advise please?
Did you copy the NetSurf archive onto the RiscPC first? It's best if
you do this and then remove the CD and reboot before doing the Boot
Merge. This sort of thing can happen if the filer has seen files or
apps on the CD and has mis-set some global variables.
> Alternatively can you expand on "drag the supplied !Boot over your
> existing boot structure"? (quoting from readme)
> I have tried one or two things but have not managed to discover the
> intended procedure.
Just drag the !Boot from the NetSurf archive into the parent directory
of !Boot on HardDisc4 (i.e. the root directory of HardDisc4). You
should have the "newer" filer option set. You don't actually need to
drop it onto the target !Boot icon, but you can.
--
Richard Porter http://www.minijem.plus.com/
mailto:ricp@minijem.plus.com
I don't want a "user experience" - I just want stuff that works.
> I am in process of installing Netsurf on my Riscpc. I downloaded the
> system on to a CD on my PC computer, then used this to transfer to the
> Risc machine. This method seems to work excpt that I have not
> succeeded with the Boot Merge. I get a well known error message saying
> that an application to load the file has not been found, etc.
> I have not been able to get round this - can you advise please?
Did you copy the NetSurf archive onto the RiscPC first? It's best if
you do this and then remove the CD and reboot before doing the Boot
Merge. This sort of thing can happen if the filer has seen files or
apps on the CD and has mis-set some global variables.
> Alternatively can you expand on "drag the supplied !Boot over your
> existing boot structure"? (quoting from readme)
> I have tried one or two things but have not managed to discover the
> intended procedure.
Just drag the !Boot from the NetSurf archive into the parent directory
of !Boot on HardDisc4 (i.e. the root directory of HardDisc4). You
should have the "newer" filer option set. You don't actually need to
drop it onto the target !Boot icon, but you can.
--
Richard Porter http://www.minijem.plus.com/
mailto:ricp@minijem.plus.com
I don't want a "user experience" - I just want stuff that works.
re: netsurf-sdl on Windows?
On Tue, 2012-05-08 at 09:44 -0400, LM wrote:
> > Why not use the native Windows frontend?
> >
> > The framebuffer frontend is totally unsuitable for use on systems that
> > have a full-featured window manager.
>
> I was hoping to use the SDL version on a couple of platforms including
> Windows. I suppose I could build each one natively though. If I can
> get it to build, long-term plan was not to use it as a web browser,
> but more of a cross-platform replacement for dialog. Was thinking I
> could possibly replace the part that sends form data to a server with
> the ability to store the information locally to a file. I wouldn't
> even need the web access capabilities, just the capabilities to
> interpret HTML and CSS and render pages. There are cross-platform
> replacements for dialog, but I'd really like to have something that
> can use standard web syntax (HTML, CSS, etc.) for creating interfaces.
I really don't think NetSurf is suited to this scenario at all. There
are many assumptions within NetSurf that there is a server to talk to.
You'd probably be better off looking for a standalone rendering engine,
which NetSurf is not (and is unlikely to be in the short to medium
term).
> When I run make TARGET=windows I see no error messages. I redirected
> standard output and standard error to a file.
The reason I asked for the exact build output is that it saves us both
time and speculation. Please do the following:
1) make TARGET=windows clean
2) make TARGET=windows Q= >buildlog-win 2>&1
3) Send the verbatim buildlog-win to me
> Checking the build-windows-windows directory, I see object files
> (108 of them). There are no libraries or executables built.
The output binary is placed in the top level of the source tree, not in
the build-* directory. For the Windows target, it is called NetSurf.exe
> When I tried running make TARGET=framebuffer, I got several errors
> back. Some of them include:
> image/mng.c:29:20: fatal error: libmng.h: No such file or directory
> [ I thought I had MNG disabled, but appears framebuffer mode still
> wants the library installed. ]
It uses the same configuration file as all the other frontends, so it's
unlikely to explicitly want it. How have you disabled MNG?
> framebuffer/thumbnail.c:66:16: error: 'NSFB_SURFACE_RAM' undeclared
> (first use in this function)
> framebuffer/thumbnail.c:71:41: error: 'NSFB_FMT_XBGR8888' undeclared
> (first use in this function)
> framebuffer/thumbnail.c:73:2: error: incompatible type for argument 1
> of 'nsfb_init'
> framebuffer/thumbnail.c:90:2: error: too many arguments to function
> 'nsfb_plot_copy'
These failures are symptomatic of building against the wrong version of
libnsfb. If you are building from a release tarball, then the correct
libraries are bundled with it. If you are building from an SVN snapshot,
please ensure that you have snapshots of the libraries from the same
revision.
> framebuffer/convert_image.c:22:17: fatal error: png.h: No such file or directory
> [ My png.h is in /usr/local/include. I'm guessing if this was
> switched from <png.h> to "png.h", it'll find it. ]
No, it won't. You need to ensure that /usr/local/include is on the
include path (or, ideally, ensure that pkg-config reports the correct
paths when asked)
> cc.exe: error: build-windows-framebuffer/caret_image.c: No such file
> or directory
> cc.exe: fatal error: no input files
> compilation terminated.
> make: *** [build-windows-framebuffer/build-windows-framebuffer_caret_image.o]
> Error 1
Again, please do the following:
1) make TARGET=framebuffer clean
2) make TARGET=framebuffer Q= >buildlog-fb 2>&1
3) Send the verbatim buildlog-fb to me
John-Mark.
> > Why not use the native Windows frontend?
> >
> > The framebuffer frontend is totally unsuitable for use on systems that
> > have a full-featured window manager.
>
> I was hoping to use the SDL version on a couple of platforms including
> Windows. I suppose I could build each one natively though. If I can
> get it to build, long-term plan was not to use it as a web browser,
> but more of a cross-platform replacement for dialog. Was thinking I
> could possibly replace the part that sends form data to a server with
> the ability to store the information locally to a file. I wouldn't
> even need the web access capabilities, just the capabilities to
> interpret HTML and CSS and render pages. There are cross-platform
> replacements for dialog, but I'd really like to have something that
> can use standard web syntax (HTML, CSS, etc.) for creating interfaces.
I really don't think NetSurf is suited to this scenario at all. There
are many assumptions within NetSurf that there is a server to talk to.
You'd probably be better off looking for a standalone rendering engine,
which NetSurf is not (and is unlikely to be in the short to medium
term).
> When I run make TARGET=windows I see no error messages. I redirected
> standard output and standard error to a file.
The reason I asked for the exact build output is that it saves us both
time and speculation. Please do the following:
1) make TARGET=windows clean
2) make TARGET=windows Q= >buildlog-win 2>&1
3) Send the verbatim buildlog-win to me
> Checking the build-windows-windows directory, I see object files
> (108 of them). There are no libraries or executables built.
The output binary is placed in the top level of the source tree, not in
the build-* directory. For the Windows target, it is called NetSurf.exe
> When I tried running make TARGET=framebuffer, I got several errors
> back. Some of them include:
> image/mng.c:29:20: fatal error: libmng.h: No such file or directory
> [ I thought I had MNG disabled, but appears framebuffer mode still
> wants the library installed. ]
It uses the same configuration file as all the other frontends, so it's
unlikely to explicitly want it. How have you disabled MNG?
> framebuffer/thumbnail.c:66:16: error: 'NSFB_SURFACE_RAM' undeclared
> (first use in this function)
> framebuffer/thumbnail.c:71:41: error: 'NSFB_FMT_XBGR8888' undeclared
> (first use in this function)
> framebuffer/thumbnail.c:73:2: error: incompatible type for argument 1
> of 'nsfb_init'
> framebuffer/thumbnail.c:90:2: error: too many arguments to function
> 'nsfb_plot_copy'
These failures are symptomatic of building against the wrong version of
libnsfb. If you are building from a release tarball, then the correct
libraries are bundled with it. If you are building from an SVN snapshot,
please ensure that you have snapshots of the libraries from the same
revision.
> framebuffer/convert_image.c:22:17: fatal error: png.h: No such file or directory
> [ My png.h is in /usr/local/include. I'm guessing if this was
> switched from <png.h> to "png.h", it'll find it. ]
No, it won't. You need to ensure that /usr/local/include is on the
include path (or, ideally, ensure that pkg-config reports the correct
paths when asked)
> cc.exe: error: build-windows-framebuffer/caret_image.c: No such file
> or directory
> cc.exe: fatal error: no input files
> compilation terminated.
> make: *** [build-windows-framebuffer/build-windows-framebuffer_caret_image.o]
> Error 1
Again, please do the following:
1) make TARGET=framebuffer clean
2) make TARGET=framebuffer Q= >buildlog-fb 2>&1
3) Send the verbatim buildlog-fb to me
John-Mark.
Re: Fw: snag on installing Netsurf
On Tue, 2012-05-08 at 16:08 +0100, DJ BRADLEY wrote:
> I am in process of installing Netsurf on my Riscpc. I downloaded the
> system on to a CD on my PC computer, then used this to transfer to the
> Risc machine. This method seems to work excpt that I have not
> succeeded with the Boot Merge. I get a well known error message saying
> that an application to load the file has not been found, etc.
>
> I have not been able to get round this - can you advise please?
The above sounds very wrong. What, exactly, did you do when attempting
the Boot Merge operation?
Resource installation is documented here:
http://www.netsurf-browser.org/documentation/resinstall
> Alternatively can you expand on "drag the supplied !Boot over your
> existing boot structure"? (quoting from readme)
As it says: drag the !Boot from the NetSurf archive over the !Boot in
the root of your hard drive.
John-Mark.
> I am in process of installing Netsurf on my Riscpc. I downloaded the
> system on to a CD on my PC computer, then used this to transfer to the
> Risc machine. This method seems to work excpt that I have not
> succeeded with the Boot Merge. I get a well known error message saying
> that an application to load the file has not been found, etc.
>
> I have not been able to get round this - can you advise please?
The above sounds very wrong. What, exactly, did you do when attempting
the Boot Merge operation?
Resource installation is documented here:
http://www.netsurf-browser.org/documentation/resinstall
> Alternatively can you expand on "drag the supplied !Boot over your
> existing boot structure"? (quoting from readme)
As it says: drag the !Boot from the NetSurf archive over the !Boot in
the root of your hard drive.
John-Mark.
Tuesday, 8 May 2012
re: netsurf-sdl on Windows?
Am Dienstag, den 08.05.2012, 09:44 +0200 schrieb LM <lmemsm@gmail.com>:
> Made a few changes to the makefile and a few other files and it looks
> like it builds the library fine on Windows.
okay... you can also run "make test" or something like that to make sure
that it is really
compiled okay.
> I don't specifically know what the issue was when I tried to build
> netsurf-2.8 versus netsurf-r13571, but I appear to have my include
...
> both, but something is different. Since I appear to be having less
> issues with the netsurf-r13571 tarball, I'm concentrating on trying to
> build using it.
I think it is best to build the release versions. stick to that.
> When I run make TARGET=windows I see no error messages. I redirected
> standard output and standard error to a file. Looking over the file,
> there are just warnings. Checking the build-windows-windows
> directory, I see object files (108 of them). There are no libraries
> or executables built. I did go in and hardcode $userinfo{USERNAME},
> $gecos and $hostname in svn-testament.pl since it was having some
> trouble running.
>
> TESTMENT: unchanged
> If I run make TARGET=windows again, those lines are all I get back.
try to make clean TARGET=windows or make clean TARGET=framebuffer.
> When I tried running make TARGET=framebuffer, I got several errors
> back. Some of them include:
> image/mng.c:29:20: fatal error: libmng.h: No such file or directory
libmng is not installed / cannot be found. Either install libmng at the
correct place, or
do the same with libpng and tell the netsurf build system to use libpng...
( study the makesfiles in top level directoy...)
> [ I thought I had MNG disabled, but appears framebuffer mode still
> wants the library installed. ]
ok... maybe you did it wrong.
> framebuffer/thumbnail.c:66:16: error: 'NSFB_SURFACE_RAM' undeclared
> (first use in this function)
maybe you modified the libnsfb makefiles a bit too much? The RAM surface
must be available, also the SDL one...
> framebuffer/thumbnail.c:71:41: error: 'NSFB_FMT_XBGR8888' undeclared
> (first use in this function)
strange. Maybe libnsfb isn't placed / installed in the correct location?
First, run the libnsfb tests ( read the docs...) - then install it properly.
> framebuffer/thumbnail.c:73:2: error: incompatible type for argument 1
> of 'nsfb_init'
> framebuffer/thumbnail.c:90:2: error: too many arguments to function
> 'nsfb_plot_copy'
> framebuffer/localhistory.c:22:23: fatal error: sys/ioctl.h: No such
> file or directory
Looks like an problem with the sources... or maybe your mingw installation
is missing something?
> The errors look fixable. Don't know whether I should try to get the
> framebuffer version to compile without errors or concentrate on why
I'm unsure if anyone has tested the SDL frontend under Windows. So that's
maybe a problem.
Greets,
Ole
> Made a few changes to the makefile and a few other files and it looks
> like it builds the library fine on Windows.
okay... you can also run "make test" or something like that to make sure
that it is really
compiled okay.
> I don't specifically know what the issue was when I tried to build
> netsurf-2.8 versus netsurf-r13571, but I appear to have my include
...
> both, but something is different. Since I appear to be having less
> issues with the netsurf-r13571 tarball, I'm concentrating on trying to
> build using it.
I think it is best to build the release versions. stick to that.
> When I run make TARGET=windows I see no error messages. I redirected
> standard output and standard error to a file. Looking over the file,
> there are just warnings. Checking the build-windows-windows
> directory, I see object files (108 of them). There are no libraries
> or executables built. I did go in and hardcode $userinfo{USERNAME},
> $gecos and $hostname in svn-testament.pl since it was having some
> trouble running.
>
> TESTMENT: unchanged
> If I run make TARGET=windows again, those lines are all I get back.
try to make clean TARGET=windows or make clean TARGET=framebuffer.
> When I tried running make TARGET=framebuffer, I got several errors
> back. Some of them include:
> image/mng.c:29:20: fatal error: libmng.h: No such file or directory
libmng is not installed / cannot be found. Either install libmng at the
correct place, or
do the same with libpng and tell the netsurf build system to use libpng...
( study the makesfiles in top level directoy...)
> [ I thought I had MNG disabled, but appears framebuffer mode still
> wants the library installed. ]
ok... maybe you did it wrong.
> framebuffer/thumbnail.c:66:16: error: 'NSFB_SURFACE_RAM' undeclared
> (first use in this function)
maybe you modified the libnsfb makefiles a bit too much? The RAM surface
must be available, also the SDL one...
> framebuffer/thumbnail.c:71:41: error: 'NSFB_FMT_XBGR8888' undeclared
> (first use in this function)
strange. Maybe libnsfb isn't placed / installed in the correct location?
First, run the libnsfb tests ( read the docs...) - then install it properly.
> framebuffer/thumbnail.c:73:2: error: incompatible type for argument 1
> of 'nsfb_init'
> framebuffer/thumbnail.c:90:2: error: too many arguments to function
> 'nsfb_plot_copy'
> framebuffer/localhistory.c:22:23: fatal error: sys/ioctl.h: No such
> file or directory
Looks like an problem with the sources... or maybe your mingw installation
is missing something?
> The errors look fixable. Don't know whether I should try to get the
> framebuffer version to compile without errors or concentrate on why
I'm unsure if anyone has tested the SDL frontend under Windows. So that's
maybe a problem.
Greets,
Ole
Fw: snag on installing Netsurf
----- Forwarded Message -----
From: DJ BRADLEY <djbradley909@btinternet.com>
To: "bugs@netsurf-browser.org" <bugs@netsurf-browser.org>
Sent: Monday, 7 May 2012, 21:50
Subject:
From: DJ BRADLEY <djbradley909@btinternet.com>
To: "bugs@netsurf-browser.org" <bugs@netsurf-browser.org>
Sent: Monday, 7 May 2012, 21:50
Subject:
I am in process of installing Netsurf on my Riscpc. I downloaded the system on to a CD on my PC computer, then used this to transfer to the Risc machine. This method seems to work excpt that I have not succeeded with the Boot Merge. I get a well known error message saying that an application to load the file has not been found, etc.
I have not been able to get round this - can you advise please?
Alternatively can you expand on "drag the supplied !Boot over your existing boot structure"? (quoting from readme)
I have tried one or two things but have not managed to discover the intended procedure.
Many thanks and regards
Don Bradley (above was undelivered)
re: netsurf-sdl on Windows?
From: Ole
> make TARGET=framebuffer
>
> the "make TARGET=windows" will attempt to build the native windows
> executable.
Thanks. TARGET=framebuffer is what I was looking for.
>
>
> The framebuffer library "libnsfb" requires X11 libraries...
> I once submitted a patch to the list that disables specific libnsfb
> backends.
> Just remove the X and maybe other not-so-important dependencies from
> the libnsfb
> makefiles and you should be able to compile the lib.
> After you have installed libnsfb, you should be able to link
> framebuffer-sdl for windows.
Made a few changes to the makefile and a few other files and it looks
like it builds the library fine on Windows.
From: John-Mark Bell
> Why not use the native Windows frontend?
>
> The framebuffer frontend is totally unsuitable for use on systems that
> have a full-featured window manager.
I was hoping to use the SDL version on a couple of platforms including
Windows. I suppose I could build each one natively though. If I can
get it to build, long-term plan was not to use it as a web browser,
but more of a cross-platform replacement for dialog. Was thinking I
could possibly replace the part that sends form data to a server with
the ability to store the information locally to a file. I wouldn't
even need the web access capabilities, just the capabilities to
interpret HTML and CSS and render pages. There are cross-platform
replacements for dialog, but I'd really like to have something that
can use standard web syntax (HTML, CSS, etc.) for creating interfaces.
>
> What issues? Please provide the exact build output.
>
I don't specifically know what the issue was when I tried to build
netsurf-2.8 versus netsurf-r13571, but I appear to have my include
paths set appropriately with the latter and I can't get them set
properly with the prior. I thought I went through the same steps in
both, but something is different. Since I appear to be having less
issues with the netsurf-r13571 tarball, I'm concentrating on trying to
build using it.
When I run make TARGET=windows I see no error messages. I redirected
standard output and standard error to a file. Looking over the file,
there are just warnings. Checking the build-windows-windows
directory, I see object files (108 of them). There are no libraries
or executables built. I did go in and hardcode $userinfo{USERNAME},
$gecos and $hostname in svn-testament.pl since it was having some
trouble running.
Last lines that are output when I run make for the windows target are:
M.CONFIG: JPEG (libjpeg) enabled (NETSURF_USE_JPEG := YES)
M.CONFIG: JNG/MNG/PNG (libmng) disabled (NETSURF_USE_MNG := NO)
M.CONFIG: PDF export (haru) disabled (NETSURF_USE_HARU_PDF := NO)
M.CONFIG: glibc internal iconv disabled (NETSURF_USE_LIBICONV_PLUG := NO)
M.CONFIG: BMP (libnsbmp) enabled (NETSURF_USE_BMP := YES)
M.CONFIG: GIF (libnsgif) enabled (NETSURF_USE_GIF := YES)
M.CONFIG: PNG (libpng) enabled (NETSURF_USE_PNG := YES)
M.CONFIG: SVG (libsvgtiny) disabled (NETSURF_USE_NSSVG := NO)
M.CONFIG: PNG/JNG/MNG (libmng) disabled (NETSURF_USE_MNG := NO)
TESTMENT: unchanged
If I run make TARGET=windows again, those lines are all I get back.
When I tried running make TARGET=framebuffer, I got several errors
back. Some of them include:
image/mng.c:29:20: fatal error: libmng.h: No such file or directory
[ I thought I had MNG disabled, but appears framebuffer mode still
wants the library installed. ]
framebuffer/thumbnail.c:66:16: error: 'NSFB_SURFACE_RAM' undeclared
(first use in this function)
framebuffer/thumbnail.c:71:41: error: 'NSFB_FMT_XBGR8888' undeclared
(first use in this function)
framebuffer/thumbnail.c:73:2: error: incompatible type for argument 1
of 'nsfb_init'
framebuffer/thumbnail.c:90:2: error: too many arguments to function
'nsfb_plot_copy'
framebuffer/localhistory.c:22:23: fatal error: sys/ioctl.h: No such
file or directory
framebuffer/gui.c:22:23: fatal error: sys/ioctl.h: No such file or directory
framebuffer/framebuffer.c:196:3: error: too many arguments to function
'nsfb_plot_copy'
framebuffer/framebuffer.c:241:4: error: too many arguments to function
'nsfb_plot_copy'
framebuffer/framebuffer.c:355:22: error: storage size of 'fbtype' isn't known
framebuffer/framebuffer.c:356:24: error: storage size of 'fbfmt' isn't known
framebuffer/framebuffer.c:361:14: error: 'NSFB_FMT_XRGB8888'
undeclared (first use in this function)
...
framebuffer/convert_image.c:22:17: fatal error: png.h: No such file or directory
[ My png.h is in /usr/local/include. I'm guessing if this was
switched from <png.h> to "png.h", it'll find it. ]
cc.exe: error: build-windows-framebuffer/caret_image.c: No such file
or directory
cc.exe: fatal error: no input files
compilation terminated.
make: *** [build-windows-framebuffer/build-windows-framebuffer_caret_image.o]
Error 1
The errors look fixable. Don't know whether I should try to get the
framebuffer version to compile without errors or concentrate on why
the windows version doesn't generate anything other that object files
(and .d files). The framebuffer version may have the same issue as
the windows version once it gets past the compiler errors.
Any further suggestions appreciated. Thanks again for the help with
the target command.
> make TARGET=framebuffer
>
> the "make TARGET=windows" will attempt to build the native windows
> executable.
Thanks. TARGET=framebuffer is what I was looking for.
>
>
> The framebuffer library "libnsfb" requires X11 libraries...
> I once submitted a patch to the list that disables specific libnsfb
> backends.
> Just remove the X and maybe other not-so-important dependencies from
> the libnsfb
> makefiles and you should be able to compile the lib.
> After you have installed libnsfb, you should be able to link
> framebuffer-sdl for windows.
Made a few changes to the makefile and a few other files and it looks
like it builds the library fine on Windows.
From: John-Mark Bell
> Why not use the native Windows frontend?
>
> The framebuffer frontend is totally unsuitable for use on systems that
> have a full-featured window manager.
I was hoping to use the SDL version on a couple of platforms including
Windows. I suppose I could build each one natively though. If I can
get it to build, long-term plan was not to use it as a web browser,
but more of a cross-platform replacement for dialog. Was thinking I
could possibly replace the part that sends form data to a server with
the ability to store the information locally to a file. I wouldn't
even need the web access capabilities, just the capabilities to
interpret HTML and CSS and render pages. There are cross-platform
replacements for dialog, but I'd really like to have something that
can use standard web syntax (HTML, CSS, etc.) for creating interfaces.
>
> What issues? Please provide the exact build output.
>
I don't specifically know what the issue was when I tried to build
netsurf-2.8 versus netsurf-r13571, but I appear to have my include
paths set appropriately with the latter and I can't get them set
properly with the prior. I thought I went through the same steps in
both, but something is different. Since I appear to be having less
issues with the netsurf-r13571 tarball, I'm concentrating on trying to
build using it.
When I run make TARGET=windows I see no error messages. I redirected
standard output and standard error to a file. Looking over the file,
there are just warnings. Checking the build-windows-windows
directory, I see object files (108 of them). There are no libraries
or executables built. I did go in and hardcode $userinfo{USERNAME},
$gecos and $hostname in svn-testament.pl since it was having some
trouble running.
Last lines that are output when I run make for the windows target are:
M.CONFIG: JPEG (libjpeg) enabled (NETSURF_USE_JPEG := YES)
M.CONFIG: JNG/MNG/PNG (libmng) disabled (NETSURF_USE_MNG := NO)
M.CONFIG: PDF export (haru) disabled (NETSURF_USE_HARU_PDF := NO)
M.CONFIG: glibc internal iconv disabled (NETSURF_USE_LIBICONV_PLUG := NO)
M.CONFIG: BMP (libnsbmp) enabled (NETSURF_USE_BMP := YES)
M.CONFIG: GIF (libnsgif) enabled (NETSURF_USE_GIF := YES)
M.CONFIG: PNG (libpng) enabled (NETSURF_USE_PNG := YES)
M.CONFIG: SVG (libsvgtiny) disabled (NETSURF_USE_NSSVG := NO)
M.CONFIG: PNG/JNG/MNG (libmng) disabled (NETSURF_USE_MNG := NO)
TESTMENT: unchanged
If I run make TARGET=windows again, those lines are all I get back.
When I tried running make TARGET=framebuffer, I got several errors
back. Some of them include:
image/mng.c:29:20: fatal error: libmng.h: No such file or directory
[ I thought I had MNG disabled, but appears framebuffer mode still
wants the library installed. ]
framebuffer/thumbnail.c:66:16: error: 'NSFB_SURFACE_RAM' undeclared
(first use in this function)
framebuffer/thumbnail.c:71:41: error: 'NSFB_FMT_XBGR8888' undeclared
(first use in this function)
framebuffer/thumbnail.c:73:2: error: incompatible type for argument 1
of 'nsfb_init'
framebuffer/thumbnail.c:90:2: error: too many arguments to function
'nsfb_plot_copy'
framebuffer/localhistory.c:22:23: fatal error: sys/ioctl.h: No such
file or directory
framebuffer/gui.c:22:23: fatal error: sys/ioctl.h: No such file or directory
framebuffer/framebuffer.c:196:3: error: too many arguments to function
'nsfb_plot_copy'
framebuffer/framebuffer.c:241:4: error: too many arguments to function
'nsfb_plot_copy'
framebuffer/framebuffer.c:355:22: error: storage size of 'fbtype' isn't known
framebuffer/framebuffer.c:356:24: error: storage size of 'fbfmt' isn't known
framebuffer/framebuffer.c:361:14: error: 'NSFB_FMT_XRGB8888'
undeclared (first use in this function)
...
framebuffer/convert_image.c:22:17: fatal error: png.h: No such file or directory
[ My png.h is in /usr/local/include. I'm guessing if this was
switched from <png.h> to "png.h", it'll find it. ]
cc.exe: error: build-windows-framebuffer/caret_image.c: No such file
or directory
cc.exe: fatal error: no input files
compilation terminated.
make: *** [build-windows-framebuffer/build-windows-framebuffer_caret_image.o]
Error 1
The errors look fixable. Don't know whether I should try to get the
framebuffer version to compile without errors or concentrate on why
the windows version doesn't generate anything other that object files
(and .d files). The framebuffer version may have the same issue as
the windows version once it gets past the compiler errors.
Any further suggestions appreciated. Thanks again for the help with
the target command.
[gccsdk] [Bug 239] New: opendir() with trailing directory separator and RISC OS filename semantics.
http://www.riscos.info/bugzilla3/show_bug.cgi?id=239
Summary: opendir() with trailing directory separator and RISC
OS filename semantics.
Product: GCC/GCCSDK
Version: other
Platform: Other
OS/Version: RISC OS
Status: NEW
Severity: normal
Priority: P1
Component: C compiler
AssignedTo: John.Tytgat@aaug.net
ReportedBy: duncan_moore@ntlworld.com
Estimated Hours: 0.0
GCCSDK GCC 4.1.2 Release 2 Development 2012-05-05
VRPC RISC OS 4.39
#include <stdio.h> // printf
#include <stdlib.h> // system
#include <dirent.h> // opendir
#include <unixlib/local.h> // __riscosify_control
int __riscosify_control;
//---------------------------------------------------------------------------
int main(void) {
system("CDir dir_u");
system("CDir dir_r");
__riscosify_control&=~__RISCOSIFY_NO_PROCESS; // Unix names
if (!opendir("dir_u/")) printf("Error with Unix names\n");
__riscosify_control|=__RISCOSIFY_NO_PROCESS; // RISC OS names
if (!opendir("dir_r.")) printf("Error with RISC OS names\n");
system("Delete dir_r");
system("Delete dir_u");
return 0;
}
opendir() works with Unix filename semantics, but fails with RISC OS semantics
when there is a trailing directory separator:
*test
Error with RISC OS names
*
It shouldn't give any output at all.
--
Configure bugmail: http://www.riscos.info/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Summary: opendir() with trailing directory separator and RISC
OS filename semantics.
Product: GCC/GCCSDK
Version: other
Platform: Other
OS/Version: RISC OS
Status: NEW
Severity: normal
Priority: P1
Component: C compiler
AssignedTo: John.Tytgat@aaug.net
ReportedBy: duncan_moore@ntlworld.com
Estimated Hours: 0.0
GCCSDK GCC 4.1.2 Release 2 Development 2012-05-05
VRPC RISC OS 4.39
#include <stdio.h> // printf
#include <stdlib.h> // system
#include <dirent.h> // opendir
#include <unixlib/local.h> // __riscosify_control
int __riscosify_control;
//---------------------------------------------------------------------------
int main(void) {
system("CDir dir_u");
system("CDir dir_r");
__riscosify_control&=~__RISCOSIFY_NO_PROCESS; // Unix names
if (!opendir("dir_u/")) printf("Error with Unix names\n");
__riscosify_control|=__RISCOSIFY_NO_PROCESS; // RISC OS names
if (!opendir("dir_r.")) printf("Error with RISC OS names\n");
system("Delete dir_r");
system("Delete dir_u");
return 0;
}
opendir() works with Unix filename semantics, but fails with RISC OS semantics
when there is a trailing directory separator:
*test
Error with RISC OS names
*
It shouldn't give any output at all.
--
Configure bugmail: http://www.riscos.info/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Monday, 7 May 2012
Re: [gccsdk] fork/exec problems
In message <4ae3418c52.Jo@hobbes.bass-software.com>
John Tytgat <John.Tytgat@aaug.net> wrote:
> In message <8372408c52.Jo@hobbes.bass-software.com>
> John Tytgat <John.Tytgat@aaug.net> wrote:
>
> > You have to be aware of fork() vs vfork() differences. One of those
> > is that the child create with vfork() is sharing the same parent's
> > address space, while fork() makes a complete copy. That explains your
> > observation of wimpslot size changes with fork() and also why this
> > can only happen in Wimp context.
>
> Also the implementation in UnixLib is relatively simple, it moves the
> parent code at the end of application space, let the child run until it
> finishes and parent continues. So this does not end up with an extra
> Wimp task, nor it allows you to have concurrently running parent & child.
>
> John.
I found this again yesterday. quote
vfork() is an obsolete optimization. Before good memory management,
fork() made a full copy of the parent's memory, so it was pretty
expensive. since in many cases a fork() was followed by exec(),
which discards the current memory map and creates a new one,
it was a needless expense. Nowadays, fork() doesn't copy the memory;
it's simply set as "copy on write", so fork()+exec() is just as
efficient as vfork()+exec().
unquote.
I dont know if unixlib's version of fork() does this "copy on
write" or not, or any other details, just that fork()/exec has been
causing unstability in the past a (few months ago) and now with
latest svn upgrades stops program completely at the fork()/exec.
The earlier situation was erratic and I think would be difficult
to build a small test case to show it happening. It possibly is
only happening while running a largish program.
I am getting a better picture of things now, for example I know
now that it is fork()/exec causing unstability, and there are
obvious advantages to using vfork()/exec in maybe all cases.
Thanks, Ron M.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
John Tytgat <John.Tytgat@aaug.net> wrote:
> In message <8372408c52.Jo@hobbes.bass-software.com>
> John Tytgat <John.Tytgat@aaug.net> wrote:
>
> > You have to be aware of fork() vs vfork() differences. One of those
> > is that the child create with vfork() is sharing the same parent's
> > address space, while fork() makes a complete copy. That explains your
> > observation of wimpslot size changes with fork() and also why this
> > can only happen in Wimp context.
>
> Also the implementation in UnixLib is relatively simple, it moves the
> parent code at the end of application space, let the child run until it
> finishes and parent continues. So this does not end up with an extra
> Wimp task, nor it allows you to have concurrently running parent & child.
>
> John.
I found this again yesterday. quote
vfork() is an obsolete optimization. Before good memory management,
fork() made a full copy of the parent's memory, so it was pretty
expensive. since in many cases a fork() was followed by exec(),
which discards the current memory map and creates a new one,
it was a needless expense. Nowadays, fork() doesn't copy the memory;
it's simply set as "copy on write", so fork()+exec() is just as
efficient as vfork()+exec().
unquote.
I dont know if unixlib's version of fork() does this "copy on
write" or not, or any other details, just that fork()/exec has been
causing unstability in the past a (few months ago) and now with
latest svn upgrades stops program completely at the fork()/exec.
The earlier situation was erratic and I think would be difficult
to build a small test case to show it happening. It possibly is
only happening while running a largish program.
I am getting a better picture of things now, for example I know
now that it is fork()/exec causing unstability, and there are
obvious advantages to using vfork()/exec in maybe all cases.
Thanks, Ron M.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Re: netsurf-sdl on Windows?
On Mon, 2012-05-07 at 15:32 -0400, LM wrote:
> Tried out netsurf-sdl on Debian and was attempting to figure out if
> something similar could be built on Windows.
Why not use the native Windows frontend?
The framebuffer frontend is totally unsuitable for use on systems that
have a full-featured window manager.
> I attempted to build netsurf-2.8-full-src.tar.gz, but hit some issues.
What issues? Please provide the exact build output.
> I invoked the make file with:
> make target=windows
This builds the native Windows frontend, not the framebuffer.
If you really want to build the framebuffer frontend, then you want
"make TARGET=framebuffer".
> It seems to build several object files in the build-windows-windows
> directory, but it never creates an executable.
Again, please provide the exact build output. Without it, there's
nothing we can do to help.
John-Mark.
> Tried out netsurf-sdl on Debian and was attempting to figure out if
> something similar could be built on Windows.
Why not use the native Windows frontend?
The framebuffer frontend is totally unsuitable for use on systems that
have a full-featured window manager.
> I attempted to build netsurf-2.8-full-src.tar.gz, but hit some issues.
What issues? Please provide the exact build output.
> I invoked the make file with:
> make target=windows
This builds the native Windows frontend, not the framebuffer.
If you really want to build the framebuffer frontend, then you want
"make TARGET=framebuffer".
> It seems to build several object files in the build-windows-windows
> directory, but it never creates an executable.
Again, please provide the exact build output. Without it, there's
nothing we can do to help.
John-Mark.
Re: netsurf-sdl on Windows?
Am Montag, den 07.05.2012, 15:32 +0200 schrieb LM <lmemsm@gmail.com>:
Hello,
> Tried out netsurf-sdl on Debian and was attempting to figure out if
> something similar could be built on Windows.
>
> I invoked the make file with:
> make target=windows
That's wrong. The framebuffer isn't OS specific, you still have to
build it this way:
make TARGET=framebuffer
the "make TARGET=windows" will attempt to build the native windows
executable.
Would be good if you could get both builds to work under windows. :)
The framebuffer library "libnsfb" requires X11 libraries...
I once submitted a patch to the list that disables specific libnsfb
backends.
Just remove the X and maybe other not-so-important dependencies from
the libnsfb
makefiles and you should be able to compile the lib.
After you have installed libnsfb, you should be able to link
framebuffer-sdl for windows.
Greets,
Ole
Hello,
> Tried out netsurf-sdl on Debian and was attempting to figure out if
> something similar could be built on Windows.
>
> I invoked the make file with:
> make target=windows
That's wrong. The framebuffer isn't OS specific, you still have to
build it this way:
make TARGET=framebuffer
the "make TARGET=windows" will attempt to build the native windows
executable.
Would be good if you could get both builds to work under windows. :)
The framebuffer library "libnsfb" requires X11 libraries...
I once submitted a patch to the list that disables specific libnsfb
backends.
Just remove the X and maybe other not-so-important dependencies from
the libnsfb
makefiles and you should be able to compile the lib.
After you have installed libnsfb, you should be able to link
framebuffer-sdl for windows.
Greets,
Ole
netsurf-sdl on Windows?
Tried out netsurf-sdl on Debian and was attempting to figure out if something similar could be built on Windows.
I have sdl, JPEG, Zlib, libxml2, openssl and curl built and installed. I'm using mingw and msys to build applications. I downloaded and built hubbub-0.1.2, libcss-0.1.2, libnsbmp-0.0.3, libnsgif-0.0.3, libparserutils-0.1.1, libwapcaplet-0.1.1. They seemed to compile and install without any complaints. I attempted to build netsurf-2.8-full-src.tar.gz, but hit some issues. So, I downloaded netsurf-r13571.tar.gz to try it. I edited makefile.config. I tried setting:
export MINGW_PREFIX='/mingw/bin/'
export MINGW_INSTALL_ENV='/usr/local'
export MINGW_INSTALL_ENV='/usr/local'
I invoked the make file with:
make target=windows
make target=windows
It seems to build several object files in the build-windows-windows directory, but it never creates an executable. It doesn't even appear to process the resource file. I'm also not sure how to tie SDL into the windows build. Any suggestions for getting further on this would be greatly appreciated. Thank you.
[gccsdk] [Bug 235] GCC 4.1.2 Rel 2 Dev 2012-05-01: Problem reading Archive image files.
http://www.riscos.info/bugzilla3/show_bug.cgi?id=235
--- Comment #5 from Duncan Moore <duncan_moore@ntlworld.com> 2012-05-07 12:15:18 PDT ---
With ds212.zip and the executable in the RAM disc I get the same result:
<output removed by hand>
textj.b9b -1
j.b9b -1
_ -1
--
Configure bugmail: http://www.riscos.info/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
--- Comment #5 from Duncan Moore <duncan_moore@ntlworld.com> 2012-05-07 12:15:18 PDT ---
With ds212.zip and the executable in the RAM disc I get the same result:
<output removed by hand>
textj.b9b -1
j.b9b -1
_ -1
--
Configure bugmail: http://www.riscos.info/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Re: [gccsdk] fork/exec problems
In message <8372408c52.Jo@hobbes.bass-software.com>
John Tytgat <John.Tytgat@aaug.net> wrote:
> You have to be aware of fork() vs vfork() differences. One of those
> is that the child create with vfork() is sharing the same parent's
> address space, while fork() makes a complete copy. That explains your
> observation of wimpslot size changes with fork() and also why this
> can only happen in Wimp context.
Also the implementation in UnixLib is relatively simple, it moves the
parent code at the end of application space, let the child run until it
finishes and parent continues. So this does not end up with an extra
Wimp task, nor it allows you to have concurrently running parent & child.
John.
--
John Tytgat, in his comfy chair at home BASS
John.Tytgat@aaug.net ARM powered, RISC OS driven
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
John Tytgat <John.Tytgat@aaug.net> wrote:
> You have to be aware of fork() vs vfork() differences. One of those
> is that the child create with vfork() is sharing the same parent's
> address space, while fork() makes a complete copy. That explains your
> observation of wimpslot size changes with fork() and also why this
> can only happen in Wimp context.
Also the implementation in UnixLib is relatively simple, it moves the
parent code at the end of application space, let the child run until it
finishes and parent continues. So this does not end up with an extra
Wimp task, nor it allows you to have concurrently running parent & child.
John.
--
John Tytgat, in his comfy chair at home BASS
John.Tytgat@aaug.net ARM powered, RISC OS driven
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Re: [gccsdk] fork/exec problems
In message <f8700c8c52.beeb@ron1954.woosh.co.nz>
Ron <beeb@woosh.co.nz> wrote:
> In message <77ebe08b52.beeb@ron1954.woosh.co.nz>
> Ron <beeb@woosh.co.nz> wrote:
>
> > I appreciate Duncan's bug report on fork/exec as I am having problems.
> > The latest release 4.1.2 (upgraded last night) reliably halts with
> > segmentation error.
> > Actually previous releases have been giving more erratic faults only
> > sometimes crashing at the some spot further into the program after
> > returning from the fork.
> >
> > In the case of Tar, simply replacing fork with vfork or setting
> > Tar$heap "" doesn't work, though I think in the long term, I'll have
> > to change the code so it will work with vfork.
> > The fork/exec is only doing a bit of file renaming for multipart tars,
> > so shouldn't need a full copy of the parent.
> >
>
> I have vfork() working with the execv command now and straight away
> it looks like any problem is gone.
>
> I was changing fork() to vfork() in a place that affected all fork()'s
> in the program and that did not work.
> It looks like the fork()/execv combination was the main problem.
>
> I notice now I'm using vfork()/execv that the application task size
> does not grow, and it stays at the allocated wimpslot size.
> With fork()/execv, about 740K was added to the displayed task size
> each time that routine was called. The older 4.1.2 would allow this
> to happen sometimes but typically would cause a crash somewhere
> after a few fork()/execv's.
What do you mean with "older 4.1.2" and "lastest release 4.1.2" ?
"http://www.riscos.info/downloads/gccsdk/gcc-4.1.2-release-1/" vs
"http://www.riscos.info/downloads/gccsdk/testing/4.1.2/" ?
You have to be aware of fork() vs vfork() differences. One of those
is that the child create with vfork() is sharing the same parent's
address space, while fork() makes a complete copy. That explains your
observation of wimpslot size changes with fork() and also why this
can only happen in Wimp context.
If you think you found a bug or any unwanted behavior, feel free to
create a bugzilly bug record but do include an as small as possible
test case. The ones Duncan makes are really superb ones.
John.
--
John Tytgat, in his comfy chair at home BASS
John.Tytgat@aaug.net ARM powered, RISC OS driven
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Ron <beeb@woosh.co.nz> wrote:
> In message <77ebe08b52.beeb@ron1954.woosh.co.nz>
> Ron <beeb@woosh.co.nz> wrote:
>
> > I appreciate Duncan's bug report on fork/exec as I am having problems.
> > The latest release 4.1.2 (upgraded last night) reliably halts with
> > segmentation error.
> > Actually previous releases have been giving more erratic faults only
> > sometimes crashing at the some spot further into the program after
> > returning from the fork.
> >
> > In the case of Tar, simply replacing fork with vfork or setting
> > Tar$heap "" doesn't work, though I think in the long term, I'll have
> > to change the code so it will work with vfork.
> > The fork/exec is only doing a bit of file renaming for multipart tars,
> > so shouldn't need a full copy of the parent.
> >
>
> I have vfork() working with the execv command now and straight away
> it looks like any problem is gone.
>
> I was changing fork() to vfork() in a place that affected all fork()'s
> in the program and that did not work.
> It looks like the fork()/execv combination was the main problem.
>
> I notice now I'm using vfork()/execv that the application task size
> does not grow, and it stays at the allocated wimpslot size.
> With fork()/execv, about 740K was added to the displayed task size
> each time that routine was called. The older 4.1.2 would allow this
> to happen sometimes but typically would cause a crash somewhere
> after a few fork()/execv's.
What do you mean with "older 4.1.2" and "lastest release 4.1.2" ?
"http://www.riscos.info/downloads/gccsdk/gcc-4.1.2-release-1/" vs
"http://www.riscos.info/downloads/gccsdk/testing/4.1.2/" ?
You have to be aware of fork() vs vfork() differences. One of those
is that the child create with vfork() is sharing the same parent's
address space, while fork() makes a complete copy. That explains your
observation of wimpslot size changes with fork() and also why this
can only happen in Wimp context.
If you think you found a bug or any unwanted behavior, feel free to
create a bugzilly bug record but do include an as small as possible
test case. The ones Duncan makes are really superb ones.
John.
--
John Tytgat, in his comfy chair at home BASS
John.Tytgat@aaug.net ARM powered, RISC OS driven
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
[gccsdk] [Bug 235] GCC 4.1.2 Rel 2 Dev 2012-05-01: Problem reading Archive image files.
http://www.riscos.info/bugzilla3/show_bug.cgi?id=235
--- Comment #4 from John Tytgat <John.Tytgat@aaug.net> 2012-05-07 11:10:54 PDT ---
I've tested this on RPC but I doubt this could be VRPC specific. Can you test
this again with ds212.zip and the executable in RAM disc ?
And no, I wasn't redirecting the output to a file.
--
Configure bugmail: http://www.riscos.info/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
--- Comment #4 from John Tytgat <John.Tytgat@aaug.net> 2012-05-07 11:10:54 PDT ---
I've tested this on RPC but I doubt this could be VRPC specific. Can you test
this again with ds212.zip and the executable in RAM disc ?
And no, I wasn't redirecting the output to a file.
--
Configure bugmail: http://www.riscos.info/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
[gccsdk] [Bug 238] New: Temporary filenames with RISC OS filename semantics
http://www.riscos.info/bugzilla3/show_bug.cgi?id=238
Summary: Temporary filenames with RISC OS filename semantics
Product: GCC/GCCSDK
Version: other
Platform: Other
OS/Version: RISC OS
Status: NEW
Severity: normal
Priority: P1
Component: C compiler
AssignedTo: John.Tytgat@aaug.net
ReportedBy: duncan_moore@ntlworld.com
Estimated Hours: 0.0
GCCSDK GCC 4.1.2 Release 2 Development 2012-05-05
VRPC RISC OS 4.39
#include <stdio.h> // printf
#include <stdlib.h> // getenv
#include <unistd.h> // sleep
#include <unixlib/local.h> // __riscosify_control
int __riscosify_control;
//---------------------------------------------------------------------------
int main(void) {
__riscosify_control|=__RISCOSIFY_NO_PROCESS; // RISC OS names
printf("%s\n",tmpnam(NULL));
printf("%s\n",tempnam(getenv("Wimp$ScrapDir"),"pip"));
FILE* tt=tmpfile();
fprintf(tt,"eee\n");
sleep(5);
return 0;
}
This program is using RISC OS filename semantics, but the temporary file
functions all ignore this, giving:
*test
/tmp/__8lbs5c
HostFS::HardDisc4.$.!boot.Resources.!Scrap.ScrapDirs.IDdisabled/pip8lbs5c
*
I would expect something more like:
*test
<Wimp$ScrapDir>.__8lbs5c
HostFS::HardDisc4.$.!boot.Resources.!Scrap.ScrapDirs.IDdisabled.pip8lbs5c
*
Note that tempnam() used '/' as the directory separator, and not '.'.
Also, tmpfile() creates a temporary file in CSD with a name like:
/tmp/__8lbs5c
--
Configure bugmail: http://www.riscos.info/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Summary: Temporary filenames with RISC OS filename semantics
Product: GCC/GCCSDK
Version: other
Platform: Other
OS/Version: RISC OS
Status: NEW
Severity: normal
Priority: P1
Component: C compiler
AssignedTo: John.Tytgat@aaug.net
ReportedBy: duncan_moore@ntlworld.com
Estimated Hours: 0.0
GCCSDK GCC 4.1.2 Release 2 Development 2012-05-05
VRPC RISC OS 4.39
#include <stdio.h> // printf
#include <stdlib.h> // getenv
#include <unistd.h> // sleep
#include <unixlib/local.h> // __riscosify_control
int __riscosify_control;
//---------------------------------------------------------------------------
int main(void) {
__riscosify_control|=__RISCOSIFY_NO_PROCESS; // RISC OS names
printf("%s\n",tmpnam(NULL));
printf("%s\n",tempnam(getenv("Wimp$ScrapDir"),"pip"));
FILE* tt=tmpfile();
fprintf(tt,"eee\n");
sleep(5);
return 0;
}
This program is using RISC OS filename semantics, but the temporary file
functions all ignore this, giving:
*test
/tmp/__8lbs5c
HostFS::HardDisc4.$.!boot.Resources.!Scrap.ScrapDirs.IDdisabled/pip8lbs5c
*
I would expect something more like:
*test
<Wimp$ScrapDir>.__8lbs5c
HostFS::HardDisc4.$.!boot.Resources.!Scrap.ScrapDirs.IDdisabled.pip8lbs5c
*
Note that tempnam() used '/' as the directory separator, and not '.'.
Also, tmpfile() creates a temporary file in CSD with a name like:
/tmp/__8lbs5c
--
Configure bugmail: http://www.riscos.info/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
[gccsdk] [Bug 237] New: Ampersands in command line arguments.
http://www.riscos.info/bugzilla3/show_bug.cgi?id=237
Summary: Ampersands in command line arguments.
Product: GCC/GCCSDK
Version: other
Platform: Other
OS/Version: RISC OS
Status: NEW
Severity: normal
Priority: P1
Component: C compiler
AssignedTo: John.Tytgat@aaug.net
ReportedBy: duncan_moore@ntlworld.com
Estimated Hours: 0.0
GCCSDK GCC 4.1.2 Release 2 Development 2012-05-05
VRPC RISC OS 4.39
#include <stdio.h>
int main(int argc,char** argv) {
while (*++argv) printf("%s\n",*argv);
return 0;
}
I've had trouble passing command line arguments with ampersands in them.
For example, I would expect:
*test &0 &0,&1 &9,&1 &-,&1 "&0,&1" &a,&b
&0
&0,&1
&9,&1
&-,&1
&0,&1
&a,&b
*
but get:
*test &0 &0,&1 &9,&1 &-,&1 "&0,&1" &a,&b
&0
&0
&9
&-
&a,&b
*
(The argument in double quotes is missing completely.) The arguments seem to be
confused with redirection file descriptors, 2>&1 2>&- etc.
With just small changes to an argument, it can be quite unpredictable as to
what's going to be passed through. The only way I've found to get what I
originally wanted was to protect the first ampersand with both "" and \
*test "\&0,&1"
&0,&1
*
--
Configure bugmail: http://www.riscos.info/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Summary: Ampersands in command line arguments.
Product: GCC/GCCSDK
Version: other
Platform: Other
OS/Version: RISC OS
Status: NEW
Severity: normal
Priority: P1
Component: C compiler
AssignedTo: John.Tytgat@aaug.net
ReportedBy: duncan_moore@ntlworld.com
Estimated Hours: 0.0
GCCSDK GCC 4.1.2 Release 2 Development 2012-05-05
VRPC RISC OS 4.39
#include <stdio.h>
int main(int argc,char** argv) {
while (*++argv) printf("%s\n",*argv);
return 0;
}
I've had trouble passing command line arguments with ampersands in them.
For example, I would expect:
*test &0 &0,&1 &9,&1 &-,&1 "&0,&1" &a,&b
&0
&0,&1
&9,&1
&-,&1
&0,&1
&a,&b
*
but get:
*test &0 &0,&1 &9,&1 &-,&1 "&0,&1" &a,&b
&0
&0
&9
&-
&a,&b
*
(The argument in double quotes is missing completely.) The arguments seem to be
confused with redirection file descriptors, 2>&1 2>&- etc.
With just small changes to an argument, it can be quite unpredictable as to
what's going to be passed through. The only way I've found to get what I
originally wanted was to protect the first ampersand with both "" and \
*test "\&0,&1"
&0,&1
*
--
Configure bugmail: http://www.riscos.info/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
[gccsdk] [Bug 235] GCC 4.1.2 Rel 2 Dev 2012-05-01: Problem reading Archive image files.
http://www.riscos.info/bugzilla3/show_bug.cgi?id=235
--- Comment #3 from Duncan Moore <duncan_moore@ntlworld.com> 2012-05-07 02:51:06 PDT ---
GCCSDK GCC 4.1.2 Release 2 Development 2012-05-05
Having looked at this again, I've noticed that if the output of the test case
is redirected to a file (i.e. test > z) rather than a TaskWindow, then the
output is different but still wrong, but at a quick glance looks correct.
This shows the difference from what I would expect, the difference being on the
second to last line:
Expected What I get
<output removed by hand>
TextfileS2 -1 TextfileS2 -1
TextfileU -1 TextfileU -1
textg.b9b -1 textg.b9b -1
texth.b9b -1 texth.b9b -1
texti.b9b -1 texti.b9b -1
textj.b9b -1 textj.b9b -1
textk.b9b -1 | textk.b9Q -1
trans.b9b -1 trans.b9b -1
Does this explain why it (seems to) works on your system John, and not mine?
--
Configure bugmail: http://www.riscos.info/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
--- Comment #3 from Duncan Moore <duncan_moore@ntlworld.com> 2012-05-07 02:51:06 PDT ---
GCCSDK GCC 4.1.2 Release 2 Development 2012-05-05
Having looked at this again, I've noticed that if the output of the test case
is redirected to a file (i.e. test > z) rather than a TaskWindow, then the
output is different but still wrong, but at a quick glance looks correct.
This shows the difference from what I would expect, the difference being on the
second to last line:
Expected What I get
<output removed by hand>
TextfileS2 -1 TextfileS2 -1
TextfileU -1 TextfileU -1
textg.b9b -1 textg.b9b -1
texth.b9b -1 texth.b9b -1
texti.b9b -1 texti.b9b -1
textj.b9b -1 textj.b9b -1
textk.b9b -1 | textk.b9Q -1
trans.b9b -1 trans.b9b -1
Does this explain why it (seems to) works on your system John, and not mine?
--
Configure bugmail: http://www.riscos.info/bugzilla3/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Re: [gccsdk] fork/exec problems
In message <77ebe08b52.beeb@ron1954.woosh.co.nz>
Ron <beeb@woosh.co.nz> wrote:
> I appreciate Duncan's bug report on fork/exec as I am having problems.
> The latest release 4.1.2 (upgraded last night) reliably halts with
> segmentation error.
> Actually previous releases have been giving more erratic faults only
> sometimes crashing at the some spot further into the program after
> returning from the fork.
>
> In the case of Tar, simply replacing fork with vfork or setting
> Tar$heap "" doesn't work, though I think in the long term, I'll have
> to change the code so it will work with vfork.
> The fork/exec is only doing a bit of file renaming for multipart tars,
> so shouldn't need a full copy of the parent.
>
I have vfork() working with the execv command now and straight away
it looks like any problem is gone.
I was changing fork() to vfork() in a place that affected all fork()'s
in the program and that did not work.
It looks like the fork()/execv combination was the main problem.
I notice now I'm using vfork()/execv that the application task size
does not grow, and it stays at the allocated wimpslot size.
With fork()/execv, about 740K was added to the displayed task size
each time that routine was called. The older 4.1.2 would allow this
to happen sometimes but typically would cause a crash somewhere
after a few fork()/execv's.
Thanks, Ron M.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Ron <beeb@woosh.co.nz> wrote:
> I appreciate Duncan's bug report on fork/exec as I am having problems.
> The latest release 4.1.2 (upgraded last night) reliably halts with
> segmentation error.
> Actually previous releases have been giving more erratic faults only
> sometimes crashing at the some spot further into the program after
> returning from the fork.
>
> In the case of Tar, simply replacing fork with vfork or setting
> Tar$heap "" doesn't work, though I think in the long term, I'll have
> to change the code so it will work with vfork.
> The fork/exec is only doing a bit of file renaming for multipart tars,
> so shouldn't need a full copy of the parent.
>
I have vfork() working with the execv command now and straight away
it looks like any problem is gone.
I was changing fork() to vfork() in a place that affected all fork()'s
in the program and that did not work.
It looks like the fork()/execv combination was the main problem.
I notice now I'm using vfork()/execv that the application task size
does not grow, and it stays at the allocated wimpslot size.
With fork()/execv, about 740K was added to the displayed task size
each time that routine was called. The older 4.1.2 would allow this
to happen sometimes but typically would cause a crash somewhere
after a few fork()/execv's.
Thanks, Ron M.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Re: [gccsdk] Autobuilder and Scummvm
On 7 May 2012 07:25, Chris Gransden <chrisg@care4free.net> wrote:
Thanks again Chris, libsdl has built now!
Rob.
In article
<CAHQJq4LUEKPtuPZJEqGkJigMWbk-AZ5MZZkjXogkTKsADadSow@mail.gmail.com>,
Most of them are due to upstream changes so I shouldn't need acess to your
> Thanks Chris, that workaround got libtiff4 to build. If you want ssh access
> to my build machine, let me know.
machine.
This should be ok now. r5801
> It looks like libsdl1.2 needs a patch now! Here is the error;
> Autobuilder: RISC OS patch:
> /home/robheaton/Downloads/gccsdk/autobuilder/libraries/sdl/libsdl1.2debian/include.os.p
> patching file src/video/riscos/SDL_wimppoll.c
> patching file src/video/riscos/SDL_riscosFullScreenVideo.c
> Autobuilder: RISC OS patch:
> /home/robheaton/Downloads/gccsdk/autobuilder/libraries/sdl/libsdl1.2debian/SDL_riscosevents.p
> patching file src/video/riscos/SDL_riscosevents.c
> Autobuilder: RISC OS patch:
> /home/robheaton/Downloads/gccsdk/autobuilder/libraries/sdl/libsdl1.2debian/video.p.elf
> patching file src/video/riscos/SDL_riscosASM.S
> Hunk #1 FAILED at 1.
> 1 out of 2 hunks FAILED -- saving rejects to file
> src/video/riscos/SDL_riscosASM.S.rej
> Autobuilder: Patches failed
> Package libsdl1.2debian: ***Failure***
> Dependency "libsdl1.2debian" failed for scummvm
> Build for package "scummvm" failed
Thanks again Chris, libsdl has built now!
Rob.
Sunday, 6 May 2012
Re: [gccsdk] Autobuilder and Scummvm
In article
<CAHQJq4LUEKPtuPZJEqGkJigMWbk-AZ5MZZkjXogkTKsADadSow@mail.gmail.com>,
> Thanks Chris, that workaround got libtiff4 to build. If you want ssh access
> to my build machine, let me know.
Most of them are due to upstream changes so I shouldn't need acess to your
machine.
> It looks like libsdl1.2 needs a patch now! Here is the error;
> Autobuilder: RISC OS patch:
> /home/robheaton/Downloads/gccsdk/autobuilder/libraries/sdl/libsdl1.2debian/include.os.p
> patching file src/video/riscos/SDL_wimppoll.c
> patching file src/video/riscos/SDL_riscosFullScreenVideo.c
> Autobuilder: RISC OS patch:
> /home/robheaton/Downloads/gccsdk/autobuilder/libraries/sdl/libsdl1.2debian/SDL_riscosevents.p
> patching file src/video/riscos/SDL_riscosevents.c
> Autobuilder: RISC OS patch:
> /home/robheaton/Downloads/gccsdk/autobuilder/libraries/sdl/libsdl1.2debian/video.p.elf
> patching file src/video/riscos/SDL_riscosASM.S
> Hunk #1 FAILED at 1.
> 1 out of 2 hunks FAILED -- saving rejects to file
> src/video/riscos/SDL_riscosASM.S.rej
> Autobuilder: Patches failed
> Package libsdl1.2debian: ***Failure***
> Dependency "libsdl1.2debian" failed for scummvm
> Build for package "scummvm" failed
This should be ok now. r5801
Chris.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
<CAHQJq4LUEKPtuPZJEqGkJigMWbk-AZ5MZZkjXogkTKsADadSow@mail.gmail.com>,
> Thanks Chris, that workaround got libtiff4 to build. If you want ssh access
> to my build machine, let me know.
Most of them are due to upstream changes so I shouldn't need acess to your
machine.
> It looks like libsdl1.2 needs a patch now! Here is the error;
> Autobuilder: RISC OS patch:
> /home/robheaton/Downloads/gccsdk/autobuilder/libraries/sdl/libsdl1.2debian/include.os.p
> patching file src/video/riscos/SDL_wimppoll.c
> patching file src/video/riscos/SDL_riscosFullScreenVideo.c
> Autobuilder: RISC OS patch:
> /home/robheaton/Downloads/gccsdk/autobuilder/libraries/sdl/libsdl1.2debian/SDL_riscosevents.p
> patching file src/video/riscos/SDL_riscosevents.c
> Autobuilder: RISC OS patch:
> /home/robheaton/Downloads/gccsdk/autobuilder/libraries/sdl/libsdl1.2debian/video.p.elf
> patching file src/video/riscos/SDL_riscosASM.S
> Hunk #1 FAILED at 1.
> 1 out of 2 hunks FAILED -- saving rejects to file
> src/video/riscos/SDL_riscosASM.S.rej
> Autobuilder: Patches failed
> Package libsdl1.2debian: ***Failure***
> Dependency "libsdl1.2debian" failed for scummvm
> Build for package "scummvm" failed
This should be ok now. r5801
Chris.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
[gccsdk] fork/exec problems
I appreciate Duncan's bug report on fork/exec as I am having problems.
The latest release 4.1.2 (upgraded last night) reliably halts with
segmentation error.
Actually previous releases have been giving more erratic faults only
sometimes crashing at the some spot further into the program after
returning from the fork.
In the case of Tar, simply replacing fork with vfork or setting
Tar$heap "" doesn't work, though I think in the long term, I'll have
to change the code so it will work with vfork.
The fork/exec is only doing a bit of file renaming for multipart tars,
so shouldn't need a full copy of the parent.
I read somewhere (in the unix world) that there is a swing back to
using fork instead of vfork where it would be used, because new
versions of fork can do it differently now.
I have noticed previously that running the program from the Supervisor
that the fork wont work unless the WindowManager is running, does this
mean the WindowManager is necessary to get a new pid/(process)?
Thanks for any enlightenment on this subject.
Ron M.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
The latest release 4.1.2 (upgraded last night) reliably halts with
segmentation error.
Actually previous releases have been giving more erratic faults only
sometimes crashing at the some spot further into the program after
returning from the fork.
In the case of Tar, simply replacing fork with vfork or setting
Tar$heap "" doesn't work, though I think in the long term, I'll have
to change the code so it will work with vfork.
The fork/exec is only doing a bit of file renaming for multipart tars,
so shouldn't need a full copy of the parent.
I read somewhere (in the unix world) that there is a swing back to
using fork instead of vfork where it would be used, because new
versions of fork can do it differently now.
I have noticed previously that running the program from the Supervisor
that the fork wont work unless the WindowManager is running, does this
mean the WindowManager is necessary to get a new pid/(process)?
Thanks for any enlightenment on this subject.
Ron M.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
Re: [Rpcemu] HostFS location
On 4 May 2012, Tony Moore <old_coaster@yahoo.co.uk> wrote:
> On 4 May 2012, Andrew Fawcett <andy@athame.co.uk> wrote:
> > On Thursday 03 May 2012 22:18:30 Tony Moore wrote:
[snip]
> > > If the HostFS folder could be located outside the RPCEmu folder,
> > > several copies of RPCEmu could access the same HostFS folder.
[snip]
> > Not sure about the Windows build, but for Linux it's easy. Just
> > create a softlink called "hostfs" to the "shared" master copy.
>
> Using the Windows build 0.8.9, I have tried creating a shortcut, in
> the RPCEmu folder, to the HostFS folder, elsewhere. On running RPCEmu,
> there is no error, but clicking on the HostFS icon opens an empty
> filer window so it seems that the shorcut is ignored.
Has anyone, using the Windows build, managed to re-locate HostFS?
Tony
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
> On 4 May 2012, Andrew Fawcett <andy@athame.co.uk> wrote:
> > On Thursday 03 May 2012 22:18:30 Tony Moore wrote:
[snip]
> > > If the HostFS folder could be located outside the RPCEmu folder,
> > > several copies of RPCEmu could access the same HostFS folder.
[snip]
> > Not sure about the Windows build, but for Linux it's easy. Just
> > create a softlink called "hostfs" to the "shared" master copy.
>
> Using the Windows build 0.8.9, I have tried creating a shortcut, in
> the RPCEmu folder, to the HostFS folder, elsewhere. On running RPCEmu,
> there is no error, but clicking on the HostFS icon opens an empty
> filer window so it seems that the shorcut is ignored.
Has anyone, using the Windows build, managed to re-locate HostFS?
Tony
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
Subscribe to:
Posts (Atom)