Monday, 31 August 2015

[gccsdk] Building libssh2 fails while building nettle with the autobuilder. Any clues?

I would like to be able to build !Nettle as I would like to see if I can add public/private key authentication to it.

I have never used the auto builder before and I am at a bit of a loss to know where to start with the error that it presented while building libssh2.

 

Along the way:

 

Autobuilder: source directory is /home/alan/build/libssh2/libssh2-1.5.0

Autobuilder: Debian patch: 0001-Add-lgpg-error-to-.pc-to-facilitate-static-linking.patch

  ...trying level 0

  ...trying level 1

patching file libssh2.pc.in

  ...success !

Autobuilder: Debian patch: 0001-Do-not-expose-private-libraries-nor-link-flags-to-us.patch

  ...trying level 0

  ...trying level 1

patching file libssh2.pc.in

  ...success !

Autobuilder: Calling setvars

Autobuilder: Calling configure command:  /home/alan/env/ro-config

RISC OS: Adding E1F filetype for ./configure

RISC OS: Adding filetype for binary tests for ./configure

RISC OS: Setting libtool library paths for ./configure

RISC OS: Adding X Paths

RISC OS: Adding SSL Path

RISC OS: Enabling shared libraries

RISC OS: Enabling static libraries

checking whether to enable maintainer-specific portions of Makefiles... no

checking for sed... /bin/sed

checking for a BSD-compatible install... /home/alan/env/ro-install

checking whether build environment is sane... yes

/home/alan/build/libssh2/libssh2-1.5.0/missing: Unknown `--is-lightweight' option

Try `/home/alan/build/libssh2/libssh2-1.5.0/missing --help' for more information

configure: WARNING: 'missing' script is too old or missing

checking for arm-unknown-riscos-strip... /home/alan/cross/bin/arm-unknown-riscos-strip

checking for a thread-safe mkdir -p... /home/alan/env/mkdir -p

checking for gawk... gawk

 

 

 

Here is how it ends

 

Making all in tests

make[1]: Entering directory `/home/alan/build/libssh2/libssh2-1.5.0/tests'

/home/alan/env/arm-unknown-riscos-gcc -DHAVE_CONFIG_H -I. -I../src -I../example  -I../src -I../include -I../src -I/home/alan/env/include  -O3 -MT ssh2.o -MD -MP -MF .deps/ssh2.Tpo -c -o ssh2.o ssh2.c

mv -f .deps/ssh2.Tpo .deps/ssh2.Po

/bin/bash ../libtool  --tag=CC   --mode=link /home/alan/env/arm-unknown-riscos-gcc  -O3   -o ssh2,e1f ssh2.o ../src/libssh2.la

libtool: link: /home/alan/env/arm-unknown-riscos-gcc -O3 -o ssh2,e1f ssh2.o  ../src/.libs/libssh2.a -L/home/alan/env/lib -lssl -lcrypto

/home/alan/env/lib/libcrypto.a(dso_dlfcn.o): In function `dlfcn_globallookup':

dso_dlfcn.c:(.text+0x7c): undefined reference to `dlopen'

dso_dlfcn.c:(.text+0x90): undefined reference to `dlsym'

dso_dlfcn.c:(.text+0x9c): undefined reference to `dlclose'

/home/alan/env/lib/libcrypto.a(dso_dlfcn.o): In function `dlfcn_bind_func':

dso_dlfcn.c:(.text+0x454): undefined reference to `dlsym'

dso_dlfcn.c:(.text+0x4c0): undefined reference to `dlerror'

/home/alan/env/lib/libcrypto.a(dso_dlfcn.o): In function `dlfcn_bind_var':

dso_dlfcn.c:(.text+0x5a0): undefined reference to `dlsym'

dso_dlfcn.c:(.text+0x60c): undefined reference to `dlerror'

/home/alan/env/lib/libcrypto.a(dso_dlfcn.o): In function `dlfcn_unload':

dso_dlfcn.c:(.text+0x6ac): undefined reference to `dlclose'

/home/alan/env/lib/libcrypto.a(dso_dlfcn.o): In function `dlfcn_load':

dso_dlfcn.c:(.text+0x750): undefined reference to `dlopen'

dso_dlfcn.c:(.text+0x7b4): undefined reference to `dlclose'

dso_dlfcn.c:(.text+0x7cc): undefined reference to `dlerror'

collect2: ld returned 1 exit status

make[1]: *** [ssh2,e1f] Error 1

make[1]: Leaving directory `/home/alan/build/libssh2/libssh2-1.5.0/tests'

make: *** [all-recursive] Error 1

Package libssh2: ***Failure***

Dependency "libssh2" failed for nettle

Build for package "nettle" failed

 

That looks to me like a disparity in expectations between static and dynamic builds for ‘tests’ but I couldn’t see how or where to influence that.  I presume that even if it can build them it can’t run them anyway.

 

Thanks in advance to anybody to suggest how to deal with this.

Alan

 

 

Re: toolchains: branch master updated. a47ec59181b5d5f7588f75517cb5c2b95ff22e08

On Mon, 31 Aug 2015 23:16:49 +0100, Commit Mailer wibbled on for an age:

> commitdiff http://git.netsurf-browser.org/toolchains.git/commit/?id=a71589395ea15a9652d0a8785af5f6d18d9b79a8
> commit a71589395ea15a9652d0a8785af5f6d18d9b79a8
> Author: Chris Young <chris@unsatisfactorysoftware.co.uk>
> Commit: Chris Young <chris@unsatisfactorysoftware.co.uk>
>
> Update to SDK 53.29

Please can somebody kick off a rebuild of the ppc-amigaos toolchain -
NetSurf will be needing the new SDK at some point soon.

Thanks
Chris

[gccsdk] [Bug 193] Build failure for libfreetype6

http://www.riscos.info/bugzilla3/show_bug.cgi?id=193

Lee <leenoar@sky.com> changed:

What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
CC| |leenoar@sky.com
Resolution| |FIXED

--- Comment #3 from Lee <leenoar@sky.com> 2015-08-31 10:36:25 PDT ---
Builds OK now.

--
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 258] gggggg

http://www.riscos.info/bugzilla3/show_bug.cgi?id=258

Lee <leenoar@sky.com> changed:

What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
CC|prataplachibtech@gmail.com |leenoar@sky.com
Resolution| |INVALID

--- Comment #1 from Lee <leenoar@sky.com> 2015-08-31 10:30:13 PDT ---
Bad title, no description. Marking as invalid.

--
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:

  • In article <55E0BE38.9060102@paymentlabs.com>, Malcolm Hussain-Gambles
    <malcolm@paymentlabs.com> wrote:

    > On 28/08/15 20:39, Dave Higton wrote:
    > > In message <1573487451.24097.1440772172686.JavaMail.zimbra@paymentlabs.com>
    > > "Malcolm A. Hussain-Gambles" <malcolm@paymentlabs.com>
    > > wrote:
    > >
    > >> I've noticed whilst doing some html work that netsurf will
    > >> barf/break where other browsers don't. It's always down to a fault
    > >> with my html, that's one of the many reasons I'm now using netsurf
    > >> for development work, then testing with other browsers. It only used
    > >> to be IE that "auto-corrected" badly formed html, it's sad Chrome
    > >> and Firefox are now up to those dirty tricks too. So I would say it
    > >> is definitely a very bad browser bug, just not in Netsurf ;-)
    > > You (all) do make use of the W3C validator, I presume?
    > >
    > > http://validator.w3.org/
    > >
    > > I find it very useful. It works with Netsurf, of course.
    > >
    > I just find Netsurf quick and easy. I've used the validation service in
    > the past, but I never thought of using it for local pages. I didn't
    > think it could do! I'll definitely do that now once I finish a page,
    > thanks for the reminder!

    As I usually use PHP to construct web pages 'on the fly' uploading local
    pages for validation is not really an option. I have WebJames running so
    the validator then fetches what are local pages to me just as if I had
    uploaded them to their permanent home.

    --
    Tim Hill
    www.timil.com

    web sites * multimedia * training

    Sunday, 30 August 2015

    Re: Disable Javascript setting

    On 30/08/15 21:31, Dave Higton wrote:
    > I forget to disable Javascript almost every time I boot NS.
    >
    > Would it be possible to let the JS system initialise, /then/ obey
    > the non-volatile "disable JS" setting?

    Not until the transition to the new Javascript engine is complete, as
    Daniel has explained elsewhere. Until then, we want it enabled to
    increase the likelihood of bug reports. We have never made claims about
    the suitability of CI builds for any given purpose so, if you would
    prefer defined behaviour, please use the most recent stable release version.


    J.

    Re: Building for RISC OS

    On 30/08/15 21:00, Dave Higton wrote:
    > In message <OUT-55E2334A.MD-1.4.17.chris.young@unsatisfactorysoftware.co.uk>
    > "Chris Young" <chris.young@unsatisfactorysoftware.co.uk> wrote:
    >
    >> On Sat, 29 Aug 2015 21:07:47 +0100, Dave Higton wrote:
    >>
    >>> All of us on the NS users ML were reminded recently that there is no
    >>> maintainer of the RISC OS port of NS.
    >>>
    >>> I have looked a couple of times at the docs for building the various
    >>> ports of NS, but, unless I've missed something, there don't seem to
    >>> be any instructions for building for RISC OS.
    >>>
    >>> I understand that it's a cross compilation on a Linux machine (or a
    >>> very long wait for every build!). I really have very little idea
    >>> what I have to install on my Ubuntu box to build NS, nor what to do
    >>> when I have the build tools.
    >>>
    >>> So, have I missed something?
    >>
    >> This, maybe?
    >> http://wiki.netsurf-browser.org/Documentation/BuildingForRISCOSQuickStart
    >
    > Ah, thanks. Now I understand why I couldn't find it - the information
    > is in a completely different place from that for the other platforms,
    > and not linked in any meaningful way from the NS website front page.

    Well, so far as building things is concerned, that wiki page contains
    the exact same information as is in the QUICK-START guide in the source
    tree. What it does do is add some background around installing the cross
    compiler binaries for those who don't want to build it from source.

    > I have two Linux machines in the house, both of which run Ubuntu 14.04
    > 32 bit. I don't believe either of them is 64 bit capable. Oh well...
    >
    You can always build the RISC OS toolchain from source for your system
    -- we only provide binaries for amd64 as that happens to be what all the
    CI buildslaves use. The toolchain binaries are a side effect of the CI
    and are not meant as a fully supported deliverable of the project.

    To build the toolchain:

    git clone git://git.netsurf-browser.org/toolchains.git

    Then follow the instructions in the README. Note that the toolchain
    build will take a number of hours to complete and is particularly
    fragile, so you will have to start from scratch in the event that
    something goes wrong.

    > Given a Linux host, is it necessary or desirable to create a separate
    > VM for NS development?

    Not if you don't want to, no.


    J.

    Disable Javascript setting

    I forget to disable Javascript almost every time I boot NS.

    Would it be possible to let the JS system initialise, /then/ obey
    the non-volatile "disable JS" setting?

    Dave

    ____________________________________________________________
    FREE 3D MARINE AQUARIUM SCREENSAVER - Watch dolphins, sharks & orcas on your desktop!
    Check it out at http://www.inbox.com/marineaquarium

    Re: Building for RISC OS

    In message <8e386bfb54.DaveMeUK@my.inbox.com>
    Dave Higton <dave@davehigton.me.uk> wrote:

    >I have two Linux machines in the house, both of which run Ubuntu 14.04
    >32 bit. I don't believe either of them is 64 bit capable.

    Correction - the one I'd prefer to use is 64 bit capable. However,
    the Ubuntu VM manager requires a 64 bit host OS, so I would have to
    reinstall Ubuntu anyway. This is not unduly difficult to do as I
    have very few personal data on the machine, so there wouldn't be a
    lot to do to be fully up and running as before.

    Maybe I can dual boot 32 bit and 64 bit Ubuntu, which would appear
    to be an alternative solution?

    Dave

    ____________________________________________________________
    Send any screenshot to your friends in seconds...
    Works in all emails, instant messengers, blogs, forums and social networks.
    TRY IM TOOLPACK at http://www.imtoolpack.com/default.aspx?rc=if2 for FREE

    Re: Building for RISC OS

    In message <OUT-55E2334A.MD-1.4.17.chris.young@unsatisfactorysoftware.co.uk>
    "Chris Young" <chris.young@unsatisfactorysoftware.co.uk> wrote:

    >On Sat, 29 Aug 2015 21:07:47 +0100, Dave Higton wrote:
    >
    >> All of us on the NS users ML were reminded recently that there is no
    >> maintainer of the RISC OS port of NS.
    >>
    >> I have looked a couple of times at the docs for building the various
    >> ports of NS, but, unless I've missed something, there don't seem to
    >> be any instructions for building for RISC OS.
    >>
    >> I understand that it's a cross compilation on a Linux machine (or a
    >> very long wait for every build!). I really have very little idea
    >> what I have to install on my Ubuntu box to build NS, nor what to do
    >> when I have the build tools.
    >>
    >> So, have I missed something?
    >
    >This, maybe?
    >http://wiki.netsurf-browser.org/Documentation/BuildingForRISCOSQuickStart

    Ah, thanks. Now I understand why I couldn't find it - the information
    is in a completely different place from that for the other platforms,
    and not linked in any meaningful way from the NS website front page.

    Anyway...

    I have two Linux machines in the house, both of which run Ubuntu 14.04
    32 bit. I don't believe either of them is 64 bit capable. Oh well...

    Given a Linux host, is it necessary or desirable to create a separate
    VM for NS development?

    Dave

    ____________________________________________________________
    FREE 3D MARINE AQUARIUM SCREENSAVER - Watch dolphins, sharks & orcas on your desktop!
    Check it out at http://www.inbox.com/marineaquarium

    Re: Do Netsurf's errors get logged? (Syslog, Wimplog)

    In article <54fad9baa7chris@chris-johnson.org.uk>,
    cj <chris@chris-johnson.org.uk> wrote:
    > In article <54f9b39313GroverR@orpheusmail.co.uk>,
    > Rod at Orpheusmail <GroverR@orpheusmail.co.uk> wrote:

    >> Would you elaborate please, as I don't fully understand. (The bit about
    >> 'kill' facility that is).

    > I haven't sorted the best UI yet, which is why it is not in the release
    > version. Basically, as part of a job you could list tasks that should be
    > killed if running (and restarted again at the end of the job). For example,
    > if you have a job to backup your Pluto directory and data files, you could
    > include Pluto in a kill file for that job. When it came to this job,
    > SyncDiscs would check whether Pluto was running, and if so, issue a task
    > close message before starting the backup. There would then be no chance of
    > the data files changing part way through the backup, and hence the backup
    > of Pluto ending up in an inconsistent state. This could be extended to
    > restart Pluto once the job has completed.

    > However, I am not sure whether such a facility would be of general interest.

    Thanks for that Chris. And I do think that it would be a useful feature.

    Kind regards,
    Rod

    --
    Rod Grover
    GroverR@orpheusmail.co.uk

    Saturday, 29 August 2015

    Re: Do Netsurf's errors get logged? (Syslog, Wimplog)

    On 29 Aug at 18:30, cj <chris@chris-johnson.org.uk> wrote:

    > In article <54f9b39313GroverR@orpheusmail.co.uk>,
    > Rod at Orpheusmail <GroverR@orpheusmail.co.uk> wrote:
    > > Would you elaborate please, as I don't fully understand. (The bit
    > > about 'kill' facility that is).
    >
    > I haven't sorted the best UI yet, which is why it is not in the
    > release version. Basically, as part of a job you could list tasks that
    > should be killed if running (and restarted again at the end of the
    > job). For example, if you have a job to backup your Pluto directory
    > and data files, you could include Pluto in a kill file for that job.
    > When it came to this job, SyncDiscs would check whether Pluto was
    > running, and if so, issue a task close message before starting the
    > backup. There would then be no chance of the data files changing part
    > way through the backup, and hence the backup of Pluto ending up in an
    > inconsistent state. This could be extended to restart Pluto once the
    > job has completed.
    >
    > However, I am not sure whether such a facility would be of general
    > interest.

    Very definitely yes. I am shortly to get an ARM6X working and it will
    need backup. I remember enormous problems with the old SuncDiscs in
    dealing with running applications so a sleep and re-awaken feature would
    be much appreciated.

    --
    Tim Powys-Lybbe tim@powys.org
    for a miscellany of bygones: http://powys.org/

    Re: Building for RISC OS

    On Sat, 29 Aug 2015 21:07:47 +0100, Dave Higton wrote:

    > All of us on the NS users ML were reminded recently that there is no
    > maintainer of the RISC OS port of NS.
    >
    > I have looked a couple of times at the docs for building the various
    > ports of NS, but, unless I've missed something, there don't seem to
    > be any instructions for building for RISC OS.
    >
    > I understand that it's a cross compilation on a Linux machine (or a
    > very long wait for every build!). I really have very little idea
    > what I have to install on my Ubuntu box to build NS, nor what to do
    > when I have the build tools.
    >
    > So, have I missed something?

    This, maybe?
    http://wiki.netsurf-browser.org/Documentation/BuildingForRISCOSQuickStart

    Chris

    Building for RISC OS

    All of us on the NS users ML were reminded recently that there is no
    maintainer of the RISC OS port of NS.

    I have looked a couple of times at the docs for building the various
    ports of NS, but, unless I've missed something, there don't seem to
    be any instructions for building for RISC OS.

    I understand that it's a cross compilation on a Linux machine (or a
    very long wait for every build!). I really have very little idea
    what I have to install on my Ubuntu box to build NS, nor what to do
    when I have the build tools.

    So, have I missed something?

    If not, I would suggest that putting the relevant docs up is a vital
    step towards getting a RISC OS maintainer.

    Dave

    ____________________________________________________________
    Can't remember your password? Do you need a strong and secure password?
    Use Password manager! It stores your passwords & protects your account.
    Check it out at http://mysecurelogon.com/manager

    Re: Do Netsurf's errors get logged? (Syslog, Wimplog)

    > In article <54f9b39313GroverR@orpheusmail.co.uk>,
    > Rod at Orpheusmail <GroverR@orpheusmail.co.uk> wrote:
    >> Would you elaborate please, as I don't fully understand. (The bit
    >> about 'kill' facility that is).

    cj wrote on 29 Aug:
    > I haven't sorted the best UI yet, which is why it is not in the
    > release version. Basically, as part of a job you could list tasks
    > that should be killed if running (and restarted again at the end of
    > the job). For example, if you have a job to backup your Pluto
    > directory and data files, you could include Pluto in a kill file for
    > that job. When it came to this job, SyncDiscs would check whether
    > Pluto was running, and if so, issue a task close message before
    > starting the backup. There would then be no chance of the data files
    > changing part way through the backup, and hence the backup of Pluto
    > ending up in an inconsistent state. This could be extended to restart
    > Pluto once the job has completed.a

    > However, I am not sure whether such a facility would be of general
    > interest.

    Err, umm, sorry to sound as if my initials are D R, but I'm lost:
    is this anything to do with the subject line?

    --
    Jim Nagel www.archivemag.co.uk

    Re: Do Netsurf's errors get logged? (Syslog, Wimplog)

    In article <54f9b39313GroverR@orpheusmail.co.uk>,
    Rod at Orpheusmail <GroverR@orpheusmail.co.uk> wrote:
    > Would you elaborate please, as I don't fully understand. (The bit
    > about 'kill' facility that is).

    I haven't sorted the best UI yet, which is why it is not in the
    release version. Basically, as part of a job you could list tasks
    that should be killed if running (and restarted again at the end of
    the job). For example, if you have a job to backup your Pluto
    directory and data files, you could include Pluto in a kill file for
    that job. When it came to this job, SyncDiscs would check whether
    Pluto was running, and if so, issue a task close message before
    starting the backup. There would then be no chance of the data files
    changing part way through the backup, and hence the backup of Pluto
    ending up in an inconsistent state. This could be extended to restart
    Pluto once the job has completed.

    However, I am not sure whether such a facility would be of general
    interest.

    --
    Chris Johnson

    Friday, 28 August 2015

    Re:

  • I just find Netsurf quick and easy.
    I've used the validation service in the past, but I never thought of
    using it for local pages.
    I didn't think it could do!
    I'll definitely do that now once I finish a page, thanks for the reminder!

    Cheers,

    Malcolm

    On 28/08/15 20:39, Dave Higton wrote:
    > In message <1573487451.24097.1440772172686.JavaMail.zimbra@paymentlabs.com>
    > "Malcolm A. Hussain-Gambles" <malcolm@paymentlabs.com> wrote:
    >
    >> I've noticed whilst doing some html work that netsurf will barf/break where
    >> other browsers don't. It's always down to a fault with my html, that's one
    >> of the many reasons I'm now using netsurf for development work, then
    >> testing with other browsers. It only used to be IE that "auto-corrected"
    >> badly formed html, it's sad Chrome and Firefox are now up to those dirty
    >> tricks too. So I would say it is definitely a very bad browser bug, just
    >> not in Netsurf ;-)
    > You (all) do make use of the W3C validator, I presume?
    >
    > http://validator.w3.org/
    >
    > I find it very useful. It works with Netsurf, of course.
    >
    > Dave
    >
    > ____________________________________________________________
    > FREE 3D EARTH SCREENSAVER - Watch the Earth right on your desktop!
    > Check it out at http://www.inbox.com/earth
    >

    Re:

  • In message <1573487451.24097.1440772172686.JavaMail.zimbra@paymentlabs.com>
    "Malcolm A. Hussain-Gambles" <malcolm@paymentlabs.com> wrote:

    > I've noticed whilst doing some html work that netsurf will barf/break where
    > other browsers don't. It's always down to a fault with my html, that's one
    > of the many reasons I'm now using netsurf for development work, then
    > testing with other browsers. It only used to be IE that "auto-corrected"
    > badly formed html, it's sad Chrome and Firefox are now up to those dirty
    > tricks too. So I would say it is definitely a very bad browser bug, just
    > not in Netsurf ;-)

    You (all) do make use of the W3C validator, I presume?

    http://validator.w3.org/

    I find it very useful. It works with Netsurf, of course.

    Dave

    ____________________________________________________________
    FREE 3D EARTH SCREENSAVER - Watch the Earth right on your desktop!
    Check it out at http://www.inbox.com/earth

    Re:

  • On Fri, 28 Aug 2015 11:03:31 -0600, Anthony J. Bentley wrote:

    > Dave Higton writes:
    > > On Fri, 28 Aug 2015 12:50:00 +0100 Richard Torrens wrote:
    > >
    > > > Doing a list of genealogical notes, I wanted to space them out so there
    > > > was a blank line between each record. So I replaced each <LI> by <P><LI>.
    > > >
    > > > This does not work in Netsurf - but does in Firefox and Chrome
    > > >
    > > > Is this a bug?
    > > >
    > > > The page in question is
    > > > www.torrens.org.uk/Genealogy/Torrens/DataScot/sasines.html
    > >
    > > <p> opens a paragraph. You need to close it again afterwards with </p>.
    >
    > No, in HTML <p> does not require a closing tag except in specific
    > circumstances. "The end of a <li>" is not one of those circumstances.

    Also, I don't think there's any guarantee that a new paragraph
    actually inserts a blank line, and starting a new paragraph in the
    middle of a list seems like an odd thing to do.

    I always use <br>, same as Rob suggested.

    > As for the original question, I would probably use CSS to add padding or
    > margin at the bottom of each li.

    That's probably the correct "modern" way to do it.

    Chris

    Re:

  • Dave Higton writes:
    > On Fri, 28 Aug 2015 12:50:00 +0100 Richard Torrens wrote:
    >
    > > Doing a list of genealogical notes, I wanted to space them out so there
    > > was a blank line between each record. So I replaced each <LI> by <P><LI>.
    > >
    > > This does not work in Netsurf - but does in Firefox and Chrome
    > >
    > > Is this a bug?
    > >
    > > The page in question is
    > > www.torrens.org.uk/Genealogy/Torrens/DataScot/sasines.html
    >
    > <p> opens a paragraph. You need to close it again afterwards with </p>.

    No, in HTML <p> does not require a closing tag except in specific
    circumstances. "The end of a <li>" is not one of those circumstances.

    http://www.w3.org/TR/html-markup/p.html

    As for the original question, I would probably use CSS to add padding or
    margin at the bottom of each li.

    --
    Anthony J. Bentley

    Re:

  • I've noticed whilst doing some html work that netsurf will barf/break where other browsers don't.
    It's always down to a fault with my html, that's one of the many reasons I'm now using netsurf
    for development work, then testing with other browsers.
    It only used to be IE that "auto-corrected" badly formed html, it's sad Chrome and Firefox are
    now up to those dirty tricks too.
    So I would say it is definitely a very bad browser bug, just not in Netsurf ;-)

    Cheers,

    Malcolm
    ----- Original Message -----
    From: "Rob Kendrick" <rjek@netsurf-browser.org>
    To: netsurf-users@netsurf-browser.org
    Sent: Friday, 28 August, 2015 13:53:18
    Subject: Re: <P><LI>

    On Fri, Aug 28, 2015 at 04:52:04AM -0800, Dave Higton wrote:
    > On Fri, 28 Aug 2015 12:50:00 +0100 Richard Torrens wrote:
    >
    > > Doing a list of genealogical notes, I wanted to space them out so there
    > > was a blank line between each record. So I replaced each <LI> by <P><LI>.
    > >
    > > This does not work in Netsurf - but does in Firefox and Chrome
    > >
    > > Is this a bug?
    > >
    > > The page in question is
    > > www.torrens.org.uk/Genealogy/Torrens/DataScot/sasines.html
    >
    > <p> opens a paragraph. You need to close it again afterwards with </p>.
    >
    > Paragraphs are rendered with blank lines between them.

    Alternatively, add padding using CSS or use <br />.

    B.

    Re:

  • On Fri, Aug 28, 2015 at 04:52:04AM -0800, Dave Higton wrote:
    > On Fri, 28 Aug 2015 12:50:00 +0100 Richard Torrens wrote:
    >
    > > Doing a list of genealogical notes, I wanted to space them out so there
    > > was a blank line between each record. So I replaced each <LI> by <P><LI>.
    > >
    > > This does not work in Netsurf - but does in Firefox and Chrome
    > >
    > > Is this a bug?
    > >
    > > The page in question is
    > > www.torrens.org.uk/Genealogy/Torrens/DataScot/sasines.html
    >
    > <p> opens a paragraph. You need to close it again afterwards with </p>.
    >
    > Paragraphs are rendered with blank lines between them.

    Alternatively, add padding using CSS or use <br />.

    B.

    RE:

  • On Fri, 28 Aug 2015 12:50:00 +0100 Richard Torrens wrote:

    > Doing a list of genealogical notes, I wanted to space them out so there
    > was a blank line between each record. So I replaced each <LI> by <P><LI>.
    >
    > This does not work in Netsurf - but does in Firefox and Chrome
    >
    > Is this a bug?
    >
    > The page in question is
    > www.torrens.org.uk/Genealogy/Torrens/DataScot/sasines.html

    <p> opens a paragraph. You need to close it again afterwards with </p>.

    Paragraphs are rendered with blank lines between them.

    Dave

    ____________________________________________________________
    FREE 3D EARTH SCREENSAVER - Watch the Earth right on your desktop!
    Check it out at http://www.inbox.com/earth

  • Doing a list of genealogical notes, I wanted to space them out so there
    was a blank line between each record. So I replaced each <LI> by <P><LI>.

    This does not work in Netsurf - but does in Firefox and Chrome

    Is this a bug?

    The page in question is
    www.torrens.org.uk/Genealogy/Torrens/DataScot/sasines.html

    --
    Richard Torrens.
    http://www.Torrens.org.uk for genealogy, natural history, wild food, walks, cats
    and more!

    Thursday, 27 August 2015

    Re: Do Netsurf's errors get logged? (Syslog, Wimplog)

    In article <54f213f81achris@chris-johnson.org.uk>,
    cj <chris@chris-johnson.org.uk> wrote:
    > In article <mpro.nsz8rv00ukqa700m9.netsurf@ypical-daemon.co.uk>,
    > Fred Bambrough <netsurf@ypical-daemon.co.uk> wrote:

    >> ... and the way to avoid it is to quit SparkFS using "FS too".

    > Agreed - that is why I have added a 'kill' facility to the version of
    > SyncDiscs I use myself.

    Hi Chris,

    Would you elaborate please, as I don't fully understand. (The bit about
    'kill' facility that is).

    Kind regards,
    Rod

    --
    Rod Grover
    GroverR@orpheusmail.co.uk

    Re: noscroll

    On Wed, Aug 26, 2015 at 11:04:57PM +0100, Jim Nagel wrote:
    >

    Although this reply is to Jim its intended to address the wider
    audience here explaining why so many sites now fail to render
    adequately.

    > http://www.allaboutcircuits.com/technical-articles/fritzing-takes-your-design-from-breadboard-to-pcb/
    >
    > If you make the window bigger, you see more of the text. But the
    > scrollbars don't change in response. Once the window fills the whole
    > screen, there's no way to view the remaining text. The cursor keys or
    > pagedown have no effect; nor does the scrollwheel.
    >
    > Never seen this phenomenon before. What's going on in the coding?
    >
    > If it's something in the site's CSS, is there any way of tweaking it
    > in Netsurf?

    Many of the sites you refer to are based on Jquery [1] or
    similar. The web has fundamentally changed in the last decade from
    declarative HTML for content and CSS for styling to almost pure code
    driven generation, including styling.

    As you are discovering this *requires* extensive, performant and
    correct JavaScript operation to display anything, let alone good
    output. Coupled to this we need a layout engine that can cope with
    dynamic updates and layered ordering.

    All the specifications for this technology is continuously changing
    and differs from day to day (yes seriously)

    In an attempt to improve the situation we are in the process of
    changing JavaScript engine and providing implementations for the
    massive DOM (207 classes and in excess of 3500 methods).

    However with only a few of us working on it part time progress is
    excruciatingly slow. For comparison the team size at Google, Mozilla
    and Microsoft working on this area number in the tens of full time
    engineers.

    I am aiming for a 3.4 release before the end of October with the
    updated Javascript engine and the multitude of other fixes that have
    happened since 3.3 [2] This release should hopefully be better for all
    supported architectures.

    However as repeatedly stated previously RISC OS will *NOT* be a
    release target for 3.4 unless someone can come forward who will be
    responsible for that OS. At best the CI master builds will be
    available but we offer no guarantee about the stability of those.

    >
    > First noticed this using Netsurf #2789 on Armini 5.22; same behaviour
    > now in Netsurf #2935 on Iyonix 5.18.
    >
    > --
    > Jim Nagel www.archivemag.co.uk
    >
    >

    [1] https://jquery.com/
    [2] http://bugs.netsurf-browser.org/mantis/changelog_page.php?version_id=16

    --
    Regards Vincent
    http://www.kyllikki.org/

    Wednesday, 26 August 2015

    noscroll

    http://www.allaboutcircuits.com/technical-articles/fritzing-takes-your-design-from-breadboard-to-pcb/

    If you make the window bigger, you see more of the text. But the
    scrollbars don't change in response. Once the window fills the whole
    screen, there's no way to view the remaining text. The cursor keys or
    pagedown have no effect; nor does the scrollwheel.

    Never seen this phenomenon before. What's going on in the coding?

    If it's something in the site's CSS, is there any way of tweaking it
    in Netsurf?

    First noticed this using Netsurf #2789 on Armini 5.22; same behaviour
    now in Netsurf #2935 on Iyonix 5.18.

    --
    Jim Nagel www.archivemag.co.uk

    black background obscures most of page

    Most of this page is obscured by a black background that overprints
    all but the first inch or two at the top of the page:
    http://www.combehillwood.co.uk/polden-way/

    This is using Netsurf #2935, the latest build, installed just now.
    It was the same on #2789 (dated about June 9).

    I'm curious to know what mechanism is at (mis)work here. Is it
    something in the CSS maybe? Or something in Javascript? I did a
    "full save" of the page in order to inspect the CSS, but it's beyond
    me. (One CSS file is nearly 700K.)

    I reported a possibly similar problem last spring about a CNET page
    that Netsurf #2644 appeared to render initially; it then immediately
    obscured most of it with a vertical grey band. Could that have been
    the same mechanism at work?
    Behaviour of that page is different now with #2935: even less of
    the text is visible.
    http://www.cnet.com/products/synology-diskstation-ds412-plus/

    As I say, I'm curious about the mechanism. If it's Javascript, I
    appreciate the vastness of the programming job.

    --
    Jim Nagel www.archivemag.co.uk

    Re: El Reg formats differently these days - why?

    In message <55DC56B6.6060505@netsurf-browser.org>
    Michael Drake <tlsa@netsurf-browser.org> wrote:

    >
    >
    > On 21/08/15 19:36, Dave Higton wrote:
    > > NS used to render The Register in three columns, like other browsers.
    > > Recently it renders with each line of three headings broken over two
    > > lines, with the additional oddity that some single items appear left
    > > justified and some appear centred. Go to http://www.theregister.co.uk
    > > and you'll see.
    >
    > They changed the way they did the layout, exposing a layout issue in
    > NetSurf.
    >
    > NetSurf CI builds from 2934 contain a fix.

    Indeed it's back to three columns now.

    My thanks to you - and it's a handy reminder to repeat my thanks to
    the whole NS team for your continued sterling work.

    Dave

    ____________________________________________________________
    Can't remember your password? Do you need a strong and secure password?
    Use Password manager! It stores your passwords & protects your account.
    Check it out at http://mysecurelogon.com/manager

    Tuesday, 25 August 2015

    Re: El Reg formats differently these days - why?

    On 25 Aug 2015, Michael Drake <tlsa@netsurf-browser.org> wrote:
    > On 21/08/15 19:36, Dave Higton wrote:

    > > NS used to render The Register in three columns, like other
    > > browsers. Recently it renders with each line of three headings
    > > broken over two lines, with the additional oddity that some single
    > > items appear left justified and some appear centred. Go to
    > > http://www.theregister.co.uk and you'll see.
    >
    > They changed the way they did the layout, exposing a layout issue in
    > NetSurf.
    >
    > NetSurf CI builds from 2934 contain a fix.

    NetSurf #2935, jsoff or json, running on a RiscPC/RO 6.20, displays 3+1
    columns, which agrees with the layout on Firefox 39, running on Win7.
    However, in NS the headline image is missing, and an inoperative
    vertical scroll bar (absent in FF) is drawn at the right of the page.

    Tony

    Re: El Reg formats differently these days - why?

    On 21/08/15 19:36, Dave Higton wrote:
    > NS used to render The Register in three columns, like other browsers.
    > Recently it renders with each line of three headings broken over two
    > lines, with the additional oddity that some single items appear left
    > justified and some appear centred. Go to http://www.theregister.co.uk
    > and you'll see.

    They changed the way they did the layout, exposing a layout issue
    in NetSurf.

    NetSurf CI builds from 2934 contain a fix.

    --
    Michael Drake http://www.netsurf-browser.org/

    Monday, 24 August 2015

    Re: [gccsdk] Converting ELF into AOF or ALF?

    An alternative is to use the older GCC-AOF compiler/toolset to build your ALF stuff and the newer GCC-ELF compiler/toolset for the ELF stuff.

    Regards,

    Nick

    On 15 August 2015 at 12:54, Steve Fryatt <lists@stevefryatt.org.uk> wrote:
    For reasons of convenience, it would useful for me to be able to generate
    libraries in ALF format alongside the ELF versions that I currently produce
    from GCCSDK. I can obviously do it via Norcroft, but that means munging the
    Linux-based source files into a structure that Norcroft will work with and
    keeping the two sets in step with each other.

    Following a similar discussion on the ROOL forums recently, I was wondering
    about a tool to convert the ELF object files into AOF (and then ALF) --
    similar in concept to elf2aif, I suppose.

    Having had a quick look at the two formats, it seems that it might be
    possible to convert one into the other, but there's a nagging worry that
    even if I handle all of the symbol and relocation details correctly, the
    actual structures of the compiled code and data might not be compatible. Any
    files would be linked to the Shared C Library in the final instance, if that
    makes a difference.

    So, before I waste too much time on this side project, is anyone already
    able to say that it /won't/ work for whatever reason? If so, there seems
    little point spending time reaching the same conclusion that other people
    have already reached before me!

    Or, even better, is there already a solution out there that Google has
    missed?

    --
    Steve Fryatt - Leeds, England

    http://www.stevefryatt.org.uk/

    _______________________________________________
    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] New VFP/NEON capable GCC release?

    > Lee Noar wrote on Thursday, August 20, 2015 1:54 PM

    > On 19/08/15 23:11, Theo Markettos wrote:

    > [snip]

    > > Today's build is number 32, where I bumped the release number to 2.
    > > This
    > > works on my Raspberry Pi with no problems (looks like my RPCEmu hostfs
    > > image
    > > is broken somehow, which is why I was a bit suspicious of the previous
    > > build)

    > I also had an issue with !PackMan in RPCEmu. I tried to install the 3
    > shared library packages (base, C, C++), but it would only let me install
    > one of them. If I tried to install a second, it would insist on
    > removing the first at the same time. On the RPi the same package files
    > installed with no problems. Possibly something to do with !PackMan's
    > interaction with hostfs.

    I've just tried this on RISC OS 5 installed on RPCEmu on Windows using
    HostFS and the installation seemed to work correctly.

    !PackMan shouldn't have a problem with it's interaction with hostfs as
    it treats it the same as any other filing system.

    I have seen problems previous on Windows where hostfs couldn't deal
    with a character in a filename which caused problems.

    If you go onto the PackMan advanced menu from the icon bar, there is
    a view log option. Go into it first time to turn the log in. Run your
    install
    and then after it fails go back to that menu option and it may show you
    some more useful information on what went wrong.

    After I had installed it I tried to build the C++ examples with GCC and
    they failed with:

    g++ -mlibscl -O3 -o template_scl template.cc
    /tmp/ccfdiE8p.o: In function `_GLOBAL__sub_I_ip':
    template.cc:(.text.startup+0x98): undefined reference to `__cxa_atexit'
    /HostFS::HostFS.$/Apps/Development/!GCC/bin/../lib/gcc/arm-unknown-riscos/4.7.4/../../../scl/hard/fpa/libstdc++.a(eh_globals.o):
    In function `__static_initialization_and_destruction_0':
    /home/jenkins/workspace/gcc-4.7-native/gcc4/srcdir/gcc/libstdc++-v3/libsupc++/eh_globals.cc:110:
    undefined reference to `__cxa_atexit'
    collect2: error: ld returned 1 exit status
    make: *** [template_scl] Error 1

    The C examples compiled OK.

    Regards,
    Alan


    _______________________________________________
    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, 23 August 2015

    Re: El Reg formats differently these days - why?

    On 23 Aug 2015, Dave Higton <dave@davehigton.me.uk> wrote:

    [snip]

    > Still the same old three columns, and it does appear to be the desktop
    > site rather than a special mobile version.

    Here, using a smart phone, running Android 2.3.6 , a Google search for
    'the register' leads to the mobile site http://m.theregister.co.uk .
    which is a single column layout.

    On the same device, http://www.theregister.co.uk displays a _four_
    column layout, the same as that on a Win7 machine, running Firefox.

    On a RiscPC, NS #2817 jsoff also displays http://m.theregister.co.uk as
    a single column.

    On a RiscPC, NS #2817 jsoff displays http://www.theregister.co.uk as
    _three_ columns (not four, as FF).

    Tony

    Re: El Reg formats differently these days - why?

    In message <3ab151f754.DaveMeUK@my.inbox.com>
    Dave Higton <dave@davehigton.me.uk> wrote:

    >In message <nsc50f37f754.Bryn@yo.rk>
    > Bryn Evans <netsurf@bryork.freeuk.com> wrote:
    >
    >>In a mad moment - Dave Higton mumbled :
    >>
    >>> NS used to render The Register in three columns, like other browsers.
    >>> Recently it renders with each line of three headings broken over two
    >>> lines, with the additional oddity that some single items appear left
    >>> justified and some appear centred. Go to http://www.theregister.co.uk
    >>> and you'll see.
    >>
    >>> I thought it was caused by a change to NS, but I just went back to a
    >>> version from June (which, I'm sure, is before it started happening),
    >>> and it's just the same.
    >>
    >>> So I have to assume that El Reg have changed their layout in some way.
    >>
    >>> The HTML is beyond my comprehension. Can anyone else explain what's
    >>> going on?
    >>
    >>it is the same using Firefox and Linux, so i guess they have made
    >>a cockup of some change - It happened mid week.
    >
    >???
    >
    >Not here, it isn't - FF 40.0, Ubuntu 14.04 fully up to date.

    Nor is it on my mobile phone (Android). Still the same old three
    columns, and it does appear to be the desktop site rather than a
    special mobile version.

    Dave

    ____________________________________________________________
    Can't remember your password? Do you need a strong and secure password?
    Use Password manager! It stores your passwords & protects your account.
    Check it out at http://mysecurelogon.com/password-manager

    Saturday, 22 August 2015

    Re: El Reg formats differently these days - why?

    In message <nsc50f37f754.Bryn@yo.rk>
    Bryn Evans <netsurf@bryork.freeuk.com> wrote:

    >In a mad moment - Dave Higton mumbled :
    >
    >> NS used to render The Register in three columns, like other browsers.
    >> Recently it renders with each line of three headings broken over two
    >> lines, with the additional oddity that some single items appear left
    >> justified and some appear centred. Go to http://www.theregister.co.uk
    >> and you'll see.
    >
    >> I thought it was caused by a change to NS, but I just went back to a
    >> version from June (which, I'm sure, is before it started happening),
    >> and it's just the same.
    >
    >> So I have to assume that El Reg have changed their layout in some way.
    >
    >> The HTML is beyond my comprehension. Can anyone else explain what's
    >> going on?
    >
    >it is the same using Firefox and Linux, so i guess they have made
    >a cockup of some change - It happened mid week.

    ???

    Not here, it isn't - FF 40.0, Ubuntu 14.04 fully up to date.

    Dave

    ____________________________________________________________
    FREE 3D EARTH SCREENSAVER - Watch the Earth right on your desktop!
    Check it out at http://www.inbox.com/earth

    Re: El Reg formats differently these days - why?

    In message <nsc50f37f754.Bryn@yo.rk>
    Bryn Evans <netsurf@bryork.freeuk.com> wrote:

    > In a mad moment - Dave Higton mumbled :
    >
    [snip]
    >
    >> So I have to assume that El Reg have changed their layout in some way.
    >
    >> The HTML is beyond my comprehension. Can anyone else explain what's
    >> going on?
    >
    > it is the same using Firefox and Linux, so i guess they have made
    > a cockup of some change - It happened mid week.
    >
    Or possibly they're trying to become more smartphone-friendly?

    --
    George

    Re: El Reg formats differently these days - why?

    In a mad moment - Dave Higton mumbled :

    > NS used to render The Register in three columns, like other browsers.
    > Recently it renders with each line of three headings broken over two
    > lines, with the additional oddity that some single items appear left
    > justified and some appear centred. Go to http://www.theregister.co.uk
    > and you'll see.

    > I thought it was caused by a change to NS, but I just went back to a
    > version from June (which, I'm sure, is before it started happening),
    > and it's just the same.

    > So I have to assume that El Reg have changed their layout in some way.

    > The HTML is beyond my comprehension. Can anyone else explain what's
    > going on?

    it is the same using Firefox and Linux, so i guess they have made
    a cockup of some change - It happened mid week.

    --
    |)    [
    |)ryn [vans mail to - BrynEvans@bryork.freeuk.com

    Friday, 21 August 2015

    Re: Error on Yahoo news

    On Fri, 21 Aug 2015 18:40:18 , Peter Slegg <p.slegg@scubadivers.co.uk> wrote:
    > > Date: Fri, 21 Aug 2015 14:18:44 +0100
    > > From: "Richard Torrens (lists)" <Lists@Torrens.org.uk>
    > > Subject: Re: Error on Yahoo news
    > > To: <netsurf-users@netsurf-browser.org>
    > > Peter Slegg <p.slegg@scubadivers.co.uk> wrote:
    > > > This url uk.news.yahoo.com just gives me the error
    > >
    > > It strikes me that that is not a URL.
    > >
    > > As this is a Netsurf list, do you mean http://uk.news.yahoo.com ? That
    > > orks here. As does https://uk.news.yahoo.com
    > >
    >
    > Is the http really necessary ? Shouldn't it just
    > direct to http://uk.news.yahoo.com in any case ?
    >
    > The link I was given wasn't https but the https page does try to
    > open, sometimes, it keeps timing out and often the css layout is
    > missing.

    PS
    I had actually followed the News link from http://uk.yahoo.com
    and that was to http://uk.news.yahoo.com and gives the GET error.

    Peter

    El Reg formats differently these days - why?

    NS used to render The Register in three columns, like other browsers.
    Recently it renders with each line of three headings broken over two
    lines, with the additional oddity that some single items appear left
    justified and some appear centred. Go to http://www.theregister.co.uk
    and you'll see.

    I thought it was caused by a change to NS, but I just went back to a
    version from June (which, I'm sure, is before it started happening),
    and it's just the same.

    So I have to assume that El Reg have changed their layout in some way.

    The HTML is beyond my comprehension. Can anyone else explain what's
    going on?

    Dave

    ____________________________________________________________
    FREE 3D MARINE AQUARIUM SCREENSAVER - Watch dolphins, sharks & orcas on your desktop!
    Check it out at http://www.inbox.com/marineaquarium

    Re: Error on Yahoo news

    In article <54f6a3f263Lists@Torrens.org.uk>, Richard Torrens (lists)
    <Lists@Torrens.org.uk> wrote:
    > In article <00110f20.031e94909f96@smtp.freeola.net>, Peter Slegg
    > <p.slegg@scubadivers.co.uk> wrote:
    > > This url uk.news.yahoo.com just gives me the error

    > It strikes me that that is not a URL.

    It is launched by Pluto and opens in Netsurf. Did you try it?

    [Snip]

    --
    Tim Hill
    www.timil.com

    web sites * multimedia * training

    Re: Error on Yahoo news

    In article <00110f20.031e94909f96@smtp.freeola.net>,
    Peter Slegg <p.slegg@scubadivers.co.uk> wrote:
    > This url uk.news.yahoo.com just gives me the error

    It strikes me that that is not a URL.

    As this is a Netsurf list, do you mean http://uk.news.yahoo.com ? That
    orks here. As does https://uk.news.yahoo.com

    --
    Richard Torrens.
    http://www.Torrens.org.uk for genealogy, natural history, wild food, walks, cats
    and more!

    Re: Error on Yahoo news

    > Date: Thu, 20 Aug 2015 21:58:34 +0100
    > From: Dave Higton
    > Subject: Re: Error on Yahoo news
    > Peter Slegg wrote:
    >
    > >
    > >This url uk.news.yahoo.com just gives me the error
    > >"Could not process this "GET" request.
    > >
    > >This is on Atari build 2915.
    > >
    > >Just wondered if it is repeatable for others.
    >
    > I'm using RISC OS build 2921, and it doesn't happen here.
    >
    > Dave
    >

    It's the same with build 2926.

    Peter

    Thursday, 20 August 2015

    Re: Error on Yahoo news

    In message <00110f20.031e94909f96@smtp.freeola.net>
    Peter Slegg <p.slegg@scubadivers.co.uk> wrote:

    >
    >This url uk.news.yahoo.com just gives me the error
    >"Could not process this "GET" request.
    >
    >This is on Atari build 2915.
    >
    >Just wondered if it is repeatable for others.

    I'm using RISC OS build 2921, and it doesn't happen here.

    Dave

    ____________________________________________________________
    FREE ONLINE PHOTOSHARING - Share your photos online with your friends and family!
    Visit http://www.inbox.com/photosharing to find out more!

    Re: [gccsdk] Converting ELF into AOF or ALF?

    On 15 Aug, John Tytgat wrote in message
    <55CF4EDA.90706@aaug.net>:

    > On 08/15/2015 01:54 PM, Steve Fryatt wrote:
    >
    > > [...] So, before I waste too much time on this side project, is anyone
    > > already able to say that it /won't/ work for whatever reason? If so,
    > > there seems little point spending time reaching the same conclusion that
    > > other people have already reached before me!
    >
    > As conclusion, I believe this is a very reasonable idea/project, not
    > straightforward but feasible.
    >
    > I'm not wearing a GCCSDK hat anymore but actually if you would be doing
    > such a project, it would be a great addition to the existing gccsdk
    > toolset.
    >
    > > Or, even better, is there already a solution out there that Google has
    > > missed?
    >
    > I'm not aware of any off-the-shelve solutions. AOF/ALF has long time been
    > left along by the non-RISC OS/Acorn community, everyone is using ELF (or
    > COFF) these days.
    >
    > Hope this helps, John.

    It does indeed; many thanks.

    I'll have a look into Ben's suggestion first, but if that fails I've now got
    a much better idea about the issues that I'd need to consider in a
    converter. If nothing else, I've gained a much better understanding of what
    the AOF and equivalent ELF files actually contain as a result!

    --
    Steve Fryatt - Leeds, England

    http://www.stevefryatt.org.uk/

    _______________________________________________
    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] New VFP/NEON capable GCC release?

    On 19/08/15 23:11, Theo Markettos wrote:

    [snip]

    > Today's build is number 32, where I bumped the release number to 2. This
    > works on my Raspberry Pi with no problems (looks like my RPCEmu hostfs image
    > is broken somehow, which is why I was a bit suspicious of the previous
    > build)

    I also had an issue with !PackMan in RPCEmu. I tried to install the 3
    shared library packages (base, C, C++), but it would only let me install
    one of them. If I tried to install a second, it would insist on
    removing the first at the same time. On the RPi the same package files
    installed with no problems. Possibly something to do with !PackMan's
    interaction with hostfs.

    Lee.


    _______________________________________________
    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

    Error on Yahoo news

    This url uk.news.yahoo.com just gives me the error
    "Could not process this "GET" request.

    This is on Atari build 2915.

    Just wondered if it is repeatable for others.

    Peter

    Wednesday, 19 August 2015

    Re: [gccsdk] New VFP/NEON capable GCC release?

    On Wed, Aug 19, 2015 at 04:53:23PM +0100, Lee Noar wrote:
    > From build #28:
    >
    > <http://ci.riscos.info/job/gcc-4.7-native/28/artifact/gcc4/release-area/riscpkg/autobuilder_website/arm/Development/>
    >
    > I've installed the packages using !PackMan on my RPi and everything
    > looks OK. I've built a simple program that uses FP, using soft-float
    > and VFP, both dynamically and statically and the resulting binaries
    > run correctly.

    Thanks. Over the last few days I did some more kicking of Jenkins and it
    now generates PackMan package lists for each GCC build, so just add

    http://ci.riscos.info/job/gcc-4.7-native/lastSuccessfulBuild/artifact/gcc4/release-area/riscpkg/autobuilder_website/pkg/autobuilt

    to your PackMan sources and it'll install the nightly build for you.
    Currently this only applies to gcc-4.7-native, but in principle can apply to
    other packages too. Nightlies won't update properly because we aren't
    bumping the package version, but that could be fixed.

    Today's build is number 32, where I bumped the release number to 2. This
    works on my Raspberry Pi with no problems (looks like my RPCEmu hostfs image
    is broken somehow, which is why I was a bit suspicious of the previous
    build)

    I announced that on the ROOL VFP thread. Unless there are any major
    showstoppers I hope to push build 32 as the official release in a few days.

    Theo

    _______________________________________________
    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] New VFP/NEON capable GCC release?

    On 15/08/15 13:46, Lee Noar wrote:
    > On 13/08/15 18:31, Theo Markettos wrote:
    >> On Tue, Aug 11, 2015 at 10:06:23AM +0100, Theo Markettos wrote:
    >>> On Mon, Aug 10, 2015 at 04:11:42PM +0100, Lee Noar wrote:
    >>>> I think this is done now. There wasn't much more to do after all.
    >>>> Perhaps someone could double check that the archives extract to a fully
    >>>> working installation. I have tried extracting to a new location and
    >>>> building a simple C program with some FP in it both with and without
    >>>> VFP and all seems well.
    >>>
    >>> Thanks. Some builds:
    >>>
    >>> http://ci.riscos.info/job/gcc-4.7-native/9/artifact/gcc4/release-area/riscpkg/
    >>>
    >>>
    >>> I'll try and give them a spin when I get a moment.
    >>
    >> FWIW I'm unlikely to get a chance to test these for the next week or so,
    >> possibly more. If anyone has the time to give them a quick spin, that
    >> would
    >> be most helpful.
    >
    > When Jenkins has finished building with the changes I've just made,
    > I'll give them a try and let you know.

    From build #28:

    <http://ci.riscos.info/job/gcc-4.7-native/28/artifact/gcc4/release-area/riscpkg/autobuilder_website/arm/Development/>

    I've installed the packages using !PackMan on my RPi and everything
    looks OK. I've built a simple program that uses FP, using soft-float
    and VFP, both dynamically and statically and the resulting binaries
    run correctly.

    Lee.

    _______________________________________________
    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

    nsgenbind can't generete needed files

    I am compiling newest sources on framebuffer.
    I get this error either on widons and linux:

    GENBIND: javascript/duktape/netsurf.bnd
    4: syntax error, unexpected TOK_IDENTIFIER
    Error: parse failed with code 1



    Test results:

    https://www.dropbox.com/s/oh7jk8bdckbew4i/testlog?dl=0

        TEST: blankidl......FAIL
        TEST: document......FAIL
        TEST: dom......FAIL
        TEST: emptyidl......FAIL
        TEST: htmldocument......PASS
        TEST: htmldocument2......PASS
        TEST: window......PASS

    Tuesday, 18 August 2015

    Re: [gccsdk] New VFP/NEON capable GCC release?

    In article <20150813173139.GH22681@chiark.greenend.org.uk>,
    Theo Markettos <theo@markettos.org.uk> wrote:
    > On Tue, Aug 11, 2015 at 10:06:23AM +0100, Theo Markettos wrote:
    > > On Mon, Aug 10, 2015 at 04:11:42PM +0100, Lee Noar wrote:
    > > > I think this is done now. There wasn't much more to do after all.
    > > > Perhaps someone could double check that the archives extract to a fully
    > > > working installation. I have tried extracting to a new location and
    > > > building a simple C program with some FP in it both with and without
    > > > VFP and all seems well.
    > >
    > > Thanks. Some builds:
    > >
    > > http://ci.riscos.info/job/gcc-4.7-native/9/artifact/gcc4/release-area/riscpkg/
    > >
    > > I'll try and give them a spin when I get a moment.

    > FWIW I'm unlikely to get a chance to test these for the next week or so,
    > possibly more. If anyone has the time to give them a quick spin, that would
    > be most helpful.

    I did a quick test by building the example C programs with softfloat and
    vfp. Everything ran ok.

    gcc -v & g++ -v still gives,

    gcc version 4.7.4 (GCCSDK GCC 4.7.4 Release 1 Development)

    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

    Monday, 17 August 2015

    Re: Fail 2890 json

    On 17/08/15 07:20, Andrew Pinder wrote:
    > If I understand the position correctly, UnixLib is part of the
    > development environment you use and isn't part of RISCOS itself? So
    > once it has been fixed I won't have to install an update on this
    > machine?

    UnixLib implements the C standard library and some of the POSIX API.
    It gets built into NetSurf, and other RISC OS software built with GCC.

    UnixLib turned out to be a red herring. The fault was actually in the
    compiler. John-Mark tracked it down, and has fixed it [1] for our
    toolchain. He's also sent the fix upstream to the GCCSDK
    maintainers [2].

    Anyway, you don't have to do anything particular to get the fix, just
    run CI build #2911 or later.

    1.
    http://git.netsurf-browser.org/toolchains.git/commit/?id=e88575dbac2075d574d65bcc91f0e345ee2110d2

    2. http://www.riscos.info/pipermail/gcc/2015-August/006389.html

    --
    Michael Drake http://www.netsurf-browser.org/

    Sunday, 16 August 2015

    Re: Fail 2890 json

    In message <20150814072335.GB10674@somnambulist.local>
    on 14 Aug 2015 Daniel Silverstone <dsilvers@netsurf-browser.org>
    wrote:

    > We've already found what we believe to be a bug in UNIXLib and we are
    > investingating it -- it can cause the browser to segfault and thus vanish
    > without apparent error.

    > Sadly I can't give you an ETA on this, though I hope we'll be fairly close to
    > our goal by the end of next week.

    > Some of us (myself and Vince included) have been spending a huge
    > amount of our
    > time on this, we're not leaving you in the lurch; and John-Mark has been
    > instrumental in helping track down the UNIXLib issue.

    Thanks, as always, for your work on NetSurf.

    If I understand the position correctly, UnixLib is part of the
    development environment you use and isn't part of RISCOS itself? So
    once it has been fixed I won't have to install an update on this
    machine?


    Regards

    Andrew
    --
    Andrew Pinder

    Re: [Rpcemu] networkimng with windows 10

    In message <df6930f454.old_coaster@old_coaster.yahoo.co.uk>
    Tony Moore <old_coaster@yahoo.co.uk> wrote:

    > On 16 Aug 2015, george greenfield <george.greenfield@tiscali.co.uk>
    > wrote:
    >> In message <dfc30df454.old_coaster@old_coaster.yahoo.co.uk>
    >> Tony Moore <old_coaster@yahoo.co.uk> wrote:
    >
    > [snip]
    >
    [snip]
    >
    > No. The Win10 installation is by informed consent.
    >
    > [snip]
    >
    That's reassuring!

    --
    George

    _______________________________________________
    Rpcemu mailing list
    Rpcemu@riscos.info
    http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu

    Re: [Rpcemu] networkimng with windows 10

    On 16 Aug 2015, george greenfield <george.greenfield@tiscali.co.uk>
    wrote:
    > In message <dfc30df454.old_coaster@old_coaster.yahoo.co.uk>
    > Tony Moore <old_coaster@yahoo.co.uk> wrote:

    [snip]

    > Did you elect to install Win10,

    I 'reserved' a copy of Win10 via the 'How to get Windows 10 for free'
    window, accessed via the icon which appears in the notification tray.
    Later, over a period of several days, the data files for Win10 began to
    arrive on my machine - their total size is several GB and, since my,
    allegedly broadband, internet connection is capable of only 40 MB/sec on
    a good day, this took a while. When all the bits were present on the
    local machine, the upgrader suggested that installation could start now,
    or could be delayed to a time of my choice. I elected to 'start now'.
    The installation took several hours, but there were no glitches.

    > or did you discover that a routine 'upgrade' to Win7 had done it for
    > you?

    No. The Win10 installation is by informed consent.

    [snip]

    > I've switched off automatic upgrade downloading/installing to my Win7
    > machine: the PC prompts me that an upgrade is available, but I have
    > the option of scrutinising each patch before installing.

    From what I have seen, Win10 will not be installed until you give the
    go-ahead.

    > > It seems that Win10 (in common with XP) doesn't preserve the
    > > network- bridge over a re-boot, so that it is necessary to re-create
    > > the bridge, for each session. Can you confirm this?

    [snip]

    > If so, this is worrying.

    Agreed.

    Tony




    _______________________________________________
    Rpcemu mailing list
    Rpcemu@riscos.info
    http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu

    Re: [Rpcemu] networkimng with windows 10

    In message <dfc30df454.old_coaster@old_coaster.yahoo.co.uk>
    Tony Moore <old_coaster@yahoo.co.uk> wrote:

    > On 14 Aug 2015, Brian D <groups@planet3.freeuk.co.uk> wrote:
    >
    >> Until recently I've had a working system with networking with windows7
    >> home premium. I've "upgraded" to windows10, and networking no longer
    >> works. I've been through all the settings and can't find anything that
    >> has changed.
    >
    > Having upgraded from Win7 Home Premium to Win10 Home, a few days ago,
    [snip]

    Did you elect to install Win10, or did you discover that a routine
    'upgrade' to Win7 had done it for you? Excuse my paranoia, but stories
    of unwanted, automated upgrades to Win10 are circulating
    http://www.theregister.co.uk/2015/08/13/two_weeks_of_windows_10_how_is_microsoft_doing/

    I've switched off automatic upgrade downloading/installing to my Win7
    machine: the PC prompts me that an upgrade is available, but I have
    the option of scrutinising each patch before installing.

    >
    > It seems that Win10 (in common with XP) doesn't preserve the network-
    > bridge over a re-boot, so that it is necessary to re-create the bridge,
    > for each session. Can you confirm this?
    [snip]

    If so, this is worrying. RPCEmu is no longer my primary RISC OS
    platform, but any decision to return it to this status would depend
    upon stable, i.e., surviving reboot, networking.

    --
    George

    _______________________________________________
    Rpcemu mailing list
    Rpcemu@riscos.info
    http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu

    Re: [Rpcemu] networkimng with windows 10

    On 14 Aug 2015, Brian D <groups@planet3.freeuk.co.uk> wrote:

    > Until recently I've had a working system with networking with windows7
    > home premium. I've "upgraded" to windows10, and networking no longer
    > works. I've been through all the settings and can't find anything that
    > has changed.

    Having upgraded from Win7 Home Premium to Win10 Home, a few days ago, I
    was pleasantly surprised to find that the upgrader had correctly copied
    the settings of RPCEmu 0811 from Win7 to Win10 - even the TAP interface
    was correct. Unfortunately, networking didn't work.

    Re-building the network-bridge fixed it but, after re-booting the
    machine, networking was dead again.

    It seems that Win10 (in common with XP) doesn't preserve the network-
    bridge over a re-boot, so that it is necessary to re-create the bridge,
    for each session. Can you confirm this?

    Tony




    _______________________________________________
    Rpcemu mailing list
    Rpcemu@riscos.info
    http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu

    Saturday, 15 August 2015

    Re: [gccsdk] Converting ELF into AOF or ALF?

    On Sat, 15 Aug 2015 12:54:44 +0100, Steve Fryatt <lists@stevefryatt.org.uk> wrote:

    > Or, even better, is there already a solution out there that Google has
    > missed?

    Well, the ARM/Norcroft linker in the DDE is supposed to be able to
    convert object files back and forth between AOF and ELF. However,
    because ELF's many relocation types are something of a moving target
    and there hasn't been much motivation to work on its ELF support for
    many years, your mileage may vary...

    Ben

    _______________________________________________
    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] Converting ELF into AOF or ALF?

    On 08/15/2015 01:54 PM, Steve Fryatt wrote:
    [...]  Following a similar discussion on the ROOL forums recently, I was wondering  about a tool to convert the ELF object files into AOF (and then ALF) --  similar in concept to elf2aif, I suppose.    Having had a quick look at the two formats, it seems that it might be  possible to convert one into the other, 

    There are several aspects to consider:
    1. ELF is a standard defined data container which gets used as intermediate object code (alike AOF), static library file (alike ALF), executables (alike AIF), shared library, core dumps, etc.  So we're talking about targeting a subset of ELF files in order to convert them to AOF/ALF.
    2. The file formats itself : I guess you'll need a good library to access ELF data structures and AOF/ALF.  For ELF, I liked ELF Tool Chain (http://elftoolchain.sf.net, copy at svn://svn.riscos.info/gccsdk/trunk/tools/elftoolchain/ with autotools support, also used in asasm for its ELF/DWARF output).  For AOF/ALF I don't think there is a general purpose one but luckily those formats are very simple so you can probably very easily write one yourself for this project.
    3. The semantics : this is where things get interesting.  ELF is much richer in semantics than AOF so this will require some attention.
      • How ELF sections (.text*, .(ro)data*, .bss, perhaps others) are mapped to AOF areas.  I assume the idea is that the AOF files will be used by the Norcroft linker so the linking order what Norcroft linker adheres to (which I believe is fixed) should be more or less in line with the default linking script in binutils link.  I don't expect this being an issue but it is worthwhile to investigate this for possible showstoppers.
      • Symbol & string table : should not be that hard to convert.
      • Relocations is going to be interesting as well.  ELF ARM has lots of relocations which need to be mapped to a few combinations in AOF.  Probably you're lucky here because in ELF relocations are very instruction specific while in AOF it's the linker which needs to recognize & support the instructions to be able to apply the generic AOF relocation to them.  So it might just be a case to bulk mapping the ELF ones to AOF ones.
      • Converting the DWARF debug information to AOF compatible one is most probably impossible, perhaps maybe it it feasible to keep the line number information (so single stepping in DDE could be done).
    4. Testing : needs some thought, you might think you're generating a valid AOF file but Norcroft's link might think otherwise.  I guess you'll need some testqueue exercising meaningful combinations and see what Norcroft's link makes of it.  I did similar things for testing asasm AOF output and got some surprises but also insights.
    Speaking of asasm, it might help to look at its sources to get up to speed : internally its data structures are basically designed for AOF and for ELF we translate AOF semantics to ELF.  It might also be a source for inspiration for the relocation part or its DWARF line number support.

    Aside, elf2aif is a very simple hack : we reserve enough space at the beginning of our first executable segment upfront and basically zap in an AIF header in that space in the already linked ELF program file and strip away the segments we're not interested in.  Simple and very effective but it won't help you for ELF->AOF/ALF.


    [...] but there's a nagging worry that  even if I handle all of the symbol and relocation details correctly, the  actual structures of the compiled code and data might not be compatible. Any  files would be linked to the Shared C Library in the final instance, if that  makes a difference.

    I thinking you're hinting at possible ABI differences for both APCS-32 implementations.  Indeed, a very valid consideration to make.  gccsdk targets APCS-32 for SCL (-mlibscl option) so in principle you should be fine here except for one known issue : gccsdk differs from Norcroft C compiler when float arguments are passed to functions.  I can't remember all the details (when to upgrade float -> double ?) but the SCL stubs have a workaround for those SCL entries where float arguments are passed for this exact reason.

    I'm not aware of any other ABI differences (as far as I know the APCS ABI never has been specified with great detail on all aspects of calling convention, data layout composite types, etc.) but I wouldn't be surprised if you start mixing Norcroft AOF files with GCCSDK AOF files that you will find such issues.  I don't think that has been done alot before with the gccsdk 3.4.x (or earlier releases) which had AOF output.

    [...] So, before I waste too much time on this side project, is anyone already  able to say that it /won't/ work for whatever reason? If so, there seems  little point spending time reaching the same conclusion that other people  have already reached before me!

    As conclusion, I believe this is a very reasonable idea/project, not straightforward but feasible.

    I'm not wearing a GCCSDK hat anymore but actually if you would be doing such a project, it would be a great addition to the existing gccsdk toolset.

    Or, even better, is there already a solution out there that Google has  missed?  
    I'm not aware of any off-the-shelve solutions.  AOF/ALF has long time been left along by the non-RISC OS/Acorn community, everyone is using ELF (or COFF) these days.

    Hope this helps,
    John.

    Re: Fail 2890 json

    In article <mpro.nt3akk00k0cd600b9.pittdj@pittdj.co.uk>,
    David Pitt <pittdj@pittdj.co.uk> wrote:
    > "Gerald Dodson", on 13 Aug, wrote:

    > > Has any one else had this version fail?

    > Try #2911 or later.

    Well, http://www.bbc.co.uk/sport/0/athletics/33867962 now seems to be OK
    with #2916. Is that the answer you were looking for?

    Re: [gccsdk] New VFP/NEON capable GCC release?

    On 11/08/15 10:06, Theo Markettos wrote:

    [snip]

    > Is there anything missing from there, or any wrong versions I should know
    > about?

    I don't think so, apart from...

    > I bumped GCC to 4.7.4-Rel2, but is DRenderer still on
    > 0.56-beta8?

    Yes, but despite the archive on riscos.info being named
    DRenderer_0.56-beta8-1.zip, the module inside is 0.55, so I think
    it does need updating.

    > Do we want to make an 'official' release of it?

    I'm not really familiar with DRenderer, but if you just mean remove
    the beta tag then I don't see why not.

    Lee.


    _______________________________________________
    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] New VFP/NEON capable GCC release?

    On 13/08/15 18:31, Theo Markettos wrote:
    > On Tue, Aug 11, 2015 at 10:06:23AM +0100, Theo Markettos wrote:
    >> On Mon, Aug 10, 2015 at 04:11:42PM +0100, Lee Noar wrote:
    >>> I think this is done now. There wasn't much more to do after all.
    >>> Perhaps someone could double check that the archives extract to a fully
    >>> working installation. I have tried extracting to a new location and
    >>> building a simple C program with some FP in it both with and without
    >>> VFP and all seems well.
    >>
    >> Thanks. Some builds:
    >>
    >> http://ci.riscos.info/job/gcc-4.7-native/9/artifact/gcc4/release-area/riscpkg/
    >>
    >> I'll try and give them a spin when I get a moment.
    >
    > FWIW I'm unlikely to get a chance to test these for the next week or so,
    > possibly more. If anyone has the time to give them a quick spin, that would
    > be most helpful.

    When Jenkins has finished building with the changes I've just made,
    I'll give them a try and let you know.

    Lee.


    _______________________________________________
    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] Converting ELF into AOF or ALF?

    For reasons of convenience, it would useful for me to be able to generate
    libraries in ALF format alongside the ELF versions that I currently produce
    from GCCSDK. I can obviously do it via Norcroft, but that means munging the
    Linux-based source files into a structure that Norcroft will work with and
    keeping the two sets in step with each other.

    Following a similar discussion on the ROOL forums recently, I was wondering
    about a tool to convert the ELF object files into AOF (and then ALF) --
    similar in concept to elf2aif, I suppose.

    Having had a quick look at the two formats, it seems that it might be
    possible to convert one into the other, but there's a nagging worry that
    even if I handle all of the symbol and relocation details correctly, the
    actual structures of the compiled code and data might not be compatible. Any
    files would be linked to the Shared C Library in the final instance, if that
    makes a difference.

    So, before I waste too much time on this side project, is anyone already
    able to say that it /won't/ work for whatever reason? If so, there seems
    little point spending time reaching the same conclusion that other people
    have already reached before me!

    Or, even better, is there already a solution out there that Google has
    missed?

    --
    Steve Fryatt - Leeds, England

    http://www.stevefryatt.org.uk/

    _______________________________________________
    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