diff --git a/resources/FatMessages b/resources/FatMessages
index d8fa5f0..bed740e 100644
--- a/resources/FatMessages
+++ b/resources/FatMessages
@@ -3222,6 +3222,8 @@ fr.all.DirectoryError:répertoire '%s' existe déjÃ
it.all.DirectoryError:La directory '%s' è già esistente
nl.all.DirectoryError:map '%s' bestaat reeds
+en.ro.SprIsNull:Unable to convert image to sprite
+
# Error messages for Amiga version only
en.ami.CompError:Unable to open
de.ami.CompError:Nicht zu öffnen
diff --git a/riscos/bitmap.c b/riscos/bitmap.c
index 384895d..162a8c6 100644
--- a/riscos/bitmap.c
+++ b/riscos/bitmap.c
@@ -43,6 +43,7 @@
#include "utils/filename.h"
#include "utils/log.h"
#include "utils/utils.h"
+#include "utils/messages.h"
#include "desktop/plotters.h"
#include "content/content.h"
#include "image/bitmap.h"
@@ -272,6 +273,11 @@ bool riscos_bitmap_save(void *vbitmap, const char *path, unsigned flags)
struct bitmap *bitmap = (struct bitmap *) vbitmap;
os_error *error;
+ if (bitmap == NULL) {
+ warn_user("SaveError", messages_get("SprIsNull"));
+ return false;
+ }
+
if (!bitmap->sprite_area) {
riscos_bitmap_get_buffer(bitmap);
}
This gives a more user-friendly error message (sadly I can't offer any translations).
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, 31 October 2015
[gccsdk] patchset to build !otterbrowser on gcc5 on linux
Enclosed patch set encapsulates all I have needed to do to build otterbrowser straight out of the gccsdk site on my Ubunti 15.10 installation which uses gcc5:-
create a folder gccsdk in your linux home directory. (ie ~/gccsdk)
copy the zipcontents over it
in ~/gccsdk type ./doall
some hours later it'll complete (unless I have more stuff in my linux than is standard -- inwhich case you may need to apt-get - or whatever - the missing tool)
I'd appreciate it if anyone else could repeat this
Could I also suggest things get committed to the repository?
Thanks
John
create a folder gccsdk in your linux home directory. (ie ~/gccsdk)
copy the zipcontents over it
in ~/gccsdk type ./doall
some hours later it'll complete (unless I have more stuff in my linux than is standard -- inwhich case you may need to apt-get - or whatever - the missing tool)
I'd appreciate it if anyone else could repeat this
Could I also suggest things get committed to the repository?
Thanks
John
--
Stop press! Just announced -
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
For full details of see https://www.utilitywarehouse.co.uk/reviews?exref=095761
Intrigued? Call me
John Ballance C.Eng MIET - jwb@macpcrepair.co.uk - 07976 295923
Re: [gccsdk] compiler selfhosted.
On 10/30/2015 03:23 PM, Michael Grunditz wrote:
> Can we try to have GCCSDK build its own compilers (including
> hostcompilers)? In this way a lot of things will run smoother.
Can you explain why a lot of things would run smoother ?
I don't see any gain doing this.
John.
_______________________________________________
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
> Can we try to have GCCSDK build its own compilers (including
> hostcompilers)? In this way a lot of things will run smoother.
Can you explain why a lot of things would run smoother ?
I don't see any gain doing this.
John.
_______________________________________________
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
Friday, 30 October 2015
[gccsdk] Otterbrowser now built with the current Qt5 stuff in autobuilder and gcc5 on ubuntu15.10
AB_GIT=https://github.com/OtterBrowser/otter-browser.git
AB_PVERSION=0.9.8
AB_ROVERSION=2
AB_DISTCLEAN="$AB_MAKE -C build clean"
ab_make() {
mkdir build && cd build
cmake -DCMAKE_TOOLCHAIN_FILE=../../../../env/toolchain-riscos.cmake \
-DCMAKE_EXE_LINKER_FLAGS='-L../../../../env/lib -lWebKit1 -lWebCore -lleveldb -lmemenv -ljpeg -lpng -lJavaScriptCore -lWTF -lsqlite3 -licui18n -licuuc -licudata -lQt5WebKit -lssl -lcrypto -lpcre16 -lz' \
..
$AB_MAKE
cd ..
}
ab_package() {
ab_create_app Otter-browser Apps/Network otter-browser
cp -T $S/build/otter-browser $A/otter-browser$AB_EXEEXT
for f in \
CHANGELOG \
CONTRIBUTING.md COPYING INSTALL \
; do
cp -prL $S/$f $A
cp -prL $S/resources $A
done
ab_docs -h man/otter-browser.1
$AB_HOME/add-riscpkg -copyrightfile COPYING -package otter-browser -name Otter-browser \
-depends Qt5Webkit
}
Hi
Have now completed build of otterbrowser from sources taken from the autobuilder yesterday.
There is 1 patch needed to the otterbrowser autobuild setvars script.
line 10 needs -lpcre16 added :
-DCMAKE_EXE_LINKER_FLAGS='-L../../../../env/lib -lWebKit1 -lWebCore -lleveldb -lmemenv -ljpeg -lpng -lJavaScriptCore -lWTF -lsqlite3 -licui18n -licuuc -licudata -lQt5WebKit -lssl -lcrypto -lpcre16 -lz' \
Amended setvars attached.
This means that there are all the mods needed to compile and build otterbrowser, (and nettle and wget incidentally) on a gcc5 based machine using GCCSDK4.7.4
I'm not doing this to be a maintainer, but it bothers me when what is available wont build from the autobuilder directly.
I'll probably put up binaries of what has been produced in case anyone is interested, and would be happy to make a zip available of the various mods/patches I have employed to get to this stage.
(they have all been reported here recently)
I strongly suspect that they will be backward compatible with the possible exception of gcc.cp.cfns.gperf.p, which tunes the definitions passed through gperf to the definitions created by gperf. gcc5 would appear to be rather more fussy about these things.
Many of the changes involve a global source change to replace __FUNCTION__ with __func__ . There are too many individual files involved to easily provide separate patches.
Feedback appreciated
Thanks
John
--
Stop press! Just announced -
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
For full details of see https://www.utilitywarehouse.co.uk/reviews?exref=095761
Intrigued? Call me
John Ballance C.Eng MIET - jwb@macpcrepair.co.uk - 07976 295923
[gccsdk] compiler selfhosted.
Hi
Can we try to have GCCSDK build its own compilers (including
hostcompilers)? In this way a lot of things will run smoother.
Br
Michael
_______________________________________________
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
Can we try to have GCCSDK build its own compilers (including
hostcompilers)? In this way a lot of things will run smoother.
Br
Michael
_______________________________________________
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: Patch: Partial fix for bug 2368 (version 2)
On Mon, Oct 26, 2015 at 10:03:27PM +0000, Dave Higton wrote:
> This patch provides a partial fix for bug 2368.
In principle the change is good (aside from the opening curly brace
should be on the same line as the if). However the error message is
not useful to the user, it is accurate of course, but anything
translated is for the general user ought to be more descriptive.
May I suggest "Unable to convert image to sprite" would be more
helpful in this context as it converys to the user exactly what action
the browser was unable to perform.
Additionally Michael has suggested that the RISC OS idiomatic thing to
do would be to grey out the save to sprite option altogether when the
action is not valid. This would need to be in addition to the error
here as "on demand" conversion could potentialy mean the image data is
not available.
>
> The second image in the cited URL is CMY, and the libraries fail to
> convert it. The resulting bitmap is null. It is rendered as a
> pink rectangle. If the user tries to export it as a sprite,
> riscos_bitmap_save() is called, which immediately attempts a null
> dereference, causing a crash (when I tried it, it was a serious
> crash that brought the whole machine down).
snipped the rest
--
Regards Vincent
http://www.kyllikki.org/
> This patch provides a partial fix for bug 2368.
In principle the change is good (aside from the opening curly brace
should be on the same line as the if). However the error message is
not useful to the user, it is accurate of course, but anything
translated is for the general user ought to be more descriptive.
May I suggest "Unable to convert image to sprite" would be more
helpful in this context as it converys to the user exactly what action
the browser was unable to perform.
Additionally Michael has suggested that the RISC OS idiomatic thing to
do would be to grey out the save to sprite option altogether when the
action is not valid. This would need to be in addition to the error
here as "on demand" conversion could potentialy mean the image data is
not available.
>
> The second image in the cited URL is CMY, and the libraries fail to
> convert it. The resulting bitmap is null. It is rendered as a
> pink rectangle. If the user tries to export it as a sprite,
> riscos_bitmap_save() is called, which immediately attempts a null
> dereference, causing a crash (when I tried it, it was a serious
> crash that brought the whole machine down).
snipped the rest
--
Regards Vincent
http://www.kyllikki.org/
Thursday, 29 October 2015
Re: [gccsdk] cross-compiling using gcc5 on the host 1
In message <56315DB9.6040409@macpcrepair.co.uk>
John Ballance <jwb@macpcrepair.co.uk> wrote:
> Its come about as my machine got taken up to ubuntu15.10,
> which uses gcc5 as a default, and mu linux skills were not really up to
> to installing/using earlier gcc for this.
>
I cant speak for 15.10 but previous versions have allowed me to chose different versions of gcc.
It required 'sudo apt-get remove gcc' then you have to see what other versions are available with 'apt-cache search gcc' or maybe use synaptic.
You need to then install the version by its package name to over-ride the distro default.
sudo apt-get install gcc-4 or whatever the appropriate name is.
HTH Ron.
_______________________________________________
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 Ballance <jwb@macpcrepair.co.uk> wrote:
> Its come about as my machine got taken up to ubuntu15.10,
> which uses gcc5 as a default, and mu linux skills were not really up to
> to installing/using earlier gcc for this.
>
I cant speak for 15.10 but previous versions have allowed me to chose different versions of gcc.
It required 'sudo apt-get remove gcc' then you have to see what other versions are available with 'apt-cache search gcc' or maybe use synaptic.
You need to then install the version by its package name to over-ride the distro default.
sudo apt-get install gcc-4 or whatever the appropriate name is.
HTH Ron.
_______________________________________________
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] toolchain-riscos.cmake
Hi
if compiling on a linux box, the content of env/toolchain-riscos.cmake needs the .exe removed from the end of lines 7 and 8
this file starts life in the gcc compile directory. Would it be better for this to decide whether it is crosscompiling on windoze or linux, or just change it?
Thanks
John
if compiling on a linux box, the content of env/toolchain-riscos.cmake needs the .exe removed from the end of lines 7 and 8
this file starts life in the gcc compile directory. Would it be better for this to decide whether it is crosscompiling on windoze or linux, or just change it?
Thanks
John
--
Stop press! Just announced -
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
For full details of see https://www.utilitywarehouse.co.uk/reviews?exref=095761
Intrigued? Call me
John Ballance C.Eng MIET - jwb@macpcrepair.co.uk - 07976 295923
Re: [gccsdk] cross-compiling using gcc5 on the host 1
--- Makefile 2015-10-28 10:05:44.839970500 +0000
+++ Makefile.fixed 2015-10-28 10:07:54.000000000 +0000
@@ -220,6 +220,7 @@
-rm -rf $(BUILDDIR) $(BUILDSTEPSDIR) $(SRCDIR) $(GCCSDK_CROSS_PREFIX) $(GCCSDK_RISCOS_PREFIX)
-svn revert -R $(SRCORIGDIR)/gcc-trunk
-svn status $(SRCORIGDIR)/gcc-trunk | grep -E "^?" | cut -b 9- | xargs rm -rf
+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -not -path '^\.' -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
for i in . riscos/asasm riscos/asasm/decaof/ riscos/asasm/elftoolchain/ ; do svn status --no-ignore $$i | grep "^I " | cut -b 9- | grep -v -E "^(gccsdk-params|srcdir\.orig|release-area)$$" | xargs rm -rf ; done
# Return to a state for doing a full fresh cross build (using the current binutils/gcc sources).
@@ -242,6 +243,7 @@
-svn revert -R $(SRCORIGDIR)/gcc-trunk
-svn status $(SRCORIGDIR)/gcc-trunk | grep -E "^?" | cut -b 9- | xargs rm -rf
cd $(SRCORIGDIR)/gcc-trunk && ./contrib/gcc_update
+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -not -path '^\.' -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
else
@echo "Makefile target updategcc is only supported when GCC sources are fetched from the GCC source control system."
endif
@@ -542,6 +544,7 @@
-rm -rf $(SRCDIR)/gcc
-svn revert -R $(SRCORIGDIR)/gcc-trunk
-svn status $(SRCORIGDIR)/gcc-trunk | grep -E "^?" | cut -b 9- | xargs rm -rf
+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -not -path '^\.' -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
ln -s $(SRCORIGDIR)/gcc-trunk $(SRCDIR)/gcc
else
src-gcc-copied: $(SRCORIGDIR)/gcc-$(GCC_VERSION).tar.bz2
@@ -686,6 +689,7 @@
-mkdir -p $(SRCORIGDIR)
cd $(SRCORIGDIR) && svn co svn://gcc.gnu.org/svn/gcc/$(GCC_SCM_PATH) gcc-trunk
cd $(SRCORIGDIR)/gcc-trunk && ./contrib/gcc_update
+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -not -path '^\.' -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
else
# Download gcc source:
$(SRCORIGDIR)/gcc-$(GCC_VERSION).tar.bz2:
<snip>
I have done this with a patch to the gcc4/Makefile. This is a sledgehammer as the gcc build tree appears to use the old format in many places.
replacing the patch lines with
+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -not -path '^\.' -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
keeps out of the svn area.
modified patcher for Makefile enclosed
Jon
+++ Makefile.fixed 2015-10-28 10:07:54.000000000 +0000
@@ -220,6 +220,7 @@
-rm -rf $(BUILDDIR) $(BUILDSTEPSDIR) $(SRCDIR) $(GCCSDK_CROSS_PREFIX) $(GCCSDK_RISCOS_PREFIX)
-svn revert -R $(SRCORIGDIR)/gcc-trunk
-svn status $(SRCORIGDIR)/gcc-trunk | grep -E "^?" | cut -b 9- | xargs rm -rf
+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -not -path '^\.' -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
for i in . riscos/asasm riscos/asasm/decaof/ riscos/asasm/elftoolchain/ ; do svn status --no-ignore $$i | grep "^I " | cut -b 9- | grep -v -E "^(gccsdk-params|srcdir\.orig|release-area)$$" | xargs rm -rf ; done
# Return to a state for doing a full fresh cross build (using the current binutils/gcc sources).
@@ -242,6 +243,7 @@
-svn revert -R $(SRCORIGDIR)/gcc-trunk
-svn status $(SRCORIGDIR)/gcc-trunk | grep -E "^?" | cut -b 9- | xargs rm -rf
cd $(SRCORIGDIR)/gcc-trunk && ./contrib/gcc_update
+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -not -path '^\.' -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
else
@echo "Makefile target updategcc is only supported when GCC sources are fetched from the GCC source control system."
endif
@@ -542,6 +544,7 @@
-rm -rf $(SRCDIR)/gcc
-svn revert -R $(SRCORIGDIR)/gcc-trunk
-svn status $(SRCORIGDIR)/gcc-trunk | grep -E "^?" | cut -b 9- | xargs rm -rf
+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -not -path '^\.' -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
ln -s $(SRCORIGDIR)/gcc-trunk $(SRCDIR)/gcc
else
src-gcc-copied: $(SRCORIGDIR)/gcc-$(GCC_VERSION).tar.bz2
@@ -686,6 +689,7 @@
-mkdir -p $(SRCORIGDIR)
cd $(SRCORIGDIR) && svn co svn://gcc.gnu.org/svn/gcc/$(GCC_SCM_PATH) gcc-trunk
cd $(SRCORIGDIR)/gcc-trunk && ./contrib/gcc_update
+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -not -path '^\.' -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
else
# Download gcc source:
$(SRCORIGDIR)/gcc-$(GCC_VERSION).tar.bz2:
<snip>
I have done this with a patch to the gcc4/Makefile. This is a sledgehammer as the gcc build tree appears to use the old format in many places.
Yes, a real sledgehammer indeed as...
[...]+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
...this recursively goes into your gcc-trunk/.svn database. :-(
replacing the patch lines with
+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -not -path '^\.' -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
keeps out of the svn area.
modified patcher for Makefile enclosed
Jon
--
Stop press! Just announced -
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
For full details of see https://www.utilitywarehouse.co.uk/reviews?exref=095761
Intrigued? Call me
John Ballance C.Eng MIET - jwb@macpcrepair.co.uk - 07976 295923
[gccsdk] Qt5Webkit built with gcc5
Hi
Have built Qt5Webkit from current autobuilder sources. It runs the QtTestBrowser doing most of what is expected. Its not perfect, (ssl fails to complete) but as it builds, debugging is possible I guess
It needs 1 patch to Setvars, again the __FUNCTION__ issue
insert at line 28:
cd $S && find -regex ".*\.\(cpp\|h\)" -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
Cheers
JB
Have built Qt5Webkit from current autobuilder sources. It runs the QtTestBrowser doing most of what is expected. Its not perfect, (ssl fails to complete) but as it builds, debugging is possible I guess
It needs 1 patch to Setvars, again the __FUNCTION__ issue
insert at line 28:
cd $S && find -regex ".*\.\(cpp\|h\)" -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
Cheers
JB
--
Stop press! Just announced -
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
For full details of see https://www.utilitywarehouse.co.uk/reviews?exref=095761
Intrigued? Call me
John Ballance C.Eng MIET - jwb@macpcrepair.co.uk - 07976 295923
[gccsdk] Qt5Base built with gcc5
--- src/corelib/kernel/qfunctions_p.h 2015-06-29 21:04:57.000000000 +0100
+++ src/corelib/kernel/qfunctions_p.h.fixed 2015-10-23 12:54:32.528764202 +0100
@@ -57,5 +57,9 @@
# include "QtCore/qfunctions_winrt.h"
#endif
+#define Q_STATIC_GLOBAL_OPERATOR static
+#define Q_STATIC_GLOBAL_INLINE_OPERATOR static inline
+
+
#endif
OK. Have now built Qt5Base from autobuilder with current sources.
It generally runs its test code, so is of interest.
patches in autobuilder as follows:
1: add following line at setvars line45 . this deals with the __FUNCTION__ issue, again in sledgehammer mode
find -regex ".*\.\(c\|C\|h\)" -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
2: deal with an undefined variable:
--- src/corelib/kernel/qfunctions_p.h 2015-06-29 21:04:57.000000000 +0100
+++ src/corelib/kernel/qfunctions_p.h.fixed 2015-10-23 12:54:32.528764202 +0100
@@ -57,5 +57,9 @@
# include "QtCore/qfunctions_winrt.h"
+++ src/corelib/kernel/qfunctions_p.h.fixed 2015-10-23 12:54:32.528764202 +0100
@@ -57,5 +57,9 @@
# include "QtCore/qfunctions_winrt.h"
#endif
+#define Q_STATIC_GLOBAL_OPERATOR static
+#define Q_STATIC_GLOBAL_INLINE_OPERATOR static inline
+
+
#endif
OK. Have now built Qt5Base from autobuilder with current sources.
It generally runs its test code, so is of interest.
patches in autobuilder as follows:
1: add following line at setvars line45 . this deals with the __FUNCTION__ issue, again in sledgehammer mode
find -regex ".*\.\(c\|C\|h\)" -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
2: deal with an undefined variable:
--- src/corelib/kernel/qfunctions_p.h 2015-06-29 21:04:57.000000000 +0100
+++ src/corelib/kernel/qfunctions_p.h.fixed 2015-10-23 12:54:32.528764202 +0100
@@ -57,5 +57,9 @@
# include "QtCore/qfunctions_winrt.h"
[gccsdk] !SharedLibs
Hi
What is the process for putting libraries in !SharedLibs
is it as simple as copy from gccsdk/env/lib?
Do the libraries themselves have to be built specifically as shared libs?
Thanks
John
What is the process for putting libraries in !SharedLibs
is it as simple as copy from gccsdk/env/lib?
Do the libraries themselves have to be built specifically as shared libs?
Thanks
John
--
Stop press! Just announced -
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
For full details of see https://www.utilitywarehouse.co.uk/reviews?exref=095761
Intrigued? Call me
John Ballance C.Eng MIET - jwb@macpcrepair.co.uk - 07976 295923
Re: [Rpcemu] RPCEmu 0.8.13
In article <EB29E69154CB44BABA1ED34B2E069E40@Laptop2>,
Gerald Holdsworth <gerald@hollypops.co.uk> wrote:
> > A new version of RPCEmu is available, 0.8.13
> Nice, well done - but not for the Mac.
> Would it be worth putting the 0.8.12 Mac build on the 'contributed
> builds' on the page?
> https://bitbucket.org/caliston/rpcemu-spoon-caliston/downloads
> I'm using it with very little issue on the latest version of Mac OS X,
> and using RISC OS 5.23.
> And while we at it, can we get the RISC OS Compatibility list updated as
> it still says that 6.20 is not compatible?
> Cheers,
> Gerald.
I've been using it for years on RISC OS 6.20, however even after this new
version 0.8.13 it still will not action the Alt + Break process to quit an
errant app, it just causes the RPCEmu to freeze up requiring a Windows [X]
to make a dirty exit.
Dave
--
Dave Triffid
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
Gerald Holdsworth <gerald@hollypops.co.uk> wrote:
> > A new version of RPCEmu is available, 0.8.13
> Nice, well done - but not for the Mac.
> Would it be worth putting the 0.8.12 Mac build on the 'contributed
> builds' on the page?
> https://bitbucket.org/caliston/rpcemu-spoon-caliston/downloads
> I'm using it with very little issue on the latest version of Mac OS X,
> and using RISC OS 5.23.
> And while we at it, can we get the RISC OS Compatibility list updated as
> it still says that 6.20 is not compatible?
> Cheers,
> Gerald.
I've been using it for years on RISC OS 6.20, however even after this new
version 0.8.13 it still will not action the Alt + Break process to quit an
errant app, it just causes the RPCEmu to freeze up requiring a Windows [X]
to make a dirty exit.
Dave
--
Dave Triffid
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
Re: [gccsdk] cross-compiling using gcc5 on the host 1
OK
I understand what is going on a bit better, and it affects stuff built via the autobuilder as well as gcc.
I suspect what is needed is a global definition something like
#ifndef __FUNCTION__
#define __FUNCTION__ __func__
I understand what is going on a bit better, and it affects stuff built via the autobuilder as well as gcc.
I suspect what is needed is a global definition something like
#ifndef __FUNCTION__
#define __FUNCTION__ __func__
[gccsdk] compiling oslib on gcc5
--- Tools/DefMod2/vapi.c 2015-10-29 08:31:47.733103350 +0000
+++ Tools/DefMod2/vapi.c.fixed 2015-10-29 08:26:30.000000000 +0000
@@ -214,7 +214,7 @@
{
int r = 0;
int w = 0;
- bool nextcaps = TRUE;
+ osbool nextcaps = TRUE;
while (name[r] != '\0')
{
@@ -237,7 +237,7 @@
vapi_name[w] = '\0';
}
-os_error *get_vapi_type(def_t t, bool ref, char* type)
+os_error *get_vapi_type(def_t t, osbool ref, char* type)
{
os_error* error = NULL;
@@ -272,7 +272,7 @@
return error;
}
-bool emit_as_const(const char* type)
+osbool emit_as_const(const char* type)
{
return
strcmp(type, "Wimp_W") == 0 ||
@@ -280,7 +280,7 @@
strcmp(type, "Toolbox_Class") == 0;
}
-bool emit_as_enum(const def_t t)
+osbool emit_as_enum(const def_t t)
{
return
t->tag == def_TYPE_ID && !emit_as_const(t->data.id);
@@ -476,9 +476,9 @@
def_t t
)
{
- bool first = TRUE;
+ osbool first = TRUE;
char array[20];
- bool multidimensional = t->tag == def_TYPE_ROW && t->data.row.base->tag == def_TYPE_ROW;
+ osbool multidimensional = t->tag == def_TYPE_ROW && t->data.row.base->tag == def_TYPE_ROW;
array[0] = '\0';
while (t->tag == def_TYPE_ROW)
@@ -504,7 +504,7 @@
const char* type,
const char* prefix,
const char* c_prefix,
- const bool skip_last
+ const osbool skip_last
);
os_error* emit_struct_member(FILE* file,
@@ -512,7 +512,7 @@
const char* type,
const char* prefix,
const char* c_prefix,
- bool ellipsis
+ osbool ellipsis
)
{
os_error* error = NULL;
@@ -614,7 +614,7 @@
const char* type,
const char* prefix,
const char* c_prefix,
- const bool skip_last
+ const osbool skip_last
)
{
os_error* error = NULL;
@@ -622,7 +622,7 @@
for (i = 0; i < t->data AS list.count - skip_last ? 1 : 0; i++)
{
- bool ellipsis =
+ osbool ellipsis =
(i == t->data AS list.count - 1 &&
t->tag == def_TYPE_STRUCT &&
t->data AS list.ellipsis);
@@ -674,7 +674,7 @@
char vapi_struct_name [def_ID_LIMIT + 1];
char c_name [def_ID_LIMIT + 1];
char base [def_ID_LIMIT + 1] = "";
- bool class;
+ osbool class;
if ((error = convert_vapi_struct_name (vapi_struct_name, prefix, type)) != NULL) goto finish;
def_as_extern (c_name, type);
@@ -801,7 +801,7 @@
{
def_as_macro(c_name, type);
if ((error = convert_vapi_const_name (vapi_name, type)) != NULL) goto finish;
- bool ref = FALSE;
+ osbool ref = FALSE;
if (t->tag == def_TYPE_REF)
{
ref = TRUE;
Hi
When compiling oslib with host gcc5 I located one set of changes needed to build the depmod binary:
Most of the source files in Tools/DefMod2/ have had all occurrences of the type 'bool' replaced with type 'osbool'
vapi.c hasn't and now complains about 'bool' being unknown. Patching 'bool' to 'osbool' resolves this.
patch file vapi.c.p to do this is attached.
Thanks
John
+++ Tools/DefMod2/vapi.c.fixed 2015-10-29 08:26:30.000000000 +0000
@@ -214,7 +214,7 @@
{
int r = 0;
int w = 0;
- bool nextcaps = TRUE;
+ osbool nextcaps = TRUE;
while (name[r] != '\0')
{
@@ -237,7 +237,7 @@
vapi_name[w] = '\0';
}
-os_error *get_vapi_type(def_t t, bool ref, char* type)
+os_error *get_vapi_type(def_t t, osbool ref, char* type)
{
os_error* error = NULL;
@@ -272,7 +272,7 @@
return error;
}
-bool emit_as_const(const char* type)
+osbool emit_as_const(const char* type)
{
return
strcmp(type, "Wimp_W") == 0 ||
@@ -280,7 +280,7 @@
strcmp(type, "Toolbox_Class") == 0;
}
-bool emit_as_enum(const def_t t)
+osbool emit_as_enum(const def_t t)
{
return
t->tag == def_TYPE_ID && !emit_as_const(t->data.id);
@@ -476,9 +476,9 @@
def_t t
)
{
- bool first = TRUE;
+ osbool first = TRUE;
char array[20];
- bool multidimensional = t->tag == def_TYPE_ROW && t->data.row.base->tag == def_TYPE_ROW;
+ osbool multidimensional = t->tag == def_TYPE_ROW && t->data.row.base->tag == def_TYPE_ROW;
array[0] = '\0';
while (t->tag == def_TYPE_ROW)
@@ -504,7 +504,7 @@
const char* type,
const char* prefix,
const char* c_prefix,
- const bool skip_last
+ const osbool skip_last
);
os_error* emit_struct_member(FILE* file,
@@ -512,7 +512,7 @@
const char* type,
const char* prefix,
const char* c_prefix,
- bool ellipsis
+ osbool ellipsis
)
{
os_error* error = NULL;
@@ -614,7 +614,7 @@
const char* type,
const char* prefix,
const char* c_prefix,
- const bool skip_last
+ const osbool skip_last
)
{
os_error* error = NULL;
@@ -622,7 +622,7 @@
for (i = 0; i < t->data AS list.count - skip_last ? 1 : 0; i++)
{
- bool ellipsis =
+ osbool ellipsis =
(i == t->data AS list.count - 1 &&
t->tag == def_TYPE_STRUCT &&
t->data AS list.ellipsis);
@@ -674,7 +674,7 @@
char vapi_struct_name [def_ID_LIMIT + 1];
char c_name [def_ID_LIMIT + 1];
char base [def_ID_LIMIT + 1] = "";
- bool class;
+ osbool class;
if ((error = convert_vapi_struct_name (vapi_struct_name, prefix, type)) != NULL) goto finish;
def_as_extern (c_name, type);
@@ -801,7 +801,7 @@
{
def_as_macro(c_name, type);
if ((error = convert_vapi_const_name (vapi_name, type)) != NULL) goto finish;
- bool ref = FALSE;
+ osbool ref = FALSE;
if (t->tag == def_TYPE_REF)
{
ref = TRUE;
Hi
When compiling oslib with host gcc5 I located one set of changes needed to build the depmod binary:
Most of the source files in Tools/DefMod2/ have had all occurrences of the type 'bool' replaced with type 'osbool'
vapi.c hasn't and now complains about 'bool' being unknown. Patching 'bool' to 'osbool' resolves this.
patch file vapi.c.p to do this is attached.
Thanks
John
--
Stop press! Just announced -
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
For full details of see https://www.utilitywarehouse.co.uk/reviews?exref=095761
Intrigued? Call me
John Ballance C.Eng MIET - jwb@macpcrepair.co.uk - 07976 295923
Re: [gccsdk] oslib source url
On 28/10/2015 19:53, John Ballance wrote:
Hi
It may be my setup, but I had to change line 2 of oslib/setvars
Problem solved.. mea culpa ... incomplete duplicate subversion on the system
the https url is fine. no changes needed
Thanks all
John
from
AB_SVN=https://svn.code.sf.net/p/ro-oslib/code/trunk/!OSLib
to
AB_SVN=svn://svn.code.sf.net/p/ro-oslib/code/trunk/!OSLib
i.e. using svn:// instead of https://
Chris had suggested the mod had taken place, but this is a fresh checkout containing https://
Re: [gccsdk] cross-compiling using gcc5 on the host 2
John Ballance wrote on Wednesday, October 28, 2015 7:31 PM:
> There is also a mod required to the cfns.gperf file, which creates cfns.h,
> to make the output in the cfns.h file consistent .. this is in
> recipe/patches/gcc
> and is named gcc/cp/cfns.gperf
> Again, I suspect, but cannot currently prove, this this will also be back
> compatible
> Index: gcc/cp/cfns.gperf
> ===================================================================
> --- gcc/cp/cfns.gperf (revision 229481)
> +++ gcc/cp/cfns.gperf (working copy)
> @@ -18,11 +18,25 @@
> <http://www.gnu.org/licenses/>. */
> #ifdef __GNUC__
> __inline
> +#else
> +#ifdef __cplusplus
> +inline
> +#if defined __GNUC_STDC_INLINE__ || defined __GNUC_GNU_INLINE__
> +__attribute__ ((__gnu_inline__))
>
> There is also a mod required to the cfns.gperf file, which creates cfns.h,
> to make the output in the cfns.h file consistent .. this is in
> recipe/patches/gcc
> and is named gcc/cp/cfns.gperf
> Again, I suspect, but cannot currently prove, this this will also be back
> compatible
> Index: gcc/cp/cfns.gperf
> ===================================================================
> --- gcc/cp/cfns.gperf (revision 229481)
> +++ gcc/cp/cfns.gperf (working copy)
> @@ -18,11 +18,25 @@
> <http://www.gnu.org/licenses/>. */
> #ifdef __GNUC__
> __inline
> +#else
> +#ifdef __cplusplus
> +inline
> +#if defined __GNUC_STDC_INLINE__ || defined __GNUC_GNU_INLINE__
> +__attribute__ ((__gnu_inline__))
>
Re: Netsurf Ported to iPhone and iPad - Distribution Available
> To whom it may concern,
>
> I noticed you have a ports and availability section at
>
> http://www.netsurf-browser.org/about/
>
> and I wanted to advise that I have ported Netsurf to iOS for iPhone
> and iPad.
>
> It's available under the QMole desktop system for iOS and can be
> installed by users using Cydia.
>
> If you are curious, below is a screenshot
>
> Netsurf
>
> Netsurf can be installed onto an iPhone as shown below.
>
> Install
>
> If you are interested, there is a Facebook page at
>
> https://web.facebook.com/qmolelinux
Shiny. If you've had to make any modifications to the sources to do
this, we'd be interested in reviewing the patches for possible
inclusion. Also, please note that the S in NetSurf is capitalised :)
B.
Wednesday, 28 October 2015
Re: [gccsdk] cross-compiling using gcc5 on the host 1
Thanks John
specific message no longer available, but what you say above seems familiar. Its come about as my machine got taken up to ubuntu15.10, which uses gcc5 as a default, and mu linux skills were not really up to to installing/using earlier gcc for this.
yes.. I guess one ought to exclude anything .svn etc. I'd greatly appreciate ' the way ' to do this.. or ideally avoid the need. I gave up trying to patch individual files as they were legion.
On 28/10/2015 22:03, John Tytgat wrote:
On 10/28/2015 08:26 PM, John Ballance wrote:
There are 2 things I have had to change to build the gccsdk on a gcc5 host
First deals with __FUNCTION__ nolonger being recognised, so a global replace of __FUNCTION__ with __func__ is needed.
Do you still have the exact error message ? __FUNCTION__ and __func__ are both still supported as far as I know. The only case I can think of which can cause an error is with a -Wpedantic -Werror combination as that results in "ISO C does not support '__FUNCTION__' predefined identifier".
specific message no longer available, but what you say above seems familiar. Its come about as my machine got taken up to ubuntu15.10, which uses gcc5 as a default, and mu linux skills were not really up to to installing/using earlier gcc for this.
I have done this with a patch to the gcc4/Makefile. This is a sledgehammer as the gcc build tree appears to use the old format in many places.
Yes, a real sledgehammer indeed as...
[...]+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
...this recursively goes into your gcc-trunk/.svn database. :-(
yes.. I guess one ought to exclude anything .svn etc. I'd greatly appreciate ' the way ' to do this.. or ideally avoid the need. I gave up trying to patch individual files as they were legion.
John.
_______________________________________________ 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
--
Stop press! Just announced -
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
For full details of see https://www.utilitywarehouse.co.uk/reviews?exref=095761
Intrigued? Call me
John Ballance C.Eng MIET - jwb@macpcrepair.co.uk - 07976 295923
Re: [gccsdk] cross-compiling using gcc5 on the host 1
On 10/28/2015 08:26 PM, John Ballance wrote:
There are 2 things I have had to change to build the gccsdk on a gcc5 host
First deals with __FUNCTION__ nolonger being recognised, so a global replace of __FUNCTION__ with __func__ is needed.
Do you still have the exact error message ? __FUNCTION__ and __func__ are both still supported as far as I know. The only case I can think of which can cause an error is with a -Wpedantic -Werror combination as that results in "ISO C does not support '__FUNCTION__' predefined identifier".
I have done this with a patch to the gcc4/Makefile. This is a sledgehammer as the gcc build tree appears to use the old format in many places.
Yes, a real sledgehammer indeed as...
[...]
+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \;
...this recursively goes into your gcc-trunk/.svn database. :-(
John.
[gccsdk] oslib source url
Hi
It may be my setup, but I had to change line 2 of oslib/setvars
from
AB_SVN=https://svn.code.sf.net/p/ro-oslib/code/trunk/!OSLib
to
AB_SVN=svn://svn.code.sf.net/p/ro-oslib/code/trunk/!OSLib
i.e. using svn:// instead of https://
Chris had suggested the mod had taken place, but this is a fresh checkout containing https://
Thanks
John
It may be my setup, but I had to change line 2 of oslib/setvars
from
AB_SVN=https://svn.code.sf.net/p/ro-oslib/code/trunk/!OSLib
to
AB_SVN=svn://svn.code.sf.net/p/ro-oslib/code/trunk/!OSLib
i.e. using svn:// instead of https://
Chris had suggested the mod had taken place, but this is a fresh checkout containing https://
Thanks
John
--
Stop press! Just announced -
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
For full details of see https://www.utilitywarehouse.co.uk/reviews?exref=095761
Intrigued? Call me
John Ballance C.Eng MIET - jwb@macpcrepair.co.uk - 07976 295923
[gccsdk] libgcrypt20 on gcc5
Again the __FUNCTION__ issue crops up
patch to fix ( into autobuilder.libraries/libgcrypt20):
--- src/g10lib.h 2015-09-07 13:05:57.000000000 +0100
+++ src/g10lib.h.fixed 2015-10-28 17:48:55.000000000 +0000
@@ -162,9 +162,9 @@
int _gcry_log_verbosity( int level );
#ifdef JNLIB_GCC_M_FUNCTION
-#define BUG() _gcry_bug( __FILE__ , __LINE__, __FUNCTION__ )
+#define BUG() _gcry_bug( __FILE__ , __LINE__, __func__ )
#define gcry_assert(expr) ((expr)? (void)0 \
- : _gcry_assert_failed (STR(expr), __FILE__, __LINE__, __FUNCTION__))
+ : _gcry_assert_failed (STR(expr), __FILE__, __LINE__, __func__))
#elif __STDC_VERSION__ >= 199901L
#define BUG() _gcry_bug( __FILE__ , __LINE__, __func__ )
#define gcry_assert(expr) ((expr)? (void)0 \
@@ -404,9 +404,9 @@
const char *description);
#ifdef JNLIB_GCC_M_FUNCTION
# define fips_signal_error(a) \
- _gcry_fips_signal_error (__FILE__, __LINE__, __FUNCTION__, 0, (a))
+ _gcry_fips_signal_error (__FILE__, __LINE__, __func__, 0, (a))
# define fips_signal_fatal_error(a) \
- _gcry_fips_signal_error (__FILE__, __LINE__, __FUNCTION__, 1, (a))
+ _gcry_fips_signal_error (__FILE__, __LINE__, __func__, 1, (a))
#else
# define fips_signal_error(a) \
_gcry_fips_signal_error (__FILE__, __LINE__, NULL, 0, (a))
--
Stop press! Just announced -
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
For full details of see https://www.utilitywarehouse.co.uk/reviews?exref=095761
Intrigued? Call me
John Ballance C.Eng MIET - jwb@macpcrepair.co.uk - 07976 295923
[gccsdk] cross-compiling using gcc5 on the host 2
There is also a mod required to the cfns.gperf file, which creates cfns.h, to make the output in the cfns.h file consistent .. this is in recipe/patches/gcc and is named gcc/cp/cfns.gperf
Again, I suspect, but cannot currently prove, this this will also be back compatible
Index: gcc/cp/cfns.gperf
===================================================================
--- gcc/cp/cfns.gperf (revision 229481)
+++ gcc/cp/cfns.gperf (working copy)
@@ -18,11 +18,25 @@
<http://www.gnu.org/licenses/>. */
#ifdef __GNUC__
__inline
+#else
+#ifdef __cplusplus
+inline
+#if defined __GNUC_STDC_INLINE__ || defined __GNUC_GNU_INLINE__
+__attribute__ ((__gnu_inline__))
Again, I suspect, but cannot currently prove, this this will also be back compatible
Index: gcc/cp/cfns.gperf
===================================================================
--- gcc/cp/cfns.gperf (revision 229481)
+++ gcc/cp/cfns.gperf (working copy)
@@ -18,11 +18,25 @@
<http://www.gnu.org/licenses/>. */
#ifdef __GNUC__
__inline
+#else
+#ifdef __cplusplus
+inline
+#if defined __GNUC_STDC_INLINE__ || defined __GNUC_GNU_INLINE__
+__attribute__ ((__gnu_inline__))
[gccsdk] cross-compiling using gcc5 on the host 1
Hi
There are 2 things I have had to change to build the gccsdk on a gcc5 host
First deals with __FUNCTION__ nolonger being recognised, so a global replace of __FUNCTION__ with __func__ is needed.
I have done this with a patch to the gcc4/Makefile. This is a sledgehammer as the gcc build tree appears to use the old format in many places.
I'm sure there is a better way, but this is needed. As far as I understand things, this should be OK for gcc4.7 and later (at least) too
There are 2 things I have had to change to build the gccsdk on a gcc5 host
First deals with __FUNCTION__ nolonger being recognised, so a global replace of __FUNCTION__ with __func__ is needed.
I have done this with a patch to the gcc4/Makefile. This is a sledgehammer as the gcc build tree appears to use the old format in many places.
I'm sure there is a better way, but this is needed. As far as I understand things, this should be OK for gcc4.7 and later (at least) too
--- Makefile 2015-10-28 10:05:44.839970500 +0000
+++ Makefile.fixed 2015-10-28 10:07:54.000000000 +0000
@@ -220,6 +220,7 @@
-rm -rf $(BUILDDIR) $(BUILDSTEPSDIR) $(SRCDIR) $(GCCSDK_CROSS_PREFIX) $(GCCSDK_RISCOS_PREFIX)
-svn revert -R $(SRCORIGDIR)/gcc-trunk
-svn status $(SRCORIGDIR)/gcc-trunk | grep -E "^?" | cut -b 9- | xargs rm -rf
+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \; for i in . riscos/asasm riscos/asasm/decaof/ riscos/asasm/elftoolchain/ ; do svn status --no-ignore $$i | grep "^I " | cut -b 9- | grep -v -E "^(gccsdk-params|srcdir\.orig|release-area)$$" | xargs rm -rf ; done
# Return to a state for doing a full fresh cross build (using the current binutils/gcc sources).
@@ -242,6 +243,7 @@
-svn revert -R $(SRCORIGDIR)/gcc-trunk
-svn status $(SRCORIGDIR)/gcc-trunk | grep -E "^?" | cut -b 9- | xargs rm -rf
cd $(SRCORIGDIR)/gcc-trunk && ./contrib/gcc_update
+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \; else
@echo "Makefile target updategcc is only supported when GCC sources are fetched from the GCC source control system."
endif
@@ -542,6 +544,7 @@
-rm -rf $(SRCDIR)/gcc
-svn revert -R $(SRCORIGDIR)/gcc-trunk
-svn status $(SRCORIGDIR)/gcc-trunk | grep -E "^?" | cut -b 9- | xargs rm -rf
+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \; ln -s $(SRCORIGDIR)/gcc-trunk $(SRCDIR)/gcc
else
src-gcc-copied: $(SRCORIGDIR)/gcc-$(GCC_VERSION).tar.bz2
@@ -686,6 +689,7 @@
-mkdir -p $(SRCORIGDIR)
cd $(SRCORIGDIR) && svn co svn://gcc.gnu.org/svn/gcc/$(GCC_SCM_PATH) gcc-trunk
cd $(SRCORIGDIR)/gcc-trunk && ./contrib/gcc_update
+ cd $(SRCORIGDIR)/gcc-trunk && find -regex ".*\.\(c\|C\|h\)" -type f -exec sed -i 's/__FUNCTION__/__func__/g' {} \; else
# Download gcc source:
$(SRCORIGDIR)/gcc-$(GCC_VERSION).tar.bz2:
--
Stop press! Just announced -
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
For full details of see https://www.utilitywarehouse.co.uk/reviews?exref=095761
Intrigued? Call me
John Ballance C.Eng MIET - jwb@macpcrepair.co.uk - 07976 295923
Tuesday, 27 October 2015
Re: Job spec of a port maintainer
On Tue, Oct 27, 2015 at 09:42:39PM +0000, Dave Higton wrote:
> In message <55162516f4webpages@sprow.co.uk>
> Sprow <webpages@sprow.co.uk> wrote:
> > http://bugs.netsurf-browser.org/mantis/view.php?id=2336
> > http://bugs.netsurf-browser.org/mantis/view.php?id=2289
> >
> >but they presumably need reviewing. I know from a day job in engineering not
> >to review my own code!
>
> Re. bugs 2336 and 2289: I'm not competent to review either of them.
> They are beyond my experience.
>
> So, devs: what to do? Would it be right for me to test them instead
> of reviewing them (I don't know if I could even be sure to test bug
> 2336 exhaustively), do we wave it through, do we put up a special
> build and ask the wider audience to test it, do we wave it through
> and specifically ask on the NS users ML for people to test these
> aspects of it; or what?
I see nothing obviously wrong with them and I am happy for them to be
merged for the wider masses to experiment with them.
B.
> In message <55162516f4webpages@sprow.co.uk>
> Sprow <webpages@sprow.co.uk> wrote:
> > http://bugs.netsurf-browser.org/mantis/view.php?id=2336
> > http://bugs.netsurf-browser.org/mantis/view.php?id=2289
> >
> >but they presumably need reviewing. I know from a day job in engineering not
> >to review my own code!
>
> Re. bugs 2336 and 2289: I'm not competent to review either of them.
> They are beyond my experience.
>
> So, devs: what to do? Would it be right for me to test them instead
> of reviewing them (I don't know if I could even be sure to test bug
> 2336 exhaustively), do we wave it through, do we put up a special
> build and ask the wider audience to test it, do we wave it through
> and specifically ask on the NS users ML for people to test these
> aspects of it; or what?
I see nothing obviously wrong with them and I am happy for them to be
merged for the wider masses to experiment with them.
B.
[gccsdk] GCCSDK474 build on ubuntu1510 (gcc5)
Any thoughts on how to proceed?
(with hindsight- in error) my linux box has been updated to ubuntu1510 which comes with gcc5. The GCCSDK474 fails to build for a number of reasons.
The first biggie, which arguably should have been dealt with a long time ago in core gcc, is that there has been a lot of use of the attribute __FUNCTION__, which was deprecated a good while ago I believe in favour of __func__. An appropriate modification to the gcc4/Makefile will do a 'sed' to resolve this..
Problems then occur in gcc-trunk/gcc/cp folder where gperf is used to generate the cfns.h from cfns.gperf.
It produces a file that is inconsistent in its __inline attributes. gperf is GNU 3.0.4.
Sorting this will be needed I guess, and any thoughts appreciated, would I be better to oblige ubuntu1510 to run with an earlier gcc, and if so, which should be best?
Many thanks
John
(with hindsight- in error) my linux box has been updated to ubuntu1510 which comes with gcc5. The GCCSDK474 fails to build for a number of reasons.
The first biggie, which arguably should have been dealt with a long time ago in core gcc, is that there has been a lot of use of the attribute __FUNCTION__, which was deprecated a good while ago I believe in favour of __func__. An appropriate modification to the gcc4/Makefile will do a 'sed' to resolve this..
Problems then occur in gcc-trunk/gcc/cp folder where gperf is used to generate the cfns.h from cfns.gperf.
It produces a file that is inconsistent in its __inline attributes. gperf is GNU 3.0.4.
Sorting this will be needed I guess, and any thoughts appreciated, would I be better to oblige ubuntu1510 to run with an earlier gcc, and if so, which should be best?
Many thanks
John
--
Stop press! Just announced -
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
For full details of see https://www.utilitywarehouse.co.uk/reviews?exref=095761
Intrigued? Call me
John Ballance C.Eng MIET - jwb@macpcrepair.co.uk - 07976 295923
Re: Job spec of a port maintainer
In article <ee3d521955.DaveMeUK@my.inbox.com>,
Dave Higton <dave@davehigton.me.uk> wrote:
> In message <55162516f4webpages@sprow.co.uk>
> Sprow <webpages@sprow.co.uk> wrote:
> > I went ahead and wrote patches for
> >
> > http://bugs.netsurf-browser.org/mantis/view.php?id=2266
>
> Bug 2266: I'm happy with the GUI patch. As for the language patch:
> Wikipedia suggests that two letters are not always enough for the
> language abbreviation. Sometimes three are necessary.
I think I'd assumed ISO3166-1 alpha 2 abbreviations were being used, but
bumping the buffer up to 8 (since it's only a stack variable) seems an
acceptable belt loosening!
Thanks for looking at them,
Sprow.
Dave Higton <dave@davehigton.me.uk> wrote:
> In message <55162516f4webpages@sprow.co.uk>
> Sprow <webpages@sprow.co.uk> wrote:
> > I went ahead and wrote patches for
> >
> > http://bugs.netsurf-browser.org/mantis/view.php?id=2266
>
> Bug 2266: I'm happy with the GUI patch. As for the language patch:
> Wikipedia suggests that two letters are not always enough for the
> language abbreviation. Sometimes three are necessary.
I think I'd assumed ISO3166-1 alpha 2 abbreviations were being used, but
bumping the buffer up to 8 (since it's only a stack variable) seems an
acceptable belt loosening!
Thanks for looking at them,
Sprow.
Re: Job spec of a port maintainer
In message <55162516f4webpages@sprow.co.uk>
Sprow <webpages@sprow.co.uk> wrote:
>After some invaluable hand holding from Rob Kendrick (I failed to locate the
>instructions to set up a VM to build for RISC OS) I went ahead and wrote
>patches for
>
> http://bugs.netsurf-browser.org/mantis/view.php?id=2266
> http://bugs.netsurf-browser.org/mantis/view.php?id=2170
> http://bugs.netsurf-browser.org/mantis/view.php?id=2336
> http://bugs.netsurf-browser.org/mantis/view.php?id=2289
>
>but they presumably need reviewing. I know from a day job in engineering not
>to review my own code!
Re. bugs 2336 and 2289: I'm not competent to review either of them.
They are beyond my experience.
So, devs: what to do? Would it be right for me to test them instead
of reviewing them (I don't know if I could even be sure to test bug
2336 exhaustively), do we wave it through, do we put up a special
build and ask the wider audience to test it, do we wave it through
and specifically ask on the NS users ML for people to test these
aspects of it; or what?
Dave
____________________________________________________________
FREE ONLINE PHOTOSHARING - Share your photos online with your friends and family!
Visit http://www.inbox.com/photosharing to find out more!
Sprow <webpages@sprow.co.uk> wrote:
>After some invaluable hand holding from Rob Kendrick (I failed to locate the
>instructions to set up a VM to build for RISC OS) I went ahead and wrote
>patches for
>
> http://bugs.netsurf-browser.org/mantis/view.php?id=2266
> http://bugs.netsurf-browser.org/mantis/view.php?id=2170
> http://bugs.netsurf-browser.org/mantis/view.php?id=2336
> http://bugs.netsurf-browser.org/mantis/view.php?id=2289
>
>but they presumably need reviewing. I know from a day job in engineering not
>to review my own code!
Re. bugs 2336 and 2289: I'm not competent to review either of them.
They are beyond my experience.
So, devs: what to do? Would it be right for me to test them instead
of reviewing them (I don't know if I could even be sure to test bug
2336 exhaustively), do we wave it through, do we put up a special
build and ask the wider audience to test it, do we wave it through
and specifically ask on the NS users ML for people to test these
aspects of it; or what?
Dave
____________________________________________________________
FREE ONLINE PHOTOSHARING - Share your photos online with your friends and family!
Visit http://www.inbox.com/photosharing to find out more!
Re: Job spec of a port maintainer
In message <55162516f4webpages@sprow.co.uk>
Sprow <webpages@sprow.co.uk> wrote:
>After some invaluable hand holding from Rob Kendrick (I failed to locate the
>instructions to set up a VM to build for RISC OS) I went ahead and wrote
>patches for
>
> http://bugs.netsurf-browser.org/mantis/view.php?id=2266
> http://bugs.netsurf-browser.org/mantis/view.php?id=2170
> http://bugs.netsurf-browser.org/mantis/view.php?id=2336
> http://bugs.netsurf-browser.org/mantis/view.php?id=2289
>
>but they presumably need reviewing. I know from a day job in engineering not
>to review my own code!
Re. bug 2170: I've not tested it but it looks plausible, so I'm happy.
Dave
____________________________________________________________
Share photos & screenshots in seconds...
TRY FREE IM TOOLPACK at http://www.imtoolpack.com/default.aspx?rc=if1
Works in all emails, instant messengers, blogs, forums and social networks.
Sprow <webpages@sprow.co.uk> wrote:
>After some invaluable hand holding from Rob Kendrick (I failed to locate the
>instructions to set up a VM to build for RISC OS) I went ahead and wrote
>patches for
>
> http://bugs.netsurf-browser.org/mantis/view.php?id=2266
> http://bugs.netsurf-browser.org/mantis/view.php?id=2170
> http://bugs.netsurf-browser.org/mantis/view.php?id=2336
> http://bugs.netsurf-browser.org/mantis/view.php?id=2289
>
>but they presumably need reviewing. I know from a day job in engineering not
>to review my own code!
Re. bug 2170: I've not tested it but it looks plausible, so I'm happy.
Dave
____________________________________________________________
Share photos & screenshots in seconds...
TRY FREE IM TOOLPACK at http://www.imtoolpack.com/default.aspx?rc=if1
Works in all emails, instant messengers, blogs, forums and social networks.
Re: Job spec of a port maintainer
In message <55162516f4webpages@sprow.co.uk>
Sprow <webpages@sprow.co.uk> wrote:
>After some invaluable hand holding from Rob Kendrick (I failed to locate the
>instructions to set up a VM to build for RISC OS) I went ahead and wrote
>patches for
>
> http://bugs.netsurf-browser.org/mantis/view.php?id=2266
> http://bugs.netsurf-browser.org/mantis/view.php?id=2170
> http://bugs.netsurf-browser.org/mantis/view.php?id=2336
> http://bugs.netsurf-browser.org/mantis/view.php?id=2289
>
>but they presumably need reviewing. I know from a day job in engineering not
>to review my own code!
Please forgive my tardiness.
Bug 2266: I'm happy with the GUI patch. As for the language patch:
Wikipedia suggests that two letters are not always enough for the
language abbreviation. Sometimes three are necessary. Along with
the slash and a terminator, I'd suggest that a 4 character buffer
for langdir is not entirely future-proof - and the extra cost of
even an 8 character buffer is minimal - so how about a bigger
reservation? In other respects I'm happy.
Dave
____________________________________________________________
Receive Notifications of Incoming Messages
Easily monitor multiple email accounts & access them with a click.
Visit http://www.inbox.com/notifier and check it out!
Sprow <webpages@sprow.co.uk> wrote:
>After some invaluable hand holding from Rob Kendrick (I failed to locate the
>instructions to set up a VM to build for RISC OS) I went ahead and wrote
>patches for
>
> http://bugs.netsurf-browser.org/mantis/view.php?id=2266
> http://bugs.netsurf-browser.org/mantis/view.php?id=2170
> http://bugs.netsurf-browser.org/mantis/view.php?id=2336
> http://bugs.netsurf-browser.org/mantis/view.php?id=2289
>
>but they presumably need reviewing. I know from a day job in engineering not
>to review my own code!
Please forgive my tardiness.
Bug 2266: I'm happy with the GUI patch. As for the language patch:
Wikipedia suggests that two letters are not always enough for the
language abbreviation. Sometimes three are necessary. Along with
the slash and a terminator, I'd suggest that a 4 character buffer
for langdir is not entirely future-proof - and the extra cost of
even an 8 character buffer is minimal - so how about a bigger
reservation? In other respects I'm happy.
Dave
____________________________________________________________
Receive Notifications of Incoming Messages
Easily monitor multiple email accounts & access them with a click.
Visit http://www.inbox.com/notifier and check it out!
Monday, 26 October 2015
Patch: Partial fix for bug 2368 (version 2)
This patch provides a partial fix for bug 2368.
The second image in the cited URL is CMY, and the libraries fail to
convert it. The resulting bitmap is null. It is rendered as a
pink rectangle. If the user tries to export it as a sprite,
riscos_bitmap_save() is called, which immediately attempts a null
dereference, causing a crash (when I tried it, it was a serious
crash that brought the whole machine down).
The simple fix is to test for a null bitmap and, if so, warn the
user and go no further.
This version has support for internationalisation of the error
message, and I've supplied messages in the three languages I can
muster.
Longer term of course we should enhance the libraries so that the
conversion does not fail.
Dave
diff --git a/resources/FatMessages b/resources/FatMessages
index e0e1682..b4aa18b 100644
--- a/resources/FatMessages
+++ b/resources/FatMessages
@@ -3217,6 +3217,10 @@ fr.all.DirectoryError:répertoire '%s' existe déjÃ
it.all.DirectoryError:La directory '%s' è già esistente
nl.all.DirectoryError:map '%s' bestaat reeds
+en.ro.ObjIsNull:Object is null
+fr.ro.ObjIsNull:L'objet est nul
+de.ro.ObjIsNull:Objekt ist null
+
# Error messages for Amiga version only
en.ami.CompError:Unable to open
de.ami.CompError:Nicht zu öffnen
diff --git a/riscos/bitmap.c b/riscos/bitmap.c
index 384895d..29665f9 100644
--- a/riscos/bitmap.c
+++ b/riscos/bitmap.c
@@ -43,6 +43,7 @@
#include "utils/filename.h"
#include "utils/log.h"
#include "utils/utils.h"
+#include "utils/messages.h"
#include "desktop/plotters.h"
#include "content/content.h"
#include "image/bitmap.h"
@@ -272,6 +273,11 @@ bool riscos_bitmap_save(void *vbitmap, const char *path, unsigned flags)
struct bitmap *bitmap = (struct bitmap *) vbitmap;
os_error *error;
+ if (bitmap == NULL)
+ {
+ warn_user("SaveError", messages_get("ObjIsNull"));
+ return false;
+ }
if (!bitmap->sprite_area) {
riscos_bitmap_get_buffer(bitmap);
}
____________________________________________________________
FREE ONLINE PHOTOSHARING - Share your photos online with your friends and family!
Visit http://www.inbox.com/photosharing to find out more!
The second image in the cited URL is CMY, and the libraries fail to
convert it. The resulting bitmap is null. It is rendered as a
pink rectangle. If the user tries to export it as a sprite,
riscos_bitmap_save() is called, which immediately attempts a null
dereference, causing a crash (when I tried it, it was a serious
crash that brought the whole machine down).
The simple fix is to test for a null bitmap and, if so, warn the
user and go no further.
This version has support for internationalisation of the error
message, and I've supplied messages in the three languages I can
muster.
Longer term of course we should enhance the libraries so that the
conversion does not fail.
Dave
diff --git a/resources/FatMessages b/resources/FatMessages
index e0e1682..b4aa18b 100644
--- a/resources/FatMessages
+++ b/resources/FatMessages
@@ -3217,6 +3217,10 @@ fr.all.DirectoryError:répertoire '%s' existe déjÃ
it.all.DirectoryError:La directory '%s' è già esistente
nl.all.DirectoryError:map '%s' bestaat reeds
+en.ro.ObjIsNull:Object is null
+fr.ro.ObjIsNull:L'objet est nul
+de.ro.ObjIsNull:Objekt ist null
+
# Error messages for Amiga version only
en.ami.CompError:Unable to open
de.ami.CompError:Nicht zu öffnen
diff --git a/riscos/bitmap.c b/riscos/bitmap.c
index 384895d..29665f9 100644
--- a/riscos/bitmap.c
+++ b/riscos/bitmap.c
@@ -43,6 +43,7 @@
#include "utils/filename.h"
#include "utils/log.h"
#include "utils/utils.h"
+#include "utils/messages.h"
#include "desktop/plotters.h"
#include "content/content.h"
#include "image/bitmap.h"
@@ -272,6 +273,11 @@ bool riscos_bitmap_save(void *vbitmap, const char *path, unsigned flags)
struct bitmap *bitmap = (struct bitmap *) vbitmap;
os_error *error;
+ if (bitmap == NULL)
+ {
+ warn_user("SaveError", messages_get("ObjIsNull"));
+ return false;
+ }
if (!bitmap->sprite_area) {
riscos_bitmap_get_buffer(bitmap);
}
____________________________________________________________
FREE ONLINE PHOTOSHARING - Share your photos online with your friends and family!
Visit http://www.inbox.com/photosharing to find out more!
Patch: Partial fix for bug 2368 (version 2)
This patch provides a partial fix for bug 2368.
The second image in the cited URL is CMY, and the libraries fail to
convert it. The resulting bitmap is null. It is rendered as a
pink rectangle. If the user tries to export it as a sprite,
riscos_bitmap_save() is called, which immediately attempts a null
dereference, causing a crash (when I tried it, it was a serious
crash that brought the whole machine down).
The simple fix is to test for a null bitmap and, if so, warn the
user and go no further.
This version has support for internationalisation of the error
message, and I've supplied messages in the three languages I can
muster.
Longer term of course we should enhance the libraries so that the
conversion does not fail.
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
The second image in the cited URL is CMY, and the libraries fail to
convert it. The resulting bitmap is null. It is rendered as a
pink rectangle. If the user tries to export it as a sprite,
riscos_bitmap_save() is called, which immediately attempts a null
dereference, causing a crash (when I tried it, it was a serious
crash that brought the whole machine down).
The simple fix is to test for a null bitmap and, if so, warn the
user and go no further.
This version has support for internationalisation of the error
message, and I've supplied messages in the three languages I can
muster.
Longer term of course we should enhance the libraries so that the
conversion does not fail.
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
Re: [Rpcemu] RPCEmu 0.8.13
On 23 Oct 2015, at 21:37, Theo Markettos <theo@markettos.org.uk> wrote:
>
> On Fri, Oct 23, 2015 at 11:49:14AM +0100, Theo Markettos wrote:
>> On Fri, Oct 23, 2015 at 10:17:07AM +0100, Gerald Holdsworth wrote:
>>>> A new version of RPCEmu is available, 0.8.13
>>>
>>> Nice, well done - but not for the Mac.
>>>
>>> Would it be worth putting the 0.8.12 Mac build on the 'contributed
>>> builds' on the page?
>>> https://bitbucket.org/caliston/rpcemu-spoon-caliston/downloads
>>
>> Let me do a test version, and if it hasn't broken anything that can be a
>> 'proper' 0.8.13.
>
> I have now built 0.8.13, that was straightforward. It's now on the above
> download page. It has received very little testing - it boots RISC OS 5 and
> seems to work for me. I believe it's still subject to the usual UI bugs -
> in particular there's no change on (lack of) Retina support.
Thanks for all your hard work getting RPCEmu for OS X going again, it's good to see a more recent build being widely available instead of just my really old ones!
Francis
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
>
> On Fri, Oct 23, 2015 at 11:49:14AM +0100, Theo Markettos wrote:
>> On Fri, Oct 23, 2015 at 10:17:07AM +0100, Gerald Holdsworth wrote:
>>>> A new version of RPCEmu is available, 0.8.13
>>>
>>> Nice, well done - but not for the Mac.
>>>
>>> Would it be worth putting the 0.8.12 Mac build on the 'contributed
>>> builds' on the page?
>>> https://bitbucket.org/caliston/rpcemu-spoon-caliston/downloads
>>
>> Let me do a test version, and if it hasn't broken anything that can be a
>> 'proper' 0.8.13.
>
> I have now built 0.8.13, that was straightforward. It's now on the above
> download page. It has received very little testing - it boots RISC OS 5 and
> seems to work for me. I believe it's still subject to the usual UI bugs -
> in particular there's no change on (lack of) Retina support.
Thanks for all your hard work getting RPCEmu for OS X going again, it's good to see a more recent build being widely available instead of just my really old ones!
Francis
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
Sunday, 25 October 2015
Re: [gccsdk] patch to build libsqlite3-0 in autobuilder
In article <562D3891.2070908@macpcrepair.co.uk>,
John Ballance <jwb@macpcrepair.co.uk> wrote:
> Hi
> Modified Makefile.in.p for libsqlite3-0 in the autobuilder. Adds -ldl to
> the link flags
This was only affecting the static build. I've checked in a fix to setvars
to disable loading extensions. I don't think it's applicable when building
the static library.
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
John Ballance <jwb@macpcrepair.co.uk> wrote:
> Hi
> Modified Makefile.in.p for libsqlite3-0 in the autobuilder. Adds -ldl to
> the link flags
This was only affecting the static build. I've checked in a fix to setvars
to disable loading extensions. I don't think it's applicable when building
the static library.
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
Re: [gccsdk] patch needed to get oslib for autobuilder
In article <562D3895.5090301@macpcrepair.co.uk>,
John Ballance <jwb@macpcrepair.co.uk> wrote:
> Hi
> S
> setvars for oslib points to an inaccessible url
> This is what is now needed:
> AB_SVN=svn://svn.code.sf.net/p/ro-oslib/code/trunk/!OSLib
Are you sure you've checked out the latest version of the autobuilder. This
change has already been done.
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
John Ballance <jwb@macpcrepair.co.uk> wrote:
> Hi
> S
> setvars for oslib points to an inaccessible url
> This is what is now needed:
> AB_SVN=svn://svn.code.sf.net/p/ro-oslib/code/trunk/!OSLib
Are you sure you've checked out the latest version of the autobuilder. This
change has already been done.
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
Re: [gccsdk] Qt5Base build error
In article <562D387B.9080103@macpcrepair.co.uk>,
John Ballance <jwb@macpcrepair.co.uk> wrote:
> Hi
> Attempting to compile Qt5Base from scratch, with latest tools and
> autobuilder, I get a failure to link to the XML stuff in libexpat.
Qt5Base is explicitly built using shared libraries. It looks like you built
some or all of the dependent libraries statically. That's why you're
getting errors.
You can default everything to build shared by saving the following in the
file 'build-setvars' in your build folder. You may need to change the paths
depending on where you are building from.
GCCSDK_INSTALL_CROSSBIN=~/gccsdk/cross/bin
GCCSDK_INSTALL_ENV=~/gccsdk/env
RO_SHAREDLIBS=yes
AB_ELFBUILD=yes
You will then need to rebuild all the dependent libraries.
If you're going to build the rest of the Qt libraries you'll also need to
revert Qt5Base to the previous 5.4.1 version.
To revert back use the following,
cd gccsdk/autobuilder/libraries/Qt/Qt5Base
svn -r 6858 update
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
John Ballance <jwb@macpcrepair.co.uk> wrote:
> Hi
> Attempting to compile Qt5Base from scratch, with latest tools and
> autobuilder, I get a failure to link to the XML stuff in libexpat.
Qt5Base is explicitly built using shared libraries. It looks like you built
some or all of the dependent libraries statically. That's why you're
getting errors.
You can default everything to build shared by saving the following in the
file 'build-setvars' in your build folder. You may need to change the paths
depending on where you are building from.
GCCSDK_INSTALL_CROSSBIN=~/gccsdk/cross/bin
GCCSDK_INSTALL_ENV=~/gccsdk/env
RO_SHAREDLIBS=yes
AB_ELFBUILD=yes
You will then need to rebuild all the dependent libraries.
If you're going to build the rest of the Qt libraries you'll also need to
revert Qt5Base to the previous 5.4.1 version.
To revert back use the following,
cd gccsdk/autobuilder/libraries/Qt/Qt5Base
svn -r 6858 update
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] patch needed to get oslib for autobuilder
Hi
S
setvars for oslib points to an inaccessible url
This is what is now needed:
AB_SVN=svn://svn.code.sf.net/p/ro-oslib/code/trunk/!OSLib
Thanks
John
S
setvars for oslib points to an inaccessible url
This is what is now needed:
AB_SVN=svn://svn.code.sf.net/p/ro-oslib/code/trunk/!OSLib
Thanks
John
--
Stop press! Just announced -
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
For full details of see https://www.utilitywarehouse.co.uk/reviews?exref=095761
Intrigued? Call me
John Ballance C.Eng MIET - jwb@macpcrepair.co.uk - 07976 295923
[gccsdk] patch to build libsqlite3-0 in autobuilder
Hi
Modified Makefile.in.p for libsqlite3-0 in the autobuilder. Adds -ldl to the link flags
Thanks
John
Modified Makefile.in.p for libsqlite3-0 in the autobuilder. Adds -ldl to the link flags
Thanks
John
--- Makefile.in 2015-05-05 19:08:54.000000000 +0100
+++ Makefile.in.fixed 2015-05-05 19:15:28.000000000 +0100
@@ -20,7 +20,7 @@
# C Compiler and options for use in building executables that
# will run on the platform that is doing the build.
#
-BCC = @BUILD_CC@ @BUILD_CFLAGS@
+BCC = /usr/bin/@BUILD_CC@ @BUILD_CFLAGS@
# TCC is the C Compile and options for use in building executables that
# will run on the target platform. (BCC and TCC are usually the
@@ -152,7 +152,7 @@
# libtool compile/link/install
LTCOMPILE = $(LIBTOOL) --mode=compile --tag=CC $(TCC) $(LTCOMPILE_EXTRAS)
-LTLINK = $(LIBTOOL) --mode=link $(TCC) $(LTCOMPILE_EXTRAS) @LDFLAGS@ $(LTLINK_EXTRAS)
+LTLINK = $(LIBTOOL) --mode=link $(TCC) $(LTCOMPILE_EXTRAS) @LDFLAGS@ $(LTLINK_EXTRAS) -ldl
LTINSTALL = $(LIBTOOL) --mode=install $(INSTALL)
# nawk compatible awk.
--
Stop press! Just announced -
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
For full details of see https://www.utilitywarehouse.co.uk/reviews?exref=095761
Intrigued? Call me
John Ballance C.Eng MIET - jwb@macpcrepair.co.uk - 07976 295923
[gccsdk] Qt5Base build error
Hi
Attempting to compile Qt5Base from scratch, with latest tools and autobuilder, I get a failure to link to the XML stuff in libexpat.
Presumably a patch needed to include libexpat, but I'm failing abysmally to find where.
this is an extract from the output@
I presume the bit in the command line :
Attempting to compile Qt5Base from scratch, with latest tools and autobuilder, I get a failure to link to the XML stuff in libexpat.
Presumably a patch needed to include libexpat, but I'm failing abysmally to find where.
this is an extract from the output@
arm-unknown-riscos-g++ -Wl,--no-undefined -Wl,-O1 -Wl,--enable-new-dtags -Wl,-rpath,/home/jb/GCCSDK474/env/lib -shared -o libqriscos.so .obj/main.o .obj/qriscosintegration.o .obj/qriscosbackingstore.o .obj/qriscoseventdispatcher.o .obj/qriscoswindow.o .obj/moc_qriscoseventdispatcher.o -L/home/jb/GCCSDK474/env/lib -Wl,-rpath-link,/home/jb/GCCSDK474/env/lib -L/home/jb/GCCSDK474/build/Qt5Base/qtbase-opensource-src-5.5.0/lib -lQt5PlatformSupport -Wl,-rpath-link,/home/jb/GCCSDK474/env/lib -lfontconfig -lfreetype -Wl,-rpath-link,/home/jb/GCCSDK474/env/lib -Wl,-rpath-link,/home/jb/GCCSDK474/env/lib -Wl,-rpath-link,/home/jb/GCCSDK474/env/lib -lQt5Gui -Wl,-rpath-link,/home/jb/GCCSDK474/env/lib -Wl,-rpath-link,/home/jb/GCCSDK474/env/lib -lQt5Core -ldl -lz -lpng12 -lOSLib32 -Wl,-rpath-link,/home/jb/GCCSDK474/env/lib
/home/jb/GCCSDK474/env/lib/libfontconfig.a(fcxml.o): In function `FcConfigMessage.constprop.11':
fcxml.c:(.text+0x2a4): undefined reference to `XML_GetCurrentLineNumber'
fcxml.c:(.text+0x2e4): undefined reference to `XML_GetCurrentLineNumber'
/home/jb/GCCSDK474/env/lib/libfontconfig.a(fcxml.o): In function `FcConfigMessage.constprop.12':
fcxml.c:(.text+0x394): undefined reference to `XML_GetCurrentLineNumber'
fcxml.c:(.text+0x3dc): undefined reference to `XML_GetCurrentLineNumber'
/home/jb/GCCSDK474/env/lib/libfontconfig.a(fcxml.o): In function `FcConfigParseAndLoad':
fcxml.c:(.text+0x1934): undefined reference to `XML_ParserCreate'
fcxml.c:(.text+0x197c): undefined reference to `XML_SetUserData'
fcxml.c:(.text+0x198c): undefined reference to `XML_SetDoctypeDeclHandler'
fcxml.c:(.text+0x199c): undefined reference to `XML_SetElementHandler'
fcxml.c:(.text+0x19a8): undefined reference to `XML_SetCharacterDataHandler'
fcxml.c:(.text+0x19b8): undefined reference to `XML_GetBuffer'
fcxml.c:(.text+0x19e8): undefined reference to `XML_ParseBuffer'
fcxml.c:(.text+0x1a24): undefined reference to `XML_ParserFree'
fcxml.c:(.text+0x1cd8): undefined reference to `XML_GetErrorCode'
fcxml.c:(.text+0x1cdc): undefined reference to `XML_ErrorString'
collect2: error: ld returned 1 exit status
I presume the bit in the command line :
-rpath-link,/home/jb/GCCSDK474/env/lib -lfontconfig -lfreetype -Wl, needs '-lexpat ' added, but I cannot locate what generates this string. Any thoughts appreciated Thanks John
--
Stop press! Just announced -
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
Most Trusted Broadband Provider in the 2014 Moneywise Customer Services Awards
For full details of see https://www.utilitywarehouse.co.uk/reviews?exref=095761
Intrigued? Call me
John Ballance C.Eng MIET - jwb@macpcrepair.co.uk - 07976 295923
Saturday, 24 October 2015
Re: Change languaje and Enable javascript
In message <4eb78d89149f3a2b14696d12b708290e@openmailbox.org>
Diego Herchhoren <lan.free1989@openmailbox.org> wrote:
>Hi friends.
>
>I'm user of Debian Sid, and I'm using Netsurf.
>
>I have a problem with languaje. I select spanish languaje in
>preferences, but the languaje is not changed.
The problem is that we don't yet have a Spanish translation for the
messages in the user interface.
Can you help? Can you translate the messages for us?
>Also, I have activated javascript, but it doesn't run.
The Javascript implementation is currently only beginning. It's
being worked on very hard, but Javascript is a big language, so it
will gradually improve over time.
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
Diego Herchhoren <lan.free1989@openmailbox.org> wrote:
>Hi friends.
>
>I'm user of Debian Sid, and I'm using Netsurf.
>
>I have a problem with languaje. I select spanish languaje in
>preferences, but the languaje is not changed.
The problem is that we don't yet have a Spanish translation for the
messages in the user interface.
Can you help? Can you translate the messages for us?
>Also, I have activated javascript, but it doesn't run.
The Javascript implementation is currently only beginning. It's
being worked on very hard, but Javascript is a big language, so it
will gradually improve over time.
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
Change languaje and Enable javascript
Hi friends.
I'm user of Debian Sid, and I'm using Netsurf.
I have a problem with languaje. I select spanish languaje in preferences, but the languaje is not changed.
Also, I have activated javascript, but it doesn't run.
What could be the problem?
Very thanks.
--
Diego Herchhoren www.marcha.org.ar
Twitter: @TwiterTarantino
n-1: thomassankara
Skype: tylerdurden495
Re: Job spec of a port maintainer
On 22 Oct, dsilvers wrote:
> > Is there a definition somewhere of what's involved?
>
> As this is open-source/free-software work there's rarely anything written
> down.
>
> My understanding of a port maintainer (for us at least) would be someone
> who can commit to being around on-channel at least a bit; is competent in
> the use of, and development of, applications on their target system, can
> respond in a timely fashion to tickets, ideally within a few days even if
> only to say thanks, I need to think about this; and will jump on front-end
> failures evident in the CI system within a day or two at most.
>
> Ideally they'd also be interested in learning about and assisting with the
> core of the NetSurf codebase, and also our libraries; since all ports rely
> on those to a greater or lesser extent.
That's a very useful outline, thanks (sorry about the slow reply, I messed up
the mail filter and it binned all your replies).
> > I went ahead and wrote patches for
> >
> > http://bugs.netsurf-browser.org/mantis/view.php?id=2266
> > http://bugs.netsurf-browser.org/mantis/view.php?id=2170
> > http://bugs.netsurf-browser.org/mantis/view.php?id=2336
> > http://bugs.netsurf-browser.org/mantis/view.php?id=2289
>
> Thank you. Hopefully Dave Higton may get to these, he has been invaluable
> recently :-)
It sounds like Dave H might be about to put his hand up as a RISC OS
maintainer then!
I can keep an eye on the tickets too, I've put the build environment on my
laptop so along with RPCEmu can pick over things when stuck on a train.
However, I'm not sure developing the core code/being on IRC is something I
can juggle as well, so some kind of job share might work better,
Sprow.
> > Is there a definition somewhere of what's involved?
>
> As this is open-source/free-software work there's rarely anything written
> down.
>
> My understanding of a port maintainer (for us at least) would be someone
> who can commit to being around on-channel at least a bit; is competent in
> the use of, and development of, applications on their target system, can
> respond in a timely fashion to tickets, ideally within a few days even if
> only to say thanks, I need to think about this; and will jump on front-end
> failures evident in the CI system within a day or two at most.
>
> Ideally they'd also be interested in learning about and assisting with the
> core of the NetSurf codebase, and also our libraries; since all ports rely
> on those to a greater or lesser extent.
That's a very useful outline, thanks (sorry about the slow reply, I messed up
the mail filter and it binned all your replies).
> > I went ahead and wrote patches for
> >
> > http://bugs.netsurf-browser.org/mantis/view.php?id=2266
> > http://bugs.netsurf-browser.org/mantis/view.php?id=2170
> > http://bugs.netsurf-browser.org/mantis/view.php?id=2336
> > http://bugs.netsurf-browser.org/mantis/view.php?id=2289
>
> Thank you. Hopefully Dave Higton may get to these, he has been invaluable
> recently :-)
It sounds like Dave H might be about to put his hand up as a RISC OS
maintainer then!
I can keep an eye on the tickets too, I've put the build environment on my
laptop so along with RPCEmu can pick over things when stuck on a train.
However, I'm not sure developing the core code/being on IRC is something I
can juggle as well, so some kind of job share might work better,
Sprow.
Friday, 23 October 2015
Re: [Rpcemu] RPCEmu 0.8.13
Excellent work. Thank you very much Theo. I'll give it a test over the weekend.
Cheers,
Gerald
Sent from one of my iPhones
> On 23 Oct 2015, at 21:37, Theo Markettos <theo@markettos.org.uk> wrote:
>
>> On Fri, Oct 23, 2015 at 11:49:14AM +0100, Theo Markettos wrote:
>> On Fri, Oct 23, 2015 at 10:17:07AM +0100, Gerald Holdsworth wrote:
>>>> A new version of RPCEmu is available, 0.8.13
>>>
>>> Nice, well done - but not for the Mac.
>>>
>>> Would it be worth putting the 0.8.12 Mac build on the 'contributed
>>> builds' on the page?
>>> https://bitbucket.org/caliston/rpcemu-spoon-caliston/downloads
>>
>> Let me do a test version, and if it hasn't broken anything that can be a
>> 'proper' 0.8.13.
>
> I have now built 0.8.13, that was straightforward. It's now on the above
> download page. It has received very little testing - it boots RISC OS 5 and
> seems to work for me. I believe it's still subject to the usual UI bugs -
> in particular there's no change on (lack of) Retina support.
>
> Theo
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
Cheers,
Gerald
Sent from one of my iPhones
> On 23 Oct 2015, at 21:37, Theo Markettos <theo@markettos.org.uk> wrote:
>
>> On Fri, Oct 23, 2015 at 11:49:14AM +0100, Theo Markettos wrote:
>> On Fri, Oct 23, 2015 at 10:17:07AM +0100, Gerald Holdsworth wrote:
>>>> A new version of RPCEmu is available, 0.8.13
>>>
>>> Nice, well done - but not for the Mac.
>>>
>>> Would it be worth putting the 0.8.12 Mac build on the 'contributed
>>> builds' on the page?
>>> https://bitbucket.org/caliston/rpcemu-spoon-caliston/downloads
>>
>> Let me do a test version, and if it hasn't broken anything that can be a
>> 'proper' 0.8.13.
>
> I have now built 0.8.13, that was straightforward. It's now on the above
> download page. It has received very little testing - it boots RISC OS 5 and
> seems to work for me. I believe it's still subject to the usual UI bugs -
> in particular there's no change on (lack of) Retina support.
>
> Theo
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
Re: [Rpcemu] RPCEmu 0.8.13
On Fri, Oct 23, 2015 at 11:49:14AM +0100, Theo Markettos wrote:
> On Fri, Oct 23, 2015 at 10:17:07AM +0100, Gerald Holdsworth wrote:
> > >A new version of RPCEmu is available, 0.8.13
> >
> > Nice, well done - but not for the Mac.
> >
> > Would it be worth putting the 0.8.12 Mac build on the 'contributed
> > builds' on the page?
> > https://bitbucket.org/caliston/rpcemu-spoon-caliston/downloads
>
> Let me do a test version, and if it hasn't broken anything that can be a
> 'proper' 0.8.13.
I have now built 0.8.13, that was straightforward. It's now on the above
download page. It has received very little testing - it boots RISC OS 5 and
seems to work for me. I believe it's still subject to the usual UI bugs -
in particular there's no change on (lack of) Retina support.
Theo
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
> On Fri, Oct 23, 2015 at 10:17:07AM +0100, Gerald Holdsworth wrote:
> > >A new version of RPCEmu is available, 0.8.13
> >
> > Nice, well done - but not for the Mac.
> >
> > Would it be worth putting the 0.8.12 Mac build on the 'contributed
> > builds' on the page?
> > https://bitbucket.org/caliston/rpcemu-spoon-caliston/downloads
>
> Let me do a test version, and if it hasn't broken anything that can be a
> 'proper' 0.8.13.
I have now built 0.8.13, that was straightforward. It's now on the above
download page. It has received very little testing - it boots RISC OS 5 and
seems to work for me. I believe it's still subject to the usual UI bugs -
in particular there's no change on (lack of) Retina support.
Theo
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
Re: [Rpcemu] Allegro not installed Ubuntu 14.04
In message <20151022233714.GB15853@spod.org>
on 23 Oct 2015 Peter Howkins wrote:
> On Thu, Oct 22, 2015 at 10:23:26PM +0100, Matthew Phillips wrote:
> > This problem is awfully similar to that which people were getting a year ago
> > (the subject line referenced Mint 17).
> >
> > I have installed liballegro4-dev, liballegro4.4 and liballegro4.4-plugin-alsa
> >
> > I have gcc and build-essential.
> >
> > When I do
> >
> > cd rpcemu/src
> > aclocal
>
> Why are you running aclocal?
>
> http://www.marutan.net/rpcemu/linuxcompile.html
>
> > ./configure
>
> On a fresh copy, does ./configure on its own work?
No, it didn't. I had followed the instructions on the page you refer to
first before I tried hunting for other suggestions. Sorry I didn't make this
clear.
I was trying aclocal as part of a sequence of suggestions from the posting a
year ago from Sean Kelly:
----------------
To save you look up time, here's what I did from start to finish.
1. Downloaded and unpacked the RPCEmu source code
http://www.marutan.net/rpcemu/cgi/download.php?sFName=0.8.11/rpcemu-0.8.11.tar.gz
2. Used the software manager in Mint to install liballegro4.2 and
liballegro4.2-plugin-jack and liballegro4 liballegro4-dev.
3. Copied (I could have moved) allegro.m4 from /usr/local/share/aclocal/
to /usr/share/aclocal/
4. Opened a terminal and issued the following commands - I know these
shouldn't all be required, but it's what worked for me, eventually, as
neither the deprecated nor the latest instructions on marutan.net
worked. I'm thinking make clean won't be required for you, since it will
all hopefully work first time.
cd rpcemu/src/
aclocal
automake
autoconf
make clean
LIBS="-lm -lpthread" ./configure --enable-dynarec
make
Then it was just a matter of copying the ROMs etc.
----------------
My symptoms had been the same as those he reported a year ago, but I had no
allegro.m4 file anywhere in /usr
I have now had a go at copying allegro.m4 from my old laptop onto the Ubuntu
10.04 laptop and RPCEmu now compiles. I am not sure which package is
supposed to supply the allegro.m4 file. I was not able to install liballegro
4.2 as Ubuntu's oldest offering is 4.4. I don't know whether that would
explain what happened.
--
Matthew Phillips
Durham
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
on 23 Oct 2015 Peter Howkins wrote:
> On Thu, Oct 22, 2015 at 10:23:26PM +0100, Matthew Phillips wrote:
> > This problem is awfully similar to that which people were getting a year ago
> > (the subject line referenced Mint 17).
> >
> > I have installed liballegro4-dev, liballegro4.4 and liballegro4.4-plugin-alsa
> >
> > I have gcc and build-essential.
> >
> > When I do
> >
> > cd rpcemu/src
> > aclocal
>
> Why are you running aclocal?
>
> http://www.marutan.net/rpcemu/linuxcompile.html
>
> > ./configure
>
> On a fresh copy, does ./configure on its own work?
No, it didn't. I had followed the instructions on the page you refer to
first before I tried hunting for other suggestions. Sorry I didn't make this
clear.
I was trying aclocal as part of a sequence of suggestions from the posting a
year ago from Sean Kelly:
----------------
To save you look up time, here's what I did from start to finish.
1. Downloaded and unpacked the RPCEmu source code
http://www.marutan.net/rpcemu/cgi/download.php?sFName=0.8.11/rpcemu-0.8.11.tar.gz
2. Used the software manager in Mint to install liballegro4.2 and
liballegro4.2-plugin-jack and liballegro4 liballegro4-dev.
3. Copied (I could have moved) allegro.m4 from /usr/local/share/aclocal/
to /usr/share/aclocal/
4. Opened a terminal and issued the following commands - I know these
shouldn't all be required, but it's what worked for me, eventually, as
neither the deprecated nor the latest instructions on marutan.net
worked. I'm thinking make clean won't be required for you, since it will
all hopefully work first time.
cd rpcemu/src/
aclocal
automake
autoconf
make clean
LIBS="-lm -lpthread" ./configure --enable-dynarec
make
Then it was just a matter of copying the ROMs etc.
----------------
My symptoms had been the same as those he reported a year ago, but I had no
allegro.m4 file anywhere in /usr
I have now had a go at copying allegro.m4 from my old laptop onto the Ubuntu
10.04 laptop and RPCEmu now compiles. I am not sure which package is
supposed to supply the allegro.m4 file. I was not able to install liballegro
4.2 as Ubuntu's oldest offering is 4.4. I don't know whether that would
explain what happened.
--
Matthew Phillips
Durham
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
Re: [Rpcemu] RPCEmu 0.8.13
Thanks for the update. I need to make an update of RISC OS FR's emulators (ad inside. sic.).
Stupid question, but does RPCEmu 0.8.13 now works with the very latest version of OpenVPN (under Windows)?
Today, I'm stuck with an old version, with security flaws. Not a big deal anyway :)Stupid question, but does RPCEmu 0.8.13 now works with the very latest version of OpenVPN (under Windows)?
No news around a JIT engine for ARM processors? (Pi Pi Pi Pi).
Bye, David.2015-10-23 14:02 GMT+02:00 David Pitt <pittdj@pittdj.co.uk>:
"Gerald Holdsworth", on 23 Oct, wrote:
> > A new version of RPCEmu is available, 0.8.13
>
> Nice, well done - but not for the Mac.
>
> Would it be worth putting the 0.8.12 Mac build on the 'contributed builds'
> on the page?
> https://bitbucket.org/caliston/rpcemu-spoon-caliston/downloads
>
> I'm using it with very little issue on the latest version of Mac OS X, and
> using RISC OS 5.23.
Not quite so lucky here, since El Capitan neither Menu nor Adjust click work
here. A bit of a show stopper. That is with a simple USB mouse. VRPC remains
OK mouse-wise.
--
David Pitt
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
--
Mobile : 06 76 67 91 60
Re: [Rpcemu] RPCEmu 0.8.13
"Gerald Holdsworth", on 23 Oct, wrote:
> > A new version of RPCEmu is available, 0.8.13
>
> Nice, well done - but not for the Mac.
>
> Would it be worth putting the 0.8.12 Mac build on the 'contributed builds'
> on the page?
> https://bitbucket.org/caliston/rpcemu-spoon-caliston/downloads
>
> I'm using it with very little issue on the latest version of Mac OS X, and
> using RISC OS 5.23.
Not quite so lucky here, since El Capitan neither Menu nor Adjust click work
here. A bit of a show stopper. That is with a simple USB mouse. VRPC remains
OK mouse-wise.
--
David Pitt
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
> > A new version of RPCEmu is available, 0.8.13
>
> Nice, well done - but not for the Mac.
>
> Would it be worth putting the 0.8.12 Mac build on the 'contributed builds'
> on the page?
> https://bitbucket.org/caliston/rpcemu-spoon-caliston/downloads
>
> I'm using it with very little issue on the latest version of Mac OS X, and
> using RISC OS 5.23.
Not quite so lucky here, since El Capitan neither Menu nor Adjust click work
here. A bit of a show stopper. That is with a simple USB mouse. VRPC remains
OK mouse-wise.
--
David Pitt
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
Re: [Rpcemu] RPCEmu 0.8.13
On Fri, Oct 23, 2015 at 10:17:07AM +0100, Gerald Holdsworth wrote:
> >A new version of RPCEmu is available, 0.8.13
>
> Nice, well done - but not for the Mac.
>
> Would it be worth putting the 0.8.12 Mac build on the 'contributed
> builds' on the page?
> https://bitbucket.org/caliston/rpcemu-spoon-caliston/downloads
>
> I'm using it with very little issue on the latest version of Mac OS
> X, and using RISC OS 5.23.
I can do a new Mac build, merging the 0.8.13 changes, but haven't had time
to do so this week. It'll still be subject to the Mac-specific bugs that
exist in 0.8.12, as I haven't done anything to the code since my previous
release. Looking at the 0.8.13 changes they appear to be mostly ISA-level,
so the Mac build shouldn't change GUI wise. Sadly OS X is not terribly
amenable to automated builds on (Linux) servers, otherwise I'd set it
up to autobuild nightly.
I'm happy for my builds to go as 'contributed builds' - I don't view my
bitbucket site as any way permanent, simply as a quick way for me to publish
the changes in a Mercurial-friendly way with a pretty GUI.
Let me do a test version, and if it hasn't broken anything that can be a
'proper' 0.8.13.
Theo
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
> >A new version of RPCEmu is available, 0.8.13
>
> Nice, well done - but not for the Mac.
>
> Would it be worth putting the 0.8.12 Mac build on the 'contributed
> builds' on the page?
> https://bitbucket.org/caliston/rpcemu-spoon-caliston/downloads
>
> I'm using it with very little issue on the latest version of Mac OS
> X, and using RISC OS 5.23.
I can do a new Mac build, merging the 0.8.13 changes, but haven't had time
to do so this week. It'll still be subject to the Mac-specific bugs that
exist in 0.8.12, as I haven't done anything to the code since my previous
release. Looking at the 0.8.13 changes they appear to be mostly ISA-level,
so the Mac build shouldn't change GUI wise. Sadly OS X is not terribly
amenable to automated builds on (Linux) servers, otherwise I'd set it
up to autobuild nightly.
I'm happy for my builds to go as 'contributed builds' - I don't view my
bitbucket site as any way permanent, simply as a quick way for me to publish
the changes in a Mercurial-friendly way with a pretty GUI.
Let me do a test version, and if it hasn't broken anything that can be a
'proper' 0.8.13.
Theo
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
Re: [Rpcemu] RPCEmu 0.8.13
> A new version of RPCEmu is available, 0.8.13
Nice, well done - but not for the Mac.
Would it be worth putting the 0.8.12 Mac build on the 'contributed builds'
on the page?
https://bitbucket.org/caliston/rpcemu-spoon-caliston/downloads
I'm using it with very little issue on the latest version of Mac OS X, and
using RISC OS 5.23.
And while we at it, can we get the RISC OS Compatibility list updated as it
still says that 6.20 is not compatible?
Cheers,
Gerald.
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
Nice, well done - but not for the Mac.
Would it be worth putting the 0.8.12 Mac build on the 'contributed builds'
on the page?
https://bitbucket.org/caliston/rpcemu-spoon-caliston/downloads
I'm using it with very little issue on the latest version of Mac OS X, and
using RISC OS 5.23.
And while we at it, can we get the RISC OS Compatibility list updated as it
still says that 6.20 is not compatible?
Cheers,
Gerald.
_______________________________________________
Rpcemu mailing list
Rpcemu@riscos.info
http://www.riscos.info/cgi-bin/mailman/listinfo/rpcemu
Re: [gccsdk] libfreetype6
In article <56283137.5010506@macpcrepair.co.uk>,
John Ballance <jwb@macpcrepair.co.uk> wrote:
> FYI this is the modified detect.p I have used...
> --- builds/unix/detect.mk.orig 2013-05-28 22:00:03.000000000 +0100
> +++ builds/unix/detect.mk 2014-01-04 18:35:52.374143596 +0000
> @@ -84,7 +84,7 @@
> CONFIG_SHELL="$(CONFIG_SHELL)" \
> $(CONFIG_SHELL) $(TOP_DIR)/builds/unix/configure $(value CFG)
> else
> - cd builds/unix; \
> + : #cd builds/unix; \
> CONFIG_SHELL="$(CONFIG_SHELL)" \
> $(CONFIG_SHELL) ./configure $(value CFG)
> endif
It just needed the two lines below commenting out too. I've checked in the
fix.
Chris.
_______________________________________________
GCCSDK mailing list gcc@gccsdk.riscos.info
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
John Ballance <jwb@macpcrepair.co.uk> wrote:
> FYI this is the modified detect.p I have used...
> --- builds/unix/detect.mk.orig 2013-05-28 22:00:03.000000000 +0100
> +++ builds/unix/detect.mk 2014-01-04 18:35:52.374143596 +0000
> @@ -84,7 +84,7 @@
> CONFIG_SHELL="$(CONFIG_SHELL)" \
> $(CONFIG_SHELL) $(TOP_DIR)/builds/unix/configure $(value CFG)
> else
> - cd builds/unix; \
> + : #cd builds/unix; \
> CONFIG_SHELL="$(CONFIG_SHELL)" \
> $(CONFIG_SHELL) ./configure $(value CFG)
> endif
It just needed the two lines below commenting out too. I've checked in the
fix.
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
RE: Job spec of a port maintainer
On Wed, 21 Oct 2015 18:31:36 +0100 Sprow wrote:
> Hi,
> In a couple of recent posts to the netsurf-users mailing list, and notes
> from
> ticket master Vincent, it's been noted there's no RISC OS maintainer at
> present in particular.
>
> Is there a definition somewhere of what's involved?
>
> For example: If it's responding to port specific tickets within ~2 weeks
> and
> maybe fixing 1 bug-a-month then that's probably background enough for one
> person to handle. If it involves writing lots of new code then that
> clearly
> needs a much bigger commitment. ie. it's maintenance in the car sense of
> "keeping something running that already runs".
>
> After some invaluable hand holding from Rob Kendrick (I failed to locate
> the
> instructions to set up a VM to build for RISC OS) I went ahead and wrote
> patches for
>
> http://bugs.netsurf-browser.org/mantis/view.php?id=2266
> http://bugs.netsurf-browser.org/mantis/view.php?id=2170
> http://bugs.netsurf-browser.org/mantis/view.php?id=2336
> http://bugs.netsurf-browser.org/mantis/view.php?id=2289
>
> but they presumably need reviewing. I know from a day job in engineering
> not
> to review my own code!
The RISC OS specific tasks don't seem all that onerous at the moment,
although perhaps I haven't been looking hard enough. But if two of us
contribute, it should be even better!
Dave
____________________________________________________________
FREE 3D EARTH SCREENSAVER - Watch the Earth right on your desktop!
Check it out at http://www.inbox.com/earth
> Hi,
> In a couple of recent posts to the netsurf-users mailing list, and notes
> from
> ticket master Vincent, it's been noted there's no RISC OS maintainer at
> present in particular.
>
> Is there a definition somewhere of what's involved?
>
> For example: If it's responding to port specific tickets within ~2 weeks
> and
> maybe fixing 1 bug-a-month then that's probably background enough for one
> person to handle. If it involves writing lots of new code then that
> clearly
> needs a much bigger commitment. ie. it's maintenance in the car sense of
> "keeping something running that already runs".
>
> After some invaluable hand holding from Rob Kendrick (I failed to locate
> the
> instructions to set up a VM to build for RISC OS) I went ahead and wrote
> patches for
>
> http://bugs.netsurf-browser.org/mantis/view.php?id=2266
> http://bugs.netsurf-browser.org/mantis/view.php?id=2170
> http://bugs.netsurf-browser.org/mantis/view.php?id=2336
> http://bugs.netsurf-browser.org/mantis/view.php?id=2289
>
> but they presumably need reviewing. I know from a day job in engineering
> not
> to review my own code!
The RISC OS specific tasks don't seem all that onerous at the moment,
although perhaps I haven't been looking hard enough. But if two of us
contribute, it should be even better!
Dave
____________________________________________________________
FREE 3D EARTH SCREENSAVER - Watch the Earth right on your desktop!
Check it out at http://www.inbox.com/earth
Subscribe to:
Posts (Atom)