OK, 0.42.1 is pretty much ready to go, except there are two remaining issues I want to get nailed down first:
1. Mac OS X folks: is it possible for us to add some workaround for the user having to run "sudo /usr/X11R6/bin/fc-cache" manually on Tiger?
2. There's a layer name crash which has popped up again that I want to get nailed down.
Anyone else got anything?
-mental
On Thu, Aug 11, 2005 at 10:01:14PM -0400, MenTaLguY wrote:
OK, 0.42.1 is pretty much ready to go, except there are two remaining issues I want to get nailed down first:
- Mac OS X folks: is it possible for us to add some workaround for the
user having to run "sudo /usr/X11R6/bin/fc-cache" manually on Tiger?
- There's a layer name crash which has popped up again that I want to
get nailed down.
Anyone else got anything?
Are we ready to go on 0.42.1? Is there a change log or release notes to go along with this release? Also, what level of PR do you think is appropriate for this release?
Bryce
On Friday 12 August 2005 22:41, Bryce Harrington wrote:
On Thu, Aug 11, 2005 at 10:01:14PM -0400, MenTaLguY wrote:
OK, 0.42.1 is pretty much ready to go, except there are two remaining issues I want to get nailed down first:
- Mac OS X folks: is it possible for us to add some workaround
for the user having to run "sudo /usr/X11R6/bin/fc-cache" manually on Tiger?
- There's a layer name crash which has popped up again that I
want to get nailed down.
Anyone else got anything?
Are we ready to go on 0.42.1? Is there a change log or release notes to go along with this release? Also, what level of PR do you think is appropriate for this release?
Bryce
Is it going to build on an FC2 system now?
SF.Net email is Sponsored by the Better Software Conference & EXPO September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
On Fri, Aug 12, 2005 at 10:57:21PM -0400, Gene Heskett wrote:
On Friday 12 August 2005 22:41, Bryce Harrington wrote:
On Thu, Aug 11, 2005 at 10:01:14PM -0400, MenTaLguY wrote:
OK, 0.42.1 is pretty much ready to go, except there are two remaining issues I want to get nailed down first:
- Mac OS X folks: is it possible for us to add some workaround
for the user having to run "sudo /usr/X11R6/bin/fc-cache" manually on Tiger?
- There's a layer name crash which has popped up again that I
want to get nailed down.
Anyone else got anything?
Are we ready to go on 0.42.1? Is there a change log or release notes to go along with this release? Also, what level of PR do you think is appropriate for this release?
Is it going to build on an FC2 system now?
Do you have the power to test this? I do not know. It would be extremely invaluable to test build capability on FC 2, 3, and 4. Do you have the ability to do testing of this? If not, would you mind accumulating a couple people to help perform this test?
Thanks, Bryce
On Fri, 2005-08-12 at 19:41 -0700, Bryce Harrington wrote:
Are we ready to go on 0.42.1? Is there a change log or release notes to go along with this release? Also, what level of PR do you think is appropriate for this release?
Pretty much. I'd like the packagers and maybe a few developers to check out the RELEASE_0_42_BRANCH and give me a verbal okay before I tag it. Once it's tagged we're done, I think.
I've updated the ChangeLog and NEWS file ... I don't know what else we might want to do. But that's why I want people to have a look.
-mental
On Fri, Aug 12, 2005 at 11:31:48PM -0400, MenTaLguY wrote:
On Fri, 2005-08-12 at 19:41 -0700, Bryce Harrington wrote:
Are we ready to go on 0.42.1? Is there a change log or release notes to go along with this release? Also, what level of PR do you think is appropriate for this release?
Pretty much. I'd like the packagers and maybe a few developers to check out the RELEASE_0_42_BRANCH and give me a verbal okay before I tag it. Once it's tagged we're done, I think.
Good call. Are there particular developers you'd like to have review it? The stock 0.42 release has been working fine on my system so my sign-off won't mean much.
Can the packagers for the major distributions report in here with their evaluation of the 0.42.1 branch? Can you let us know if you run into any troubles?
I've updated the ChangeLog and NEWS file ... I don't know what else we might want to do. But that's why I want people to have a look.
Generally, we also do a ReleaseNotes, which I think would be especially valuable in this situation. Perhaps this can be adapted from the NEWS file?
Bryce
On Fri, 12 Aug 2005, Bryce Harrington wrote:
On Thu, Aug 11, 2005 at 10:01:14PM -0400, MenTaLguY wrote:
OK, 0.42.1 is pretty much ready to go, except there are two remaining issues I want to get nailed down first:
- Mac OS X folks: is it possible for us to add some workaround for
the user having to run "sudo /usr/X11R6/bin/fc-cache" manually on Tiger?
Sorry for my silence on this, I've been devoting all my time to the Summer of Code stuff. I had a look at this on Friday night and a little bit of today on the two Tiger machines I have access to. I can't reproduce the symptoms at all -- new local font config caches *always* get regenerated for me on these machines, even without the presence of /etc/fonts/fonts.conf. All I can think is that maybe there are different versions of Apple's X11 floating around with different behaviour.
So, I've done the following (commited to CVS). On Tiger, when Inkscape is launched, after checking X11 is available, a dialog is brought up informing users of the fontconfig problem. The dialog has a button labelled "run fc-cache". If chosen, an OS X authentication dialog in launched and fc-cache is run following the authentication. A file is stored in ~/.inkscape to mark that fc-cache has been run. Thus, once you you've gone though the process it won't try this again.
Cheers, Michael
On Sat, 2005-08-13 at 18:08 +1000, Michael Wybrow wrote:
So, I've done the following (commited to CVS). On Tiger, when Inkscape is launched, after checking X11 is available, a dialog is brought up informing users of the fontconfig problem. The dialog has a button labelled "run fc-cache". If chosen, an OS X authentication dialog in launched and fc-cache is run following the authentication. A file is stored in ~/.inkscape to mark that fc-cache has been run. Thus, once you you've gone though the process it won't try this again.
Ok, good. This is exactly what I've been waiting for.
However, could you please make one small change before I merge to 0.42.1?
Instead of checking for ~/.inkscape/.fccache, check for /etc/fonts/fonts.conf directly.
There's no reason to bug them if it already exists, is there?
-mental
Mental,
Cyreve just fixed another crash bug - personally I do not completely understand the fix, but it seems safe in any case, so I think it should go to 42.1. Patch attached.
On Sat, 13 Aug 2005, MenTaLguY wrote:
On Sat, 2005-08-13 at 18:08 +1000, Michael Wybrow wrote:
So, I've done the following (commited to CVS). On Tiger, when Inkscape is launched, after checking X11 is available, a dialog is brought up informing users of the fontconfig problem. The dialog has a button labelled "run fc-cache". If chosen, an OS X authentication dialog in launched and fc-cache is run following the authentication. A file is stored in ~/.inkscape to mark that fc-cache has been run. Thus, once you you've gone though the process it won't try this again.
Ok, good. This is exactly what I've been waiting for.
However, could you please make one small change before I merge to 0.42.1?
Instead of checking for ~/.inkscape/.fccache, check for /etc/fonts/fonts.conf directly.
There's no reason to bug them if it already exists, is there?
Unfortunately there doesn't seem to be a good way to check if the font caches need to be regenerated. Or at least I can't code a check for this since I can't recreate the problem on any of the Tiger machines I have access to. The problem is not /etc/fonts/fonts.conf, that should always exist and is not created by fc-cache command. fc-cache creates .fonts.cache-1 files in users home directories if run by a user, or font.cache-1 files in the system font directories if run by root. These files contain some information about fonts available fontconfig to speed up application launching. Incidently, the automatic generation of these files was the reason we were seeing very slow start times for some users on Tiger.
Some users that have reported this error even had the cache files present in the system fonts directories but they just didn't seem to be up to date or valid. Thus, until I find a machine on which I can detect the actual problem the next best thing is to just show this warning to everyone the first time Inkscape is run on Tiger.
Cheers, Michael
Mental, so what's up with 42.1? It's long overdue IMHO.
On Thu, 2005-08-18 at 15:14 -0300, bulia byak wrote:
Mental, so what's up with 42.1? It's long overdue IMHO.
1. somebody needs to write release notes
2. haven't heard anything back from anyone else checking out the branch like I asked to say it's OK
3. I'm away for Otakon
Once at least the first two are resolved, we can tag RELEASE_0_42_1 and on the branch and go.
-mental
On 8/18/05, MenTaLguY <mental@...3...> wrote:
On Thu, 2005-08-18 at 15:14 -0300, bulia byak wrote:
Mental, so what's up with 42.1? It's long overdue IMHO.
- somebody needs to write release notes
Hmm, I thought you already did it. Anyway, if you expected me (for example) to do it, you should have contacted me directly about it.
- haven't heard anything back from anyone else checking out the branch
like I asked to say it's OK
As for the patches that went into it, I'm totally OK, as I said before.
As for testing the end result, it's a major hassle to set up a branch compile. Windows users can't do it at all. So, can we instead tag it, make win/linux/osx binaries and ask people to test it? And do the release if they're ok?
On Thu, 2005-08-18 at 16:17 -0300, bulia byak wrote:
On 8/18/05, MenTaLguY <mental@...3...> wrote:
On Thu, 2005-08-18 at 15:14 -0300, bulia byak wrote:
Mental, so what's up with 42.1? It's long overdue IMHO.
- somebody needs to write release notes
Hmm, I thought you already did it. Anyway, if you expected me (for example) to do it, you should have contacted me directly about it.
I planned on doing it myself, but I haven't had time to do it in the past couple weeks. So somebody needs to at this point. Probably won't be me. Doesn't have to be you.
As for testing the end result, it's a major hassle to set up a branch compile. Windows users can't do it at all.
How come?
-mental
On 8/18/05, MenTaLguY <mental@...3...> wrote:
On Thu, 2005-08-18 at 16:17 -0300, bulia byak wrote:
On 8/18/05, MenTaLguY <mental@...3...> wrote:
On Thu, 2005-08-18 at 15:14 -0300, bulia byak wrote:
Mental, so what's up with 42.1? It's long overdue IMHO.
- somebody needs to write release notes
Hmm, I thought you already did it. Anyway, if you expected me (for example) to do it, you should have contacted me directly about it.
I planned on doing it myself, but I haven't had time to do it in the past couple weeks. So somebody needs to at this point. Probably won't be me. Doesn't have to be you.
Can you just resend the complete changelog / list of patches you have? I'll then massage it into a section in NEWS.
As for testing the end result, it's a major hassle to set up a branch compile. Windows users can't do it at all.
How come?
Most Windows users don't compile at all. They use compiled binaries. So, I think we should either:
- decide that the patches are safe enough and just do the release
- decide that some testing won't hurt, in which case make binaries for people to test
Whereas lingering at the pre-tagging stage does not make much sense at all, imho.
On Thu, 2005-08-18 at 16:50 -0300, bulia byak wrote:
Can you just resend the complete changelog / list of patches you have? I'll then massage it into a section in NEWS.
I sort of did do that (see the "Important Bugfixes" section). The descriptions are just skeletal and not very useful though.
-mental
On Thu, 2005-08-18 at 16:50 -0300, bulia byak wrote:
Can you just resend the complete changelog / list of patches you have? I'll then massage it into a section in NEWS.
FWIW the list of patches is reflected in the branch's ChangeLog.
I guess if you're satisfied with the current NEWS I'll go ahead and tag. It has taken long enough...
-mental
On 8/18/05, MenTaLguY <mental@...3...> wrote:
I guess if you're satisfied with the current NEWS I'll go ahead and tag. It has taken long enough...
Looking at
http://cvs.sourceforge.net/viewcvs.py/inkscape/inkscape/NEWS?view=markup&...
I don't see a mention of the OSX issues. If they are fixed, they should be in "fixed bugs", otherwise in "known problems".
Other than that, I'm OK with this.
On 8/18/05, MenTaLguY <mental@...3...> wrote:
I've tagged RELEASE_0_42_1. Hopefully I've not forgotten anything.
Ted, will you please make & post the tarball?
Packagers, please note, we need you again!
On Fri, 2005-08-19 at 23:45 -0300, bulia byak wrote:
On 8/18/05, MenTaLguY <mental@...3...> wrote:
I've tagged RELEASE_0_42_1. Hopefully I've not forgotten anything.
Ted, will you please make & post the tarball?
Yup, it is uploaded:
https://sourceforge.net/project/showfiles.php?group_id=93438&package_id=...
MD5 Sums:
b520179d459a3b49af56d58ed6981762 inkscape-0.42.1.tar.bz2 5c3218d8cb2542904332ff0d3fe14624 inkscape-0.42.1.tar.bz2.sig d8f13877b2c07bbdc4452e9f0c68be93 inkscape-0.42.1.tar.gz 9d04c717d49221d6a06e547ad1ab4c2c inkscape-0.42.1.tar.gz.sig
Should we hide the 0.42 release when there are packages for this one?
--Ted
On Sat, 20 Aug 2005, Ted Gould wrote:
On Fri, 2005-08-19 at 23:45 -0300, bulia byak wrote:
On 8/18/05, MenTaLguY <mental@...3...> wrote:
I've tagged RELEASE_0_42_1. Hopefully I've not forgotten anything.
Ted, will you please make & post the tarball?
Yup, it is uploaded:
https://sourceforge.net/project/showfiles.php?group_id=93438&package_id=...
The OSX package is up.
As an aside, what's the feeling on stripping libraries and binaries? This is worth talking about for the OSX package since it is bundled with all the necessary libraries. Stripping the libraries and binary reduces the Inkscape install footprint by about 40MB. In addition, the dmg download size drops by almost half from 23MB to 12MB.
What happens with autopackage and the windows installer? Are these stripped?
FWIW, there haven't been any OSX crash-bugs reported that have benefitted from all the debugging symbols, the problems have mostly been packaging problems.
I didn't strip the 0.42.1 package I uploaded, but I have a stripped dmg which I have tested, and have committed a trivial patch to the OSX build script to allow for the possibility of creating stripped dmgs.
Cheers, Michael
Michael Wybrow wrote:
What happens with autopackage and the windows installer? Are these stripped?
Autopackage is stripped by virtue of being an autopackage and there is no other way. If possible it would be nice to have both stripped and unstripped. I'm sure there will be a time when the debugging symbols will help solve a problem.
Aaron Spike
Ted Gould wrote:
bulia byak wrote:
Ted, will you please make & post the tarball?
up, it is uploaded:
Hello,
could you please add the files that are required to compile it for Win-32 via MinGW? They were gone in the .42-tarball but present in CVS. It would be nice if one could compile the .42.1-tarball "out of the box" with MinGW.
(There is at least config.h.mingw, po/Makefile.mingw, mingwenv.bat missing, most likely more. Donno how they are generated)
thanks for that great program & greetings
Bertram Drost schrieb: ...
Hello,
could you please add the files that are required to compile it for Win-32 via MinGW? They were gone in the .42-tarball but present in CVS. It would be nice if one could compile the .42.1-tarball "out of the box" with MinGW.
(There is at least config.h.mingw, po/Makefile.mingw, mingwenv.bat missing, most likely more. Donno how they are generated)
thanks for that great program & greetings
Ciao, there is this Bug: http://sourceforge.net/tracker/index.php?func=detail&aid=1245838&gro...
I wrote some comments what is missing. Currently nobody is assigned. Need one who knows autotool. But I don't think this should have solved in a 42.1 version we should try to fix that for a proper .43.
Adib.
On Sun, 2005-08-21 at 14:10 +0200, Bertram Drost wrote:
could you please add the files that are required to compile it for Win-32 via MinGW? They were gone in the .42-tarball but present in CVS. It would be nice if one could compile the .42.1-tarball "out of the box" with MinGW.
(There is at least config.h.mingw, po/Makefile.mingw, mingwenv.bat missing, most likely more. Donno how they are generated)
Yes, that is the plan. It didn't happen for the 0.42.x's, but I'm hoping to get it done for 0.43. Sorry about that.
--Ted
Ted Gould wrote:
On Sun, 2005-08-21 at 14:10 +0200, Bertram Drost wrote:
could you please add the files that are required to compile it for Win-32 via MinGW?
Yes, that is the plan. It didn't happen for the 0.42.x's, but I'm hoping to get it done for 0.43. Sorry about that.
np :) I just thought it was forgotten and would only require a "configure --for-mingw" or sth. similiary simple.
Thanks and keep up the good work!
participants (10)
-
unknown＠example.com -
Adib Taraben -
Bertram Drost -
Bryce Harrington -
bulia byak -
Gene Heskett -
MenTaLguY -
Michael Wybrow -
Ralf Stephan -
Ted Gould