Looking over the remaining level-9 bugs, here are my thoughts:
1230742 etc/gtk-2.0/gtkrc broken in windows build *** ======================================================================== Sounds like it should be fixed for this release. Doesn't sound too hard. Not sure what the cause is, though.
1227186 0.41 won't start; uses 90%+ of CPU ======================================================================== This may be a packaging issue unrelated to Inkscape, and not a must-fix bug.
1227185 Layer Add Then Move Crash *** ======================================================================== This sounds fairly important, but it appears to be Win32-specific and we do not have a backtrace. I think someone on Win32 should test it and report some additional debugging info, and we should treat it as a must fix.
1226498 WIN: crash on open/import tiff ======================================================================== This bug simply requires use of libtiff3.dll in the windows package. If that's included in the latest package, then we can close this bug. However, I don't think I'd rank this one as a must fix, since it's an uncommon function and apparently occurs just on Windows.
1221401 WIN 2003: Loading the Tutorial crash Inkscape ======================================================================== Sounds like it's an issue with Gtk on Windows 2003 SP1, rather than Inkscape. It works on other Windows releases, so I think this is rare enough that it doesn't qualify as must fix.
1221398 WIN 2003: Saving a file crashes Inkscape ======================================================================== Possibly the same situation as 1221401; also it lacks enough debugging info to troubleshoot. If it can be replicated more widely, and if we can get better debug output, maybe this could be a must fix, but as it stands currently it doesn't qualify.
1217361 freeze on quitting with transform dialog *** ======================================================================== Sounds like it's pretty pervasive, and causes a crash, so this is a must fix.
1210100 Selecting corrupted embedded images crashes *** ======================================================================== I think this may be fixed now, but we need someone on Win32 to confirm. Since it only happens on Windows, and only when using corrupted files, I'm not sure it qualifies as must fix, but since we have a patch to fix it, I think we may as well put priority on getting it done.
1202453 CRASH: effects ======================================================================== We do not have a reliable way to reproduce this. If someone can find a way to reliably reproduce it, I think it should be a must fix, but without that I don't think we should let it hold up the release.
1198534 The XML windows stops in a loop-search blocking program. ======================================================================== No one seems able to reproduce this one. It might be windows-only. Without some way to reproduce it, or a back trace, we can't consider it a must fix bug.
1169396 Crash on startup (win32) ======================================================================== We need someone on Win32 to verify this still exists; it may be fixed now. If it does still exist, it's a must fix, but unless someone can confirm and provide some updated debug info, I don't think it should be listed as must fix.
1167897 missing units.txt *** ======================================================================== We really need someone on Win32 to sort this one out, but I think it qualifies as a must fix.
1157125 crashes on startup - document.cpp line 367 ======================================================================== Possibly just a packaging issue; not enough info to debug it. It's also old and we haven't heard from the reporter for a while, so since no one can reproduce it, I don't think it qualifies.
1119649 Inkscape-0.41-1 One or more extensions failed to load ======================================================================== The bug that would not die. :-) The changes we've done for 0.42 will reduce the visibility of this bug, but since people are still reporting it, this bug should stay open to give them a place to add their comments to.
1108732 New releases still ~ARCH in Gentoo ======================================================================== I've been pushing to get 0.41 listed as stable in Gentoo, and have succeeded for 4 different architectures. Ironically, though, it's still not listed as stable for x86. In any case, this is completely external to the Inkscape codebase, so not really relevant to our bug fixing cycle.
995815 markers: markers don't take object color ======================================================================== This is more of a feature request, for an SVG 1.2 feature, so doesn't qualify as a must fix. Sure would be a nice feature to have though...
1220959 Scripts with Dependencies Crash Inkscape CVS ======================================================================== Since extensions are an optional feature, I don't think this is a must fix bug. Also, I suspect this one is going to require a LOT of work to get sorted out, so not something we can achieve in the near term anyway.
Okay, those are my thoughts on the critical bugs that are must-fixes. Please speak up if you disagree. I think we may be able to get these fixed within a week or two.
From reviewing the bugs it is clear that we are extremely desperate for
more Win32 people to get more heavily involved in the QA process.
Bryce
On Sun, Jul 03, 2005 at 03:48:49PM -0700, Bryce Harrington wrote:
Looking over the remaining level-9 bugs, here are my thoughts:
Bulia, would you mind reviewing the remaining bugs and giving your thoughts on which should be must-fixers? If we can narrow this list down really tightly, I think we may be looking at a release as early as this next weekend. :-)
Here's an update on the list I posted. The '***' indicates ones which I felt may be must-fixers.
1230742 etc/gtk-2.0/gtkrc broken in windows build ***
Sounds like it should be fixed for this release. Doesn't sound too hard. Not sure what the cause is, though.
No info yet. Ishmal, can you update us on where we stand with this one?
1217361 freeze on quitting with transform dialog ***
Sounds like it's pretty pervasive, and causes a crash, so this is a must fix.
bpfowler has added some troubleshooting info on this one. I looked into this a bit but wasn't able to sort it out; ishmal, have you had any luck with it?
1210100 Selecting corrupted embedded images crashes ***
I think this may be fixed now, but we need someone on Win32 to confirm. Since it only happens on Windows, and only when using corrupted files, I'm not sure it qualifies as must fix, but since we have a patch to fix it, I think we may as well put priority on getting it done.
bpfowler posted a patch for this today. I think we're close to having this one closed.
1167897 missing units.txt ***
We really need someone on Win32 to sort this one out, but I think it qualifies as a must fix.
I need some help from someone who understands how the internal paths work on Windows. The units.txt file is getting installed to the correct place, but the #define for INKSCAPE_DATADIR appears to not be set on Windows. We need to have a way to communicate where the share files are on Windows.
1227185 Layer Add Then Move Crash ***
This sounds fairly important, but it appears to be Win32-specific and we do not have a backtrace. I think someone on Win32 should test it and report some additional debugging info, and we should treat it as a must fix.
1227186 0.41 won't start; uses 90%+ of CPU
This may be a packaging issue unrelated to Inkscape, and not a must-fix bug.
1226498 WIN: crash on open/import tiff
This bug simply requires use of libtiff3.dll in the windows package. If that's included in the latest package, then we can close this bug. However, I don't think I'd rank this one as a must fix, since it's an uncommon function and apparently occurs just on Windows.
1221401 WIN 2003: Loading the Tutorial crash Inkscape
Sounds like it's an issue with Gtk on Windows 2003 SP1, rather than Inkscape. It works on other Windows releases, so I think this is rare enough that it doesn't qualify as must fix.
1221398 WIN 2003: Saving a file crashes Inkscape
Possibly the same situation as 1221401; also it lacks enough debugging info to troubleshoot. If it can be replicated more widely, and if we can get better debug output, maybe this could be a must fix, but as it stands currently it doesn't qualify.
1202453 CRASH: effects
We do not have a reliable way to reproduce this. If someone can find a way to reliably reproduce it, I think it should be a must fix, but without that I don't think we should let it hold up the release.
1198534 The XML windows stops in a loop-search blocking program.
No one seems able to reproduce this one. It might be windows-only. Without some way to reproduce it, or a back trace, we can't consider it a must fix bug.
1169396 Crash on startup (win32)
We need someone on Win32 to verify this still exists; it may be fixed now. If it does still exist, it's a must fix, but unless someone can confirm and provide some updated debug info, I don't think it should be listed as must fix.
1157125 crashes on startup - document.cpp line 367
Possibly just a packaging issue; not enough info to debug it. It's also old and we haven't heard from the reporter for a while, so since no one can reproduce it, I don't think it qualifies.
1119649 Inkscape-0.41-1 One or more extensions failed to load
The bug that would not die. :-) The changes we've done for 0.42 will reduce the visibility of this bug, but since people are still reporting it, this bug should stay open to give them a place to add their comments to.
1108732 New releases still ~ARCH in Gentoo
I've been pushing to get 0.41 listed as stable in Gentoo, and have succeeded for 4 different architectures. Ironically, though, it's still not listed as stable for x86. In any case, this is completely external to the Inkscape codebase, so not really relevant to our bug fixing cycle.
995815 markers: markers don't take object color
This is more of a feature request, for an SVG 1.2 feature, so doesn't qualify as a must fix. Sure would be a nice feature to have though...
1220959 Scripts with Dependencies Crash Inkscape CVS
Since extensions are an optional feature, I don't think this is a must fix bug. Also, I suspect this one is going to require a LOT of work to get sorted out, so not something we can achieve in the near term anyway.
Okay, those are my thoughts on the critical bugs that are must-fixes. Please speak up if you disagree. I think we may be able to get these fixed within a week or two.
From reviewing the bugs it is clear that we are extremely desperate for more Win32 people to get more heavily involved in the QA process.
Bryce
There is just one single bug left on our must fix list...
On Mon, Jul 04, 2005 at 05:22:32PM -0700, Bryce Harrington wrote:
1230742 etc/gtk-2.0/gtkrc broken in windows build ***
Sounds like it should be fixed for this release. Doesn't sound too hard. Not sure what the cause is, though.
No info yet. Ishmal, can you update us on where we stand with this one?
Ishmal reports this one is irrelevant with the new libs, so I think it can be closed now.
1217361 freeze on quitting with transform dialog ***
Sounds like it's pretty pervasive, and causes a crash, so this is a must fix.
bpfowler has added some troubleshooting info on this one. I looked into this a bit but wasn't able to sort it out; ishmal, have you had any luck with it?
Sounds like bpfowler is continuing to make good progress, but he needs help. There is a good comment thread on this bug. It sounds like this is fairly close to being fixed. Ishmal, could you review the comments and see if you can give answers to the questions, or if not, pull in someone who can?
1210100 Selecting corrupted embedded images crashes ***
I think this may be fixed now, but we need someone on Win32 to confirm. Since it only happens on Windows, and only when using corrupted files, I'm not sure it qualifies as must fix, but since we have a patch to fix it, I think we may as well put priority on getting it done.
bpfowler posted a patch for this today. I think we're close to having this one closed.
Sounds like this patch stops the crash, but may not be the ideal solution to the bad data problem. Regardless, this patch reduces the severity enough that I think we can de-list it from the must fix list.
1167897 missing units.txt ***
We really need someone on Win32 to sort this one out, but I think it qualifies as a must fix.
I need some help from someone who understands how the internal paths work on Windows. The units.txt file is getting installed to the correct place, but the #define for INKSCAPE_DATADIR appears to not be set on Windows. We need to have a way to communicate where the share files are on Windows.
Based on the discussion in the thread, it sounds like this one can be closed as unreproduceable.
Bryce
On 7/12/05, Bryce Harrington <bryce@...260...> wrote:
There is just one single bug left on our must fix list...
We got a couple others since then:
https://sourceforge.net/tracker/index.php?func=detail&aid=1235382&gr...
plus Aaron's startup crash reported in the list.
Also what about
https://sourceforge.net/tracker/index.php?func=detail&aid=1202453&gr...
Ted can reproduce it, and it's not fixed.
1230742 etc/gtk-2.0/gtkrc broken in windows build ***
Ishmal reports this one is irrelevant with the new libs, so I think it can be closed now.
I think it was with the new libs that it _became_ relevant. I'll test the latest build today and report.
1167897 missing units.txt ***
Based on the discussion in the thread, it sounds like this one can be closed as unreproduceable.
Again I haven't tried the absolute latest build, but at the time of pre1 it was VERY reproduceable. I run Inkscape and the first thing I get is the console window with the error message about units.txt. This definitely must be fixed.
1167897 missing units.txt
Based on the discussion in the thread, it sounds like this one can
be
closed as unreproduceable.
Again I haven't tried the absolute latest build, but at the time of pre1 it was VERY reproduceable. I run Inkscape and the first thing I get is the console window with the error message about units.txt. This definitely must be fixed.
At this time my only thought is that it may be a Win98 bug. As I stated in the thread, I've installed inkscape multiple times on over a dozen different win boxes w/ no issues (however, they've all been 2000/XP). Last I knew, Bulia was running 98 for win32 stuff. Can anyone else with a 98 box test this one out?
-Josh
I just tried http://inkscape.org/win32-snap/Inkscape0507120200.zip. It now uses gtk-wimp. Does it mean the release will use it too? If no, why the snapshots are different from the (pre)releases?
Apart from that, however, the snapshot is quite good. The bug that prevented using gtk-wimp before:
https://sourceforge.net/tracker/index.php?func=detail&aid=992667&gro...
is no more. Is this a new verion of gtk-wimp? Or gtk-wimp integrated into gtk?
1230742 etc/gtk-2.0/gtkrc broken in windows build ***
This one is fixed in that build.
1167897 missing units.txt ***
This one is fixed too.
So, now we need to do at least one more prerelease, pre2 (especially since cyreve and me did quite a lot of fixes in flowtext recently). If these problems do not reappear in it, we can close them. Also please make sure to commit Adib's installer patch before making pre2.
As for the freeze, sure, why not. But given that we have at least three unfixed crashes, this freeze may last a week (time to fix, verify, test etc). And we'll need to make pre3 after all these crashes are fixed, and wait at least a few days with pre3 before the final release.
What about screenshots? Can't anyone show off something impressive?
BTW, yesterday I updated keys.xml for 0.42 (quite a big update that was). Translators may want to take notice.
BTW, yesterday I updated keys.xml for 0.42 (quite a big update that was). Translators may want to take notice.
I saw this on CVS mailing-list. July 14 is our national day in France, which means I ll have some time for updating keys.fr.xml this WE.
BTW, do you wonder if I submit a patch (I hope this WE) for some additional context prefix to some short strings ? or would it be too late for .42 ?
Regards,
Matiphas
On Tue, Jul 12, 2005 at 10:47:51PM +0200, matiphas@...8... wrote:
BTW, yesterday I updated keys.xml for 0.42 (quite a big update that was). Translators may want to take notice.
I saw this on CVS mailing-list. July 14 is our national day in France, which means I ll have some time for updating keys.fr.xml this WE.
BTW, do you wonder if I submit a patch (I hope this WE) for some additional context prefix to some short strings ? or would it be too late for .42 ?
This weekend would probably be too late; we're already in string freeze, so unless it gets in extremely soon I don't think translators would be able to take advantage of the extra context anyway. However, definitely send in the patch, and even if it's too late for 0.42, that'll give us an early headstart for 0.43. :-)
Bryce
MenTaLguY wrote:
On Tue, 2005-07-12 at 10:04 -0300, bulia byak wrote:
plus Aaron's startup crash reported in the list.
Please file a bug for that one so we don't lose track of it..
If you want, but I'm gonna switch distros today and I won't be able to offer any more information. Since no one else was experiencing such a problem, I thought we could just forget about it.
Aaron Spike
On Tue, Jul 12, 2005 at 10:04:55AM -0300, bulia byak wrote:
On 7/12/05, Bryce Harrington <bryce@...260...> wrote:
There is just one single bug left on our must fix list...
We got a couple others since then:
https://sourceforge.net/tracker/index.php?func=detail&aid=1235382&gr...
Closed now.
plus Aaron's startup crash reported in the list.
Did this get resolved?
Also what about
https://sourceforge.net/tracker/index.php?func=detail&aid=1202453&gr...
Closed
Ted can reproduce it, and it's not fixed.
1230742 etc/gtk-2.0/gtkrc broken in windows build ***
Ishmal reports this one is irrelevant with the new libs, so I think it can be closed now.
I think it was with the new libs that it _became_ relevant. I'll test the latest build today and report.
You mentioned this one is fixed. Can it be closed in the tracker?
1167897 missing units.txt ***
Based on the discussion in the thread, it sounds like this one can be closed as unreproduceable.
Since you said it is working now, I've gone ahead and closed this one too.
It sounds like all of the must fix bugs are closed now pretty much.
Over the last couple of days I've also gone through everything in the ReleaseNotes and tried out all of the features. There's a few quirks here and there, but a) it never crashed, b) things worked pretty much as advertised, and c) compared to 0.41 it feels like there's a whole heck of a lot of good new stuff in it. :-)
I think we are ready to go into the next stage of the release process. I'll send a follow up note on that.
Bryce
On 7/14/05, Bryce Harrington <bryce@...260...> wrote:
plus Aaron's startup crash reported in the list.
Did this get resolved?
Probably unreproducible, looks like different-compiler-version-on-libs again.
https://sourceforge.net/tracker/index.php?func=detail&aid=1202453&gr...
Closed
Tentatively, since this but was hard to catch I think we need to wait some time to make sure it does not resurface.
1230742 etc/gtk-2.0/gtkrc broken in windows build ***
You mentioned this one is fixed. Can it be closed in the tracker?
I'd like to test the windows build of pre2 and then I'll close it.
1167897 missing units.txt ***
Based on the discussion in the thread, it sounds like this one can be closed as unreproduceable.
Since you said it is working now, I've gone ahead and closed this one too.
OK
It sounds like all of the must fix bugs are closed now pretty much.
Not really, you forgot
https://sourceforge.net/tracker/index.php?func=detail&aid=1217361&gr...
It's still there and it's a mustfix.
Bryce Harrington wrote:
Looking over the remaining level-9 bugs, here are my thoughts:
1230742 etc/gtk-2.0/gtkrc broken in windows build ***
Sounds like it should be fixed for this release. Doesn't sound too hard. Not sure what the cause is, though.
With the new libs, this one is likely irrelevant now.
1226498 WIN: crash on open/import tiff
This bug simply requires use of libtiff3.dll in the windows package. If that's included in the latest package, then we can close this bug. However, I don't think I'd rank this one as a must fix, since it's an uncommon function and apparently occurs just on Windows.
Fixed with the new libs.
1221401 WIN 2003: Loading the Tutorial crash Inkscape
Sounds like it's an issue with Gtk on Windows 2003 SP1, rather than Inkscape. It works on other Windows releases, so I think this is rare enough that it doesn't qualify as must fix.
Again, likely irrelevant now.
1221398 WIN 2003: Saving a file crashes Inkscape
Possibly the same situation as 1221401; also it lacks enough debugging info to troubleshoot. If it can be replicated more widely, and if we can get better debug output, maybe this could be a must fix, but as it stands currently it doesn't qualify.
I'm not sure what this one means. Is this maybe the network folder thing that is fixed with gtk2.6?
1217361 freeze on quitting with transform dialog ***
Sounds like it's pretty pervasive, and causes a crash, so this is a must fix.
I will look at this one this weekend. Something about widget destruction at shutdown.
1169396 Crash on startup (win32)
We need someone on Win32 to verify this still exists; it may be fixed now. If it does still exist, it's a must fix, but unless someone can confirm and provide some updated debug info, I don't think it should be listed as must fix.
Again, I have no idea.
1167897 missing units.txt ***
We really need someone on Win32 to sort this one out, but I think it qualifies as a must fix.
I have no idea what this means. Is this a gtk file?
From reviewing the bugs it is clear that we are extremely desperate for
more Win32 people to get more heavily involved in the QA process.
We've made a lot of progress in the week since you sent this. With the new libs we have a 'clean slate' that seems to be magically fixing a lot of issues for us.
Bob
On 7/8/05, Bob Jamison <rjamison@...357...> wrote:
1230742 etc/gtk-2.0/gtkrc broken in windows build ***
With the new libs, this one is likely irrelevant now.
I'll test today, but last time it happened to me with the 2.6 libs.
1167897 missing units.txt
***
=======================================================================
=
We really need someone on Win32 to sort this one out, but I think
it
qualifies as a must fix.
I have no idea what this means. Is this a gtk file?
Me neither... I've done a ton of installs and opened a number of SVGs from Explorer w/o ever running into this. I did comment on the bug though because I _may_ know what the problem was and have a potential solution.
We've made a lot of progress in the week since you sent this. With
the
new libs we have a 'clean slate' that seems to be magically fixing a lot of issues for us.
The new libs are awesome! ;) I'm curious, do we know if you still need to manually install Perl & Python on win32 for the extensions to work?
-Josh
On Fri, Jul 08, 2005 at 04:24:01PM -0500, Bob Jamison wrote:
1167897 missing units.txt ***
We really need someone on Win32 to sort this one out, but I think it qualifies as a must fix.
I have no idea what this means. Is this a gtk file?
No, it's a file that is in inkscape's share directory, that provides a list of unit definitions. The transformation dialog uses this information.
Bryce
participants (6)
-
unknown@example.com
-
Bob Jamison
-
Bryce Harrington
-
bulia byak
-
Joshua A. Andler
-
MenTaLguY