Am 16.07.2017 um 07:23 schrieb Bryce Harrington:

Let me know if there is anything else that should be done or accounted for in the above plan.

Hi Bryce,

thanks for pinning down a schedule!

I have a list of three items that we should address before a 0.92.2 release. The first two are "simple" bugs which I think are too bad to ship another release with them (i.e. "blockers"):

* Windows print bug ( https://bugs.launchpad.net/inkscape/+bug/1665768 ). Our mingw-w64 based builds (devlibs64 and MSYS2) currently crash for many people when attempting to print. I was finally able to reproduce the bug on a machine I control, so I hope I'll be able to debug it soon. I'll keep you updated. (Any help is obviously appreciated) * Line height bugs (there were more than one if I'm not mistaken): What's the current status on them? I was not actively involved in the discussions and have never touched the relevant code, but from my own experiences and comments from a lot of people at work this is a serious deal-breaker for any productive work. I'd like to ask everybody who knows that part of our codebase and/or tracked the bugs/reports to make 100% sure they're all sorted out and report back!

Finally there's one important thing we'd need to figure out for Windows builds which I don't want to decide myself, so I hope it's fine if I use this possibility to get peoples input:

* I was working hard on improving MSYS2 builds on Windows and at this point I'd say they are superior to our old devlibs builds in every respect:

* Many bugs that were caused by outdated libraries were fixed (plus we obviously can profit from all improvements that were made to those libraries in the meantime) * Many things that were broken due to missing / misconfigured libraries could be sorted out * Some new features can be provided that depend on optional libraries (e.g. gtkspell)

That's the reason why I want to provide a 0.92.2 release built with MSYS2. However (and that's the big "but") they obviously had far less testing so far (but that's yet another reason why we need to get a release out there!), so I see four options:

1. Release MSYS2-only and hope it works out as well as I want to believe (this option makes sure people test it ;-) ) 2. Release MSYS2 as the main release but provide a backup built with devlibs in case we see unexpected issues (this is the option I'd prefer) 3. Release devlibs, but offer MSYS2 as a "public beta" (sounds not bad, but I'm pretty sure most people will stick to what they know for which we can't blame them) 4. Release devlibs-only for the 0.92.x branch (which would be a shame in my opinion. Personally I'd prefer an MSYS2 build over a devlibs build for productive work at any time and I think users should profit from the improvements, also - or especially - in the stable branch).

Let me know what you think (preferably soon-ish) so I can prioritize my work for the release!

Regards, Eduard

Am 17.07.2017 um 21:16 schrieb the Adib:

Hello All,

regarding Windows build using MSYS. I can confirm that building works. Some adjustments in the guides still needed (references to bzr.). Thanks for the guide to Eduard. Building uses ninja backend of cmake in fafour to make. I haven't timed the build but my impression is far better than cmake/make combination.

I would prefer option 1) MSYS2 only release. We need to see the impact on installer.

my 2 ct,

Adib.

Hi Adib,

fixing the instructions in the Wiki is on my todo list but I didn't get to it yet. Also I recently changed some things in the build system that also will slightly change instructions, but that work is finished now, so I hope to provide an update soon.

Any particular reason for not providing a devlibs build at all? (To clarify: I don't intend to provide a full set of installers in that case - only a .7z archive as fallback in case we really messed something up in the MSYS2 builds).

Installers certainly have to be updated but it should be doable with just a few path adjustments (I hope it can be done in a way that does not break compatibility with devlibs builds for the time being).

Regards, Eduard

Tav is busy this week, but he and I are mostly in charge of this bug at the moment. I'll bug him Monday about it again.

-C

On 20 Jul 2017 11:10, "Eduard Braun" <eduard.braun2@...173...> wrote:

So, quick update

Am 16.07.2017 um 13:38 schrieb Eduard Braun:

I have a list of three items that we should address before a 0.92.2 release. The first two are "simple" bugs which I think are too bad to ship another release with them (i.e. "blockers"):

• Windows print bug ( https://bugs.launchpad.net/inkscape/+bug/1665768

). Our mingw-w64 based builds (devlibs64 and MSYS2) currently crash for many people when attempting to print. I was finally able to reproduce the bug on a machine I control, so I hope I'll be able to debug it soon. I'll keep you updated. (Any help is obviously appreciated)

I was able to come up with a cairo patch that avoids the crash: https://bugs.freedesktop.org/show_bug.cgi?id=101833 Unfortunately I have no experience with cairo, so I don't really know if what I'm doing makes sense and there was no answer by cairo-devs yet, so I'm unsure on how to proceed... The prints I tested with the patched cairo show some weird issues with clipping which might highlight a more fundamental issue though, so there's still work to do, but at least we can provide a build that does not crash when printing...

• Line height bugs (there were more than one if I'm not mistaken):

What's the current status on them? I was not actively involved in the discussions and have never touched the relevant code, but from my own experiences and comments from a lot of people at work this is a serious deal-breaker for any productive work. I'd like to ask everybody who knows that part of our codebase and/or tracked the bugs/reports to make 100% sure they're all sorted out and report back!

No comments on this issue? Who was looking into the line-height bug(s)? I don't want to nag but this is one of these issues that have the potential to produce very bad PR if we don't get it right for yet another release!

Regards, Eduard

---

---

Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel

Just as an update, I don't yet have a pre0 I'm happy with. I've been wrestling with some make dist issues relating to the move from bzr to git, and digesting a number of recent cmake changes. I think I'm close but will have to pick it up tomrrow.

In the meantime, consider the 0.92.x frozen at this point. Do not land code or cmake changes directly - notify me about them as described below. (Translations are still ok for now.)

Bryce

On Sat, Jul 15, 2017 at 10:23:07PM -0700, Bryce Harrington wrote:

We were holding off for 0.92.2 until the git migration was done. Now with that behind us, I'd like to suggest some plans for doing another point release:

Jul 22: 0.92.x Pre-release + String Freeze (0.92.x branch only) + Packagers package pre0 release as desired + Translators work on translations + Revert changes that introduced regressions + Code changes must be reviewed and landed by Bryce (0.92.x branch only) Aug 4: Final Freeze + Final deadline for all translation work + Only Release Warden may commit (0.92.x branch only) Aug 5: 0.92.x Release Tarball + Packagers package final tarball Aug 7: Announce

Backporting of fixes has been going strong (good work everyone!) Please keep doing as you've been doing for one more week.

After the 22nd, I will be reviewing and landing all further code-changing patches. You can send me merge requests, lists of cherrypickable SHA's from trunk, or patchsets generated via `git format-patch`, as is convenient for you. Stuff I'll be watching for:

☑ Patch is short ☑ Fix is already in trunk ☑ Has been tested and confirmed by at least one other person ☑ References a bug ID or url in commit message ☑ No translatable strings are changed ☑ Bugfix only - no refactoring, no feature changes

On the 22nd, I'll package a pre0 tarball for testing purposes. The switch to git will require updating some release procedures so this will be a good dry run. It would be nice to see packages made.

Testing of the pre-release will be appreciated, however for release purposes we will be principally looking for _regressions_ compared with 0.92.1, so we can revert whatever patch caused the regression. Other bugs are welcome to be filed and worked on, of course, but I'd like to keep on schedule.

For the actual release, please take Aug 5th as a pencilled in date, that may need to change depending on how things go. When the tarball is ready I'll send a private note to the release team (packagers + announcers).

Vector Team, if you would like to take the announcement writing and/or dissemination tasks, let me know - I'd be very appreciative of the help.

Let me know if there is anything else that should be done or accounted for in the above plan.

Bryce

---

Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel