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* Many bugs that were caused by outdated libraries were fixed (plus we obviously can profit from all improvements that were made to those libraries in the meantime) * Many things that were broken due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