Summary: | The process for the imaps://{server} protocol died unexpectedly. | ||
---|---|---|---|
Product: | [Unmaintained] kmail | Reporter: | Ivan D Vasin <ivan> |
Component: | IMAP | Assignee: | kdepim bugs <kdepim-bugs> |
Status: | RESOLVED WORKSFORME | ||
Severity: | normal | ||
Priority: | NOR | ||
Version: | 1.13.6 | ||
Target Milestone: | --- | ||
Platform: | Gentoo Packages | ||
OS: | Linux | ||
Latest Commit: | Version Fixed In: | ||
Sentry Crash Report: |
Description
Ivan D Vasin
2011-02-12 22:38:17 UTC
this bug disappeared from my system. after the last time i verified it as reproducible, i performed the following possibly relevant steps. note that most of these are Gentoo-specific. 1. rescue my somewhat broken installation of GCC 4.5 + Graphite (broken in that it causes ICEs at seemingly random times when building various packages, including GCC itself). 1. disable Graphite loop optimizations in CFLAGS in /etc/make.conf. 3. # USE=-gtk emerge sys-devel/gcc:4.4 5. # gcc-config x86_64-pc-linux-gnu-4.4.5 4. # USE=-gtk emerge -e system 6. # USE=-gtk emerge sys-devel/gcc:4.5 5. # gcc-config x86_64-pc-linux-gnu-4.5.2 7. # emerge -e gtk+ 8. # emerge gcc 2. do a total rebuild of my packages now that GCC is somewhat sane. 1. # emerge -e system 2. # emerge -e world 3. update and upgrade packages, including KDE 4.6.0 -> 4.6.1. 1. # emerge --sync 2. # emerge -uND world without knowing the real source of this problem, i'm guessing it could've been any of the following: * a bug that got fixed in KDE 4.6.1. * KMail or one of its libraries doesn't support GCC 4.5 + Graphite. i say GCC 4.5 because this bug wasn't present on the same system with GCC 4.4 + Graphite. then again, given the apparently insane nature of GCC 4.5 + Graphite in general, it's probably not fair to blame it on KMail. * undefined behavior caused by a reverse dependency rebuild requirement that, for whatever reason, didn't get caught by past runs of revdep-rebuild, but naturally got caught by ``emerge -e``. but usually these manifest as "cannot load shared object [...]" when launching, not as buggy runtime behavior. |