[KDE/Mac] The meinproc4 segfault is finally REPRODUCIBLE

[René J.V. Bertin]

On Thursday May 07 2015 10:36:56 Marko Käning wrote:
Hi

>in this thread and in the related ticket we were discussing this and the
>result was, that it isn’t a simple double-use of a cache file. At least
>not the one specified on the command line.

Simple, no. Otherwise Linux would have been affected too, and the bug squashed a long time ago.

As I said in an earlier email, the null workaround would be to use a lockfile at the start of meinproc4, so that only a single instance is active at a given time. If the command takes a cache file argument and that file is created before any work is done, that file could serve as the lock (if it's the same for all potentially concurrent invocations).

>I just realised that the submitter of that issue still needs to build +docs.
>Perhaps this will produce the crash on his end…

How long does it take to build kmymoney4 +docs? I build my kdelibs4 with a bit more than default effort to preserve debugging info, so I actually may stand a better chance to produce a useful backtrace.
Also, I have a very recent kdelibs4 version (4.14.7.16).

R