[KDE/Mac] The meinproc4 segfault is finally REPRODUCIBLE

[Ian Wadham]

Hi Marko,

On 07/05/2015, at 8:09 AM, Marko Käning wrote:
> On 29 Apr 2015, at 09:51 , Ian Wadham <iandw.au at gmail.com> wrote:
> 
>> Let's see what vazspam comes back with when he tries build.jobs=1.
> 
> he verified that building only one job at a time solves this issue.

Good to hear.

> What next?

1. Set up a script, in a separate build or user work directory, something like:

    <pathTo>meinproc4 <parameters for building KMyMoney Handbook> &
    <pathTo>meinproc4 <parameters for building KMyMoney manpage> &

2. if it does not crash, put a loop around the above (would need to keep
    deleting or renaming the output files).

3. If the crash output does not show anything obvious, as it probably will
    not, we will need to get log output from meinproc4.

4. If there is no debug output already in meinproc4 (and even if there is),
    we will probably need to get into the meinproc4 source code and insert
    diagnostic messages to help narrow down the crash circumstances and
    develop a fix.

Any volunteers?  I do not think I have time to get involved in this at the
moment, but I might be able to try steps 1 and 2 in the next few days.

And I would only be able to look at the KDE 4 version of meinproc4.

We are unlikely to get much help from KDE developers, sad to say, but you
can try… ;-)  The code of meinproc4 is very old and uncommented and nobody
knew much about it or had done anything much with it for several years, last
time this came up.  I presume it has now been ported to Frameworks and
renamed to meinproc5(?), in doctools or some name like that(?).

Cheers, Ian W.