Re: Regressions in Open Source projects ...

Hi Michael, all

I'm looking at another few Windows specific bugs that
> are of interest. Particularly with the new drmemory tool and Jesus'
> windows / debug builds - we should be able to progress here quickly.
> It'd be wonderful if we could get these traces for Windows specific
> bugs.

I'm quite curious to test that on a 100% reproducible crash as soon as I
find some time ;)


> Clearly that is true of some hypothetical user, for whom some serious
> regression blocks them from updating.


It is not so hypothetical. I had to keep a Portable LibreOffice 3.3.4 copy
around just to be able to open documents that included equations. And yes I
know it was a docx only bug but I do need to exchange manuscript revisions
with my co-authors and I haven't been able to convince them to drop
Microsoft Office 2007...

So this is a real life example for a single specific bug, as I'm sure there
are many others ;)


> There is however an easy solution
> for them - pay to have their (apparently un-interesting to the
> community) bugs fixed: then they can have their regression-free release,
> supported indefinately and everyone is happy :-)

This sounds like a pitch for users to switch to SUSE LibreOffice :P
BTW the code optimizations under Windows are fantastic ;)
But SUSE released version 3.4.2 while TDF is already at 3.5.1. Doesn't that
show a little about the importance of stability? ;)

As a straw-man (and I don't think anyone suggests this) - suggesting
> that we never ship until there are zero regressions would not meet this
> goal: we typically find regressions only after we ship.

I'm not sure I understand the straw-man concept but no one in QA is
advocating not to ship until zarro boogs are found... But we (at least I
am) are asking to give higher priority to fixing regressions and crashers
(even if they happen to only ONE user as long as it is reproducible) than
to adding new features. This doesn't mean AT ALL to stop adding features.
Just to change priorities ;)

As a general conclusion I think we are all doing the best given the limited
human resources and the scarcity of our most valuable resource: time!
It is because we do CARE about this project that we complain about the
short testing times AND the regressions ;)

Regards,
Pedro