[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gnash-dev] Stable branch of Gnash

From: Benjamin Wolsey
Subject: Re: [Gnash-dev] Stable branch of Gnash
Date: Sat, 22 May 2010 08:13:43 +0200

Am Freitag, den 21.05.2010, 23:52 +0200 schrieb strk:
> On Fri, May 21, 2010 at 02:31:12PM -0500, Rob Savoye wrote:
> >   The problem with the "stable branch" concept is that stable in this
> > case means stagnant, with no real development of major new features.
> ... and the problem with "experimental branch" is that it breaks easily
> has more bugs and may blow up your computer :)
> Seriously, that's the whole point of having different branches.
> Let's use this capability...

The stable branch will not be stagnant and have no new features. It is
intended to be a development branch that has no commits that break the
testsuite, so develops in a stable, regression-free way.

The problem with the current branch is that Rob makes commits of such
low quality and in ignorance of failing tests that after three weeks of
work Gnash regularly crashes, while other parts are 'adapted' to fix
crashes without any idea of what they're for.

Other developers and I have always tried to point out these errors in
great detail (no "vague accusations" here, unless you mean backtraces,
code analysis, valgrind reports and design issues!) and I fix some of
them myself, but it's precisely the lack of developer time that now
makes this a problem. I'd like to spend my limited time adding new
features, not trying to contain problems caused by other code and
reviewing code that's evidently been written by guesscoding.

The current code Rob is working on, for instance, would be useful if it
worked. Maybe it does even work now when it doesn't crash. If it were
only committed after it's been reviewed, rewritten, and the testsuite
passes there would be no problem. But intermediate commits that break
things have never been acceptable, and *no other developer* has done
this as consistently over many years.

There seems to be no prospect of getting the maintainer to understand
this, as the same problems have arisen again and again, and as I've been
removed from the project on Savannah, it seems there is no choice anyway
but to move development of Gnash somewhere else.


Free Flash, use Gnash

Benjamin Wolsey, Software Developer -
C++ and Open-Source Flash blog -


Attachment: signature.asc
Description: Dies ist ein digital signierter Nachrichtenteil

reply via email to

[Prev in Thread] Current Thread [Next in Thread]