Re: Memory leak a "fix" for stable tree?

[Paul Edwards]

"Derek Robert Price" <derek@ximbiot.com> wrote in message 
news:mailman.7673.1055268807.21513.bug-cvs@gnu.org...
> Just wondering if anyone had an opinion on this.  I have a six or seven
> line fix for a memory leak (usually only a few bytes leak, and not more
> than what has already been allocated for argv) that I'm getting ready to
> commit to the feature branch.  Is that a valid fix for the stable tree,
> or are we only fixing errors that cause user-perceivable problems there?

As a principle, I believe that anything that is considered a
defect, even minor, should be fixed in the stable one.  The
goal as I see it is to create a quality product, as opposed to
a feature-filled product.  The goal is to give people an
alternative place to add their "you-beaut latest features"
without interfering with quality development.  ie to protect
against introducing bugs while introducing new features.  If
bugs are introduced while fixing defects, that's life in software.
But if bugs are introduced while someone makes CVS
auto-updates web pages via WAP, that is unacceptable.

Web via WAP is not a defect or design error or something
that makes basic operations unusable or inconsistent or
anything at all.

I agree it is a grey area that warrants asking these questions.

BFN.  Paul.