Is it worth reporting build errors on obsolete platforms?

[Andy Dougherty]

On Tue, 20 Apr 2010, jerry gay wrote:

> On Tue, Apr 20, 2010 at 16:55, LocoPollo <parrot at sol42.com> wrote:
> > Hello.
> > The platform is Solaris 9/SPARC, running on a SparcStation 5.  I'm using the bundled gcc 2.95 compiler.
> > There are some new problems with 2.3.0 (Parrot_{gm|local|asc}time_r), and one known error (Parrot_unsetenv) from 2.2.0 and earlier remains.  Is it worth reporting them?
> >
> we don't have (as far as i'm aware) active porters for gcc 2.95 on
> sol9/sparc, but parrot strives to be portable to as many platforms as
> possible.  we'd love to have your bug reports via trac.parrot.org, and
> if possible, regular test runs or smoke results on any platforms you
> have available.  if you can test parrot and report results regularly,
> it can only improve our product.

Yes, such reports can be helpful.  Even though the code may be in a 
section not tested by usual parrot developers, some of these problems 
could have been avoided ahead of time, and can be easily fixed by someone 
without access to that platform.  For example, the unsetenv problem is 
already reported as TT #1140, and is due to someone changing the signature 
for Parrot_setenv without grep-ing the sources for other instances of that 
function.  It could easily have been found ahead of time that way.  
Similar issues have arisen quite recently in TT #1567 and in TT #1571.

Of course, if you are able to test and report (and even submit patches) 
regularly, that's even more helpful, but even just reporting the items can 
still be useful.

-- 
    Andy Dougherty		doughera at lafayette.edu