Re: Questian about Architecture
On Thu, Sep 12, 2002 at 08:01:40AM -0600, J. Scott Edwards wrote:
> I agree that one shouldn't restrict the sources.
*shrug* The build daemons override it anyway. I don't see a point in
inflating it into an issue of political correctness like Geert seems to
want to do; if it doesn't work for some deep-seated reason, then it
doesn't work. In a limited class of cases it *may* be worth saying that
to avoid wasting effort. It's not an insult to the port in question or
whatever.
Of course, there's always this:
http://www.gag.com/~bdale/talks/2002/debconf2/porting/html/mgp00014.html
The impression I had from that particular part of Bdale's talk is that
including only the architectures you know to work (the old
"Architecture: i386" problem) is a dreadful idea. On the other hand,
excluding architectures you know not to work, after talking to porters,
may be sensible. Perhaps.
> But I am curious how one tests on various platforms. All of my
> machines are x86 and PowerPC (and it's dead at the moment). Then of
> course I could use the compile farm at SourceForge. Do maintainers
> usually test on various platforms?
Not usually. The autobuilders ensure that it compiles; for the rest, we
rely on bug reports from users and the occasional blitz by porters to
fix common classes of problems.
--
Colin Watson [cjwatson@flatline.org.uk]
Reply to: