Re: GCC 3.0 Release Criteria

["Martin v. Loewis" <martin at loewis dot home dot cs dot tu-berlin dot de>]

> IMO we should test instead just for Linux

You cannot do that. There is not "just Linux". For a strict criterion,
you'd have to say something like "we require it to work on the
computer of Andreas Jäger, on the evening of the release". That is
something that gives a clear yes/no answer.

Of course, if it works on the evening of the release, and nothing is
changed, it probably works also the next day. Also, when the computer
of Mark Mitchell has the same Linux version installed, it probably
works on his computer as well. If I have a *different* version of
Linux installed, it may fail for me, but work for you - maybe it is a
bug in the kernel, or a glitch in the C library that causes it to
fail. If you have different Linux distributions, there is no 100%
guarantee that it works on all of them if it works on one of
them. That makes it a bad release criterion: If it fails for 1 person,
but works for ten, on identical compiler sources - to release or not
to release? Perhaps that person just had misconfigured the compiler,
or there is a serious problem. To reproduce it, you have to know
*exactly* what operating system was used.

Of course, Linux will be a supported platform, anyhow.

Regards,
Martin