Clive Page wrote:
> > > Yes, of course. But in turn vendors should help potential bug reporters
> > > by giving them information on known bugs.
> That's all I am asking. Some vendors used to do that so it's clearly not
> impossible and though some descriptions were too vague to be useful, many
> were not.
The problem of listing "known bugs" is not that the descriptions might
be too vague to be useful, but that they might be *misleading*.
Misleading in giving a false impression of completeness.
It is impossible, even for someone intimately acquainted with the
internals of the compiler in question, to derive from the given
(failing) example *and a complete description of the bug*, which _class_
of source code constructions will be vulnerable to *that* bug. So
certainly the users won't be able to do that.
Therefore, such a list *won't* help you to "recognise" errant behaviour
as a known bug _before_ you actually isolate the problematic code and
hand it over to the compiler writer. It is not necessary for the
failing code to be small - it just has to be self-contained.
> But the number willing to list known bugs seems to be smaller
> now, perhaps because they don't like washing their dirty linen in public.
> But I have much more confidence in vendors who are willing to admit to
> their (fixed) past mistakes, than in those who prefer not to admit to
> having any.
Well, I can only encourage you to try GNU Fortran 77. We have a
completely open bug report database (see the entry "GNATS" on the GCC
home page http://gcc.gnu.org). For fixed past mistakes you can just
look at the testsuite ...
<evil grin>
I'm sorry, Clive - I'm sympathetic to your wishes, but based on my
experience they're completely unrealistic.
--
Toon Moene - mailto:[log in to unmask] - phoneto: +31 346 214290
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
GNU Fortran 95: http://g95.sourceforge.net/ (under construction)
|