This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [C/C++] PR 13358 long long and C++ do not mix well


2009/4/20 Mark Mitchell <mark@codesourcery.com>:
>
> It's in no way material to this patch, but I wish you would not put so
> much of the error text into test-cases. ?I realize that you feel it's
> important to test the exact message, but I feel that this just creates a
> disincentive to make small changes to improve the text of a message.
>

After fixing the C testsuite to distinguish between errors and
warnings, I have the completely opposite opinion. If one wants to
change an error text, it is better to have as much text as possible to
do a find + sed and replace everything in one go. If one does not have
any text, a logic change can start producing nonsense and no one will
notice.

We already have testcases (particularly in g++.old-deja) that do not
match what they were supposed to match. Ask Janis how many C++ tests
she had to guess or just ignore what it was being tested when she
fixed the C++ testsuite to distinguish between errors and warnings.

The largest disincentive to make small changes to improve the text of
a message is the waiting and the slow discussion about the merits and
demerits of the change. People do not have enough patience to follow
up such small changes (c.f. the circular dependency patch+thread which
is still unresolved). On the other hand, spelling changes are quickly
reviewed and committed.

Nonetheless, you are the maintainer, so if you want me to cut the test
pattern, tell me what I should match and I will update the patch.

Cheers,

Manuel.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]