This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ 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: [patch] debug mode


Hi

Here is my new patch proposition with the modifications to the unit tests and performance tests. As signaled in a previous mail I kept the __f and __l for the moment to respect the 80 columns line width.

Regards


On 08/18/2010 11:17 PM, Paolo Carlini wrote:
Hi,
No, there were AFAIR 22 failures, the same one with or without the
tests. Do you mean that I should undef _GLIBCXX_DEBUG when the test
don't work with it ? If so I could perhaps do so in an other patch
dedicated to fixing the debug-check target.
Ok, if you see only the synopsis.cc, the pb_ds (there are some latent
problems there), and the ext/profile failures, that's fine in
debug-mode. Actually, I should clean the ext/profile fails, in
particular profiler_algos.cc, which is new, nothing to do with your
work, anyway.
- 80 columns

      Sorry about that I looked for the number of chars and not the
number of columns, because of tabulations it is not equal. By the way,
are tabulations ok ? Most of projects I work on was forbidding tabs,
should I consider that tabs are 8 spaces long ? To fix this problem I
replace many __first, __last by respectively __f, __l, I hope it is
fine.

Certainly tabs are not forbidden, but the indentations must be right, 2
chars wide each, essentially.
Really, with such a width indentation is not right in files I work on,
with width 8 it looks better.
Well, don't ask my opinion here: I'm so used to the GNU, and libstdc++
in particular, conventions, that *anything* else looks ugly to me ;)

Paolo

Attachment: debug.patch
Description: Text document


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