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: Install four PCH files instead of one


In article <20030416205137 dot 7d5193b3 dot bkoz at redhat dot com>,
Benjamin Kosnik<bkoz at redhat dot com> writes:

> Yeah. Sadly, the patch posted last night redid this completely. I'm
> assuming things are mostly working on BSD? Should I just check that in?

Yes, it is basically working on BSD (all remaining bugs are
non-critical; and reporting them daily will force us to fix them).
Yes, AFAIC, please check in your patch.  I'll tweak it with the agreed
4-shot update.

> I think the fallout will be pretty minimal, but I'd like to get a round
> of testing in before Friday and the weekend sets in.

Agreed.

>> libstdc++-v3): -O0, -O2, "-O0 -g", "-O2 -g".

> Hmmm. Had not thought of this. 

I had to read the new PCH sections of the gcc manual to see how do to
it.  (Every occasional gcc developer, such as myself, should reread
the entire gcc manual every now and again. ;-)

To ensure that flags are properly overrode in all cases, I changed the
list to: "-O0 -g0", "-O2 -g0", "-O0 -g", "-O2 -g".

>> Also, note, please don't touch something in ${glibcpp_srcdir}.
> I think the stamp file thing in general is a last resort. I redid the
> patch to remove this part entirely. The 'touch ${pch_output}' was

ACK.  FYI, I bungled it: 'touch ${pch_input};' wasn't touching
${glibcpp_srcdir}.  Please forgive me for making a dumb observation.
How about this (perhaps better than adding -Winvalid-pch when you have
multiple PCH options to reduce the noise; and documented as a valid
technique):

	echo '#error Failure to use PCH file as requested.' > ${pch_input}; \

A #warning might be better still.  Perhaps, when multiple PCH files
are consider, no reason should be presented why 3/4 don't work.  If
4/4 don't work then dump all reasons.

>> (I'll update it and commit it once I see yours go in and you say you
>> agree; or else free feel to include it in your before you first commit.)

> Let's do it after?

Ah, I never reported to the public list.  I see a similar speedup on
FreeBSD (testsuite run is ~twice as fast).  I have a patch from Geoff
(only required for dwarf2 machines) to fix some but not all remaining
current PCH issues visible on this port.

Regards,
Loren


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