[cfe-dev] (Partial) Success compiling Adium
David Smith
catfish.man at gmail.com
Tue Feb 10 17:32:38 PST 2009
Hmmm... it's possible it's actually working, but not helping as much
as it helps GCC. I reran a number of times just now with:
time xcodebuild -parallelizeTargets CC=$CCC LD=$CCC > /dev/null
vs
time xcodebuild -parallelizeTargets CC=$CCC LD=$CCC
GCC_PRECOMPILE_PREFIX_HEADER=NO > /dev/null
The typical results were around 28s for the first and 33s for the
latter. GCC was about 15s (I didn't have the redirect to /dev/null
before, so I suspect the 19s time earlier was largely due to
xcodebuild spamming my terminal).
David
On Feb 10, 2009, at 5:20 PM, Daniel Dunbar wrote:
> Nice!
>
> Any idea what is wrong with PTH? The driver transparently tries to
> translate the gcc PCH options into PTH compatible options, so this
> should "just work". Feel free to file a bug if you think it is a
> driver issue; if you attach a log file of the xcodebuild (preferably
> with run with
> CCC_ADD_ARGS=-ccc-echo
> also set in the environment) I can investigate.
>
> - Daniel
>
> On Tue, Feb 10, 2009 at 3:37 PM, David Smith <catfish.man at gmail.com>
> wrote:
>> After andersca's exception handling changes recently,
>> AIUtilities (a
>> framework of various GUI widgets, categories, and utility classes
>> that
>> we use in Adium) compiles successfully with clang. This time I
>> doublechecked that it was actually running clang before getting
>> excited ;)
>>
>> Timing results indicate that I don't have PTH working yet, but
>> ignoring that it's looking good:
>> GCC with PCH: real 0m19.376s
>> GCC without PCH: real 0m50.649s
>> CCC without PTH: real 0m33.290s
>>
>> The main Adium application isn't compiling yet, but I'm not
>> entirely
>> certain why. It seems to be running into assertion failures in
>> xcodebuild among other things, which makes it hard to determine
>> what's
>> actually going on.
>>
>>
>> David
>> _______________________________________________
>> cfe-dev mailing list
>> cfe-dev at cs.uiuc.edu
>> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev
>>
More information about the cfe-dev
mailing list