On 2013-11-14 7:46 PM, Mike Hommey wrote:
On Thu, Nov 14, 2013 at 07:01:03PM -0500, Ehsan Akhgari wrote:
On 2013-11-14 6:29 PM, Chris Peterson wrote:
On 11/14/13, 2:49 PM, Ehsan Akhgari wrote:
The way that unified builds work is
by using the UNIFIED_SOURCES instead of the SOURCES variable in moz.build
files. With that, the build system creates files such as:
// Unified_cpp_path_0.cpp
#include "Source1.cpp"
#include "Source2.cpp"
// ...
Are the UNIFIED_SOURCES from one moz.build file ever unified with
another moz.build file's UNIFIED_SOURCES?
I don't think so, which is good. You don't want to do that since
then adding a file in one directory can break the build in another!
I see that UNIFIED_SOURCES only supports .c and .cpp (and .cc) sources.
Is there any reason why .mm files are not also supported? Do
Objective-C's #imports cause problems?
Ironically I fixed that before seeing your email! See bug 938844.
One piece of interesting data point. I have a whole bunch of
patches locally which makes us build all of layout/ in unified mode.
With those patches applied, and after touching all of the .cpp files
in layout/, I can rebuild that directory in about 4 seconds on my
machine. How long this takes without those patches is left as an
exercise for the readers' imagination. :-)
That sounds incredibly low, which got me curious, so I ran the test,
without unified sources, because if I am to believe my attempts last
year, layout doesn't build this way without modifications.
$ find objdir/layout -name '*.o' -delete
$ time make -s -C objdir/layout -j12 compile
real 1m36.949s
user 11m37.836s
sys 0m32.834s
With ccache and no cache miss.
$ find objdir/layout -name '*.o' -delete
$ time make -s -C objdir/layout -j12 compile
real 0m1.816s
user 0m9.273s
sys 0m2.616s
There is *no way* 1 minute and a half can get down to 4s. So I'll assume
your 4 seconds claim is with ccache with no cache miss. I doubt unified
vs. non unified makes much difference there. But really, ccache is not
the right metric. What's interesting is the actual compilation time, and
I'm sure it goes down. Just not down to 4s.
You're right, I did in fact use ccache because I was using the wrong
mozconfig! (Rookie mistake.) With that fixed, we go down from 1m51s on
my machine to 31s, which is only about 4x faster!
Cheers,
Ehsan
_______________________________________________
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform
