very slow icedtea6 builds

[Andrew John Hughes]

2008/7/2 Andrew Haley <aph at redhat.com>:
> Matthias Klose wrote:
>> Kelly O'Hair schrieb:
>>> The construction of the rt.jar (or maybe rt-orig.jar in the Makefiles)
>>> can be very slow, lots of disk activity in creating it.
>>>
>>> The hotspot build will likely be the CPU intensive one (parallel make,
>>> high opt),
>>> but the rest of the jdk (images in particular) is much more disk intensive.
>>> The faster the disk drive and the faster the local disk, the better.
>>>
>>> The number of bytes just moving from one place on the disk to another
>>> is pretty intense.
>>
>> the build is done on a striped two disk array (WD Raptor 10k disks), so I don't
>> think the build is limited by IO (plus I didn't change the config during the
>> last six months).  Somebody had the suspicion that the build doesn't use the
>> jit, but I didn't find out yet if this is true.
>
> What do you mean by "the building of the jar"?  The building of the components
> of the jar, or only the jar command itself?  This isn't clear.
>
> Andrew.
>

Specifically it's the creation of rt.jar.

Something is clearly wrong, because my java -version output is now:

OpenJDK 64-Bit Core VM (build 1.6.0_0-b10, mixed mode

on x86_64 -- it's usually:

OpenJDK 64-Bit Server VM (build 1.6.0-b08, mixed mode)

(from my installed system version).  My builds have increased about
ten-fold.  I have a feeling
this is related to Gary's recent commit to be honest but I can't see
why just yet. I also
also now get a 'fastdebug' version but that may be a side-effect of
make distcheck.
-- 
Andrew :-)

Support Free Java!
Contribute to GNU Classpath and the OpenJDK
http://www.gnu.org/software/classpath
http://openjdk.java.net

PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
Fingerprint: F8EF F1EA 401E 2E60 15FA 7927 142C 2591 94EF D9D8