Eclipse 3.5.1

[Stephane E. Potvin]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Vladimir Grebenschikov wrote:
> Hi 
> 
> Now it verify destribution, but fails later:
> 
> 
> # make
> 
>  eclipse is using libxul for gecko support, but you can
>  change that by defining WITH_GECKO to the following values:
> 
>    libxul 
> 
> ===>  Found saved configuration for eclipse-3.5.1.M20090917.0800
> ===>  Extracting for eclipse-3.5.1.M20090917.0800
> => MD5 Checksum OK for eclipse/eclipse-build-0_4_RC4.tar.gz.
> => SHA256 Checksum OK for eclipse/eclipse-build-0_4_RC4.tar.gz.
> => MD5 Checksum OK for eclipse/eclipse-R3_5_1-fetched-src.tar.bz2.
> => SHA256 Checksum OK for eclipse/eclipse-R3_5_1-fetched-src.tar.bz2.
> ===>  Patching for eclipse-3.5.1.M20090917.0800
> ===>   eclipse-3.5.1.M20090917.0800 depends on executable: ant - found
> ===>   eclipse-3.5.1.M20090917.0800 depends on executable: gpatch - found
> ===>   eclipse-3.5.1.M20090917.0800 depends on executable: zip - found
> ===>  Applying FreeBSD patches for eclipse-3.5.1.M20090917.0800
> Picked up _JAVA_OPTIONS: -Xmx2048m
> Error occurred during initialization of VM
> Could not reserve enough space for object heap
> Could not create the Java virtual machine.
> *** Error code 1
> 
> Stop in /usr/ports/java/eclipse.
> *** Error code 1
> 
> Stop in /usr/ports/java/eclipse.
> # swapinfo 
> Device          512-blocks     Used    Avail Capacity
> /dev/ad0s2b        4194304        0  4194304     0%
> /dev/md0           2097152        0  2097152     0%
> Total              6291456        0  6291456     0%
> #
> 
> It is i386 system with 2Gb RAM, but it should be enough RAM+swap (5 GB total)
> 
> after change to -Xmx1024m it start building (will see results later)
> 
> 
> PS. second run of make after that error gives error about already
> applied patches (need to clean work to fix)
> 
Hi Vladimir,

I guess that asking java to use a 2Gb heap was a little aggressive... I've updated my patch locally to use 1Gb. I just wonder why it
didn't fail in my tinderbox. I would have expected a 32-bit executable to have the same data segment limitations whether the host is
64-bit or not.

Thanks for the report!

Regards,

Steph
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.14 (FreeBSD)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAks708EACgkQmdOXtTCX/numygCgp6FOLfArIUCKvpzbgta11skb
ijUAoN5nP5HGbniOnvfXxxGQif1pxLGx
=JiSf
-----END PGP SIGNATURE-----