Pavel's diary

Issue #i73858# is used for tracking 2.2 related status - it references 7 blockers, but we still have 24 issues with target 2.2 (including the tracking bug and QA tasks).

Other interesting numbers follow.

Issues with target 2.2.1: 24 issues.
Issues with target 2.3: 236 issues.
Issues with target 2.x: 1908 issues.
Issues with target 3.0: 155 issues.
Issues with target OOoLater: 7554 issues.
Issues with target OOoPleaseHelp: 70 issues.

We need more discipline in managing issues and we need more developers.

I'm still not completely OK (as well as my wife and my son :-(), so I'm at home these days. I was able to do some real work today, so I tried to workaround X11's offscreen rendering bug on Mac OS X (OOo's issue #i68019#). First patch proves that we are able to workaround the bug.

The real question is why it works correctly in xterm, xlogo and similar applications... I have to test it in some graphical application as well.

Starting from today, the channel #openoffice.org where all developers meet is now renamed to #dev.openoffice.org so the original channel is freed for generic user support channel.

Please join us on #dev.openoffice.org!

The winter is weird this year. No snow, 12 deg. Celsius in the morning and similar unusual stuff.

We took the chance to do wonderful though small trip to Prague, to the hill Petřín. My son especially enjoyed local funicular, parents and friends enjoyed lunch at Cantina followed by desert at Nebozízek.

Hamburg RE has just branched OOF680 from the SRC680 development tree and we are going into the stabilization phase just before releasing 2.2. OOF680_m1 was announced as ready (contains new module xt). Rüdiger announced rules for nominations of child workspaces for 2.2 (see dev@ mailing list archive).

Current release plan is available at our Wiki. But we still have 120 issue with target 2.2 (many of them are missing localizations though). We still have 44 child workspaces to be integrated...

We are approaching the release phase of OpenOffice.org 2.2.

The current release plan is available here.

We currently have 208 issues with target 2.2. A bit too much. At least according to our typical numbers...

Please review your issues and especially those with target set to 2.2.

I took the chance of early availability of SRC680_m199 milestone to do some cleanup and measurements on my build system.

First some data about cleanup: I removed approx 130GB of .ccache data from all my systems using ccache (both GNU/Linux boxes, PowerPC Mini and MacBook Pro). The removal of 37GB directory on my Mini took approximately 4 hours and as it was not finished in time, I renamed the directory and will remove it later ;-)

I'll add some data about all builds when they finish into this blog entry. I was building with no ccache data, 28 languages (only en-US, cs and da on Solaris/SPARC because of storage capacity).

Mac OS X/Intel, X11 build

The fastest machine I have now. My laptop, ugh.

Build time without ccache: 12 hours 13 minutes
Build time with ccache: 8 hours 13 minutes
.ccache size after the build: 444 MB
.ccache data after the first build: cache hit 343, cache miss 13905, files in cache 27810
.ccache data after the second build: cache hit +13958, cache miss +290, files in cache +580
du -sh of the complete build tree (ooo_...): 18 GB
du -sh of the solver: 960 MB
du -sh of the data to be uploaded: 4.1 GB

The above data are for building all languages in my build system. I collected the data for one language (en-US) as well: 6 hours 49 minutes with cold ccache, 3 hours 51 minutes with hot ccache.

GNU/Linux x86

One of the oldest machines I have - PIV 2.8 GHz, 1G.

Build time without ccache: 14 hours 18 minutes
Build time with ccache: 8 hours 53 minutes
.ccache size after the build: 871 MB
.ccache data after the first build: cache hit 198, cache miss 14681, files in cache 29362
.ccache data after the second build: ccache hit +14579, cache miss +299, files in cache +598
du -sh of the complete build tree (ooo_...): 19 GB
du -sh of the solver: 954 MB
du -sh of the data to be uploaded: 7.8 GB (RPMS and DEBS)

The above data are for building all languages in my build system. I collected the data for one language (en-US) as well: 7 hours 48 minutes with cold ccache, 1 hour 23 minutes with hot ccache.

GNU/Linux x86_64

Dual Opteron 242 (1600MHz) machine, 2GB.

Build time without ccache: 15 hours 5 minutes
Build time with ccache: 10 hours 34 minutes
.ccache size after the build: 884 MB
.ccache data after the first build: cache hit 167, cache miss 14619, files in cache 29238
.ccache data after the second build: cache hit +14468, cache miss +299, files in cache +598
du -sh of the complete build tree (ooo_...): 20 GB
du -sh of the solver: 1.1 GB
du -sh of the data to be uploaded: 7.8 GB (RPMS and DEBS)

The above data are for building all languages in my build system. I collected the data for one language (en-US) as well: 7 hours with cold ccache, 1 hour 35 minutes with hot ccache.

Windows

The same machine (different instance of the same configuration ;-) as I use for GNU/Linux x86.

Build time: 22 hours 37 minutes
du -sh of the complete build tree (ooo_...): 19 GB
du -sh of the solver: 1.9 GB
du -sh of the data to be uploaded: 3.4 GB

Solaris/SPARC

I forgot what type of the machine it is 8)

Build time: 30 hours
du -sh of the complete build tree (ooo_...): 14 GB
du -sh of the solver: 627 MB
du -sh of the data to be uploaded: 490 MB

Mac OS X/PowerPC

Mac Mini/PowerPC, 512 MB.

Build time without ccache: 3 days 11 hours 20 minutes
Build time with ccache: 18 hours 47 minutes
.ccache size after the build: 515 MB
.ccache data after the first build: cache hit 344, cache miss 13905, files in cache 27810
.ccache data after the second build: ccahe hit +13959, ccache miss +290, files in cache +580
du -sh of the complete build tree (ooo_...): 18G
du -sh of the solver: 986 MB
du -sh of the data to be uploaded: 4.1 GB

As I wrote above, I'll update with the data from other builds as soon as I have them.