Joined: May 25, 2003 Posts: 4752 Location: Santa Barbara, CA
Posted: Tue Sep 30, 2003 11:32 pm Post subject: Beginning move to 1.1 source on HEAD
Ok, Dan, I started doing it even though I haven't finished going through applying patches and the like. For those who are curious...
The eventual migration strategy that was decided upon was the following:
1) Determine a set of patches that contain the changes applied in the Neo source base from OOo 1.0.1 (OOo1.0.1->Neo0.0.1 patch set)
2) Import a fresh copy of OOo 1.1 source with gcc3 compile support into the Neo CVS repository
3) Apply the Neo patches back into the source.
There turned out to be an enormous amount of code drift within the OpenOffice.org source from 1.0.x to 1.1, and this made my original approach of generating OOo 1.0 -> OOo 1.1 patches and integrating with the older Neo sources really time consuming.
All work on the 1.1 migration is occurring on HEAD, so if you're trying to grab source that's buildable you definitely want to use the "Pre11Source" tag that roughly contains the NeoOffice IG code.
Thus far, on HEAD, old 1.0.x files that are not in 1.1 have been removed. I'll post again when I've finished committing the new 1.1 code (may be tomorrow or Thurs. as I have pressing work getting preso ready for conference deadline)
Joined: May 25, 2003 Posts: 4752 Location: Santa Barbara, CA
Posted: Wed Oct 01, 2003 9:35 pm Post subject:
So you're the one griding the CVS server down
(cooking dinner)
Dan, just be aware that the update you're doing right now will leave everything in an intermediate state...it'll remove the 1.0.x files that aren't in 1.1, but it *won't* put in the new 1.1 files since I haven't gotten a chance to commit them yet I am still verifying the magic CVS directory moving voodoo to commit them.
So...you'll have to do yet another update to get the full 1.1 branched source. I'll post here when all of the files are committed. I've got some Fosters, so we'll see if after a few I neglect my osxcon stuff and try to get this junk committed finally
Joined: May 25, 2003 Posts: 4752 Location: Santa Barbara, CA
Posted: Wed Oct 01, 2003 10:46 pm Post subject:
Right now I'm finished CVS adding the directory structure, and am busy CVS adding the files. It may not complete by the time I go to sleep CVS is notoriously slow, even when on a local network...for each add it's opening a new connection, establishing lock file, executing add, closing connection, etc. and there's network traffic for all.
Once the next CVS add step finishes I'll be able to run a CVS commit. It should take about 5-8 hrs. from past experience as it does not only its checks but also merges from 1.0.x to 1.1.x source in its internal RCS files.
CVS is god awful IO bound on my serer. The bottleneck isn't the processor of the server, but the slow performance of the Ultra66 ATA and WD drive connected to it as it creates and destroys thousands of lock files and directories. I'm sure the blazing performance of NTFS just helps it along. The only time it's really CPU bound is when commits occur and it does the file diffs.
Ah, the joys of CVS. But I will say this...
It's still a hell of a lot faster then sourcesafe.
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum You cannot attach files in this forum You cannot download files in this forum
I am still verifying the magic CVS directory moving voodoo to commit them.
