[00:12:10] if you want one sngle commit per delta is quite simple .. give me one day , but repo is ti ok ? [00:13:39] for me the main problem is the branch to find the forking point .. [00:14:17] --- dev-zero@jabber.org has become available [00:40:28] --- dragos.tatulea has become available [00:40:33] hi [00:40:38] hi [00:40:40] --- dragos.tatulea has left [00:40:42] --- dragos.tatulea has become available [00:40:49] hi [00:41:12] test? [00:41:17] hi [00:42:19] hi [00:42:36] http://www.beolink.org/cgi-bin/gitweb.cgi [00:43:00] i'd like to know if the problem is only in the multi commit per DELTA [00:43:19] :-) I need to ask to derrick [01:06:12] --- manfred furuholmen has left [01:07:57] --- manfred furuholmen has become available [01:16:26] --- dev-zero@jabber.org has left [01:43:52] --- dev-zero@jabber.org has become available [01:48:37] --- abo has left [01:51:11] --- Simon Wilkinson has left [02:53:37] --- abo has become available [03:31:52] --- dev-zero@jabber.org has left [04:06:56] --- dev-zero@jabber.org has become available [04:34:40] --- Simon Wilkinson has become available [05:22:55] --- Simon Wilkinson has left [05:41:59] --- Simon Wilkinson has become available [06:01:50] --- Simon Wilkinson has left [06:03:07] --- edgester has become available [06:16:59] --- manfred furuholmen has left [06:22:39] --- manfred furuholmen has become available [08:24:58] --- dev-zero@jabber.org has left [08:57:34] manfred: there are several issues. DELTAs are made up of multiple commits across multiple files. A DELTA can also have multiple commits to the same file which may be interleaved with other DELTAs when the affected range does not overlap. DELTAs also cross branch boundaries in some cases. [09:02:58] --- reuteras has left [09:05:49] i can collect delta in a single commit [09:06:00] (I run now a new import) [09:06:35] but I think you use the same delta name on different branch .. [09:12:23] where there is a DELTA that was committed on separate branches with the same name it is an error that needs to be fixed [09:13:02] UNLESS the DELTA spanned a branch point [09:13:05] thn there are a lot of them i think [09:13:40] DELTAs should be foobar on the head and PREFIX-foobar on the branch [09:14:18] what are you using to import cvs+DELTAs into git? [09:14:58] in the first i have made a tool for a cvps log creation [09:15:40] now i use the log of cvsps with a traslator a collect all the delta .. [09:16:04] the git cvs import can read the cvsps log file [09:16:09] the approach that Derrick and now Mike+Max are following is exporting each DELTA as a patchset. Combining multiple commits to the same file, and then applying the patchset to a new git tree to build up the repository [09:16:53] --- stevegt has left: Replaced by new connection [09:16:53] --- stevegt has become available [09:17:02] the same is made by cvsimport .. it takes the file list from cvsps log file [09:17:19] and it makes cvs export on each file [09:17:39] the problem is that as the process goes on, a mistake in the repository is found such as a incorrect DELTA name, an improperly tagged file, etc.. the repository must be *fixed* and then the process is repeated [09:18:15] from the log of cvsps could be simple .. [09:18:40] Derrick will have to comment. He did 99% of the work [09:19:28] he did a better work ! [09:22:31] --- dragos.tatulea has left [09:22:45] In the repository there are also some file not togged ( like test) [09:24:13] files not tagged [10:09:19] http://www.beolink.org/cgi-bin/gitweb.cgi [10:09:26] with single commit per delta per branch [10:24:05] i... will look. [10:24:27] how did you generate it? [10:24:59] i take the cvsps log file [10:25:28] and i write a small parser for it .. in this way I create a new cvps file [10:25:39] with all delta in the same commit [10:25:53] the rest of the work is made with cvs improt [10:26:05] using which cvsps? [10:26:20] the last with some path around (2.1.2) [10:26:29] if you used git cvsimport, and parsed cvsps, your result is still wrong [10:26:43] binary ? [10:26:46] no [10:27:47] git cvsimport checks out discrete file revisions. we need to apply deltas. otherwise changes get lost. you can verify this simply. check out different versions from your tree and compare them with the same versions from cvs, like, say, 1.4.7 or 1.2.8 [10:27:48] i build a file with the same format of cvsps [10:28:32] that's fine. the issue is not the cvsps file. the issue is how you then processed it [10:29:00] as simply as i can put it: if you used git cvsimport, unmodified from the way it shipped, your result is not useful. [10:29:21] the issue is not what you fed it. the issue is git cvsimport does not cope with openafs deltas. [10:31:44] basically, you don't need me to verify, either. after you convert, diff your trunk to a cvs head. diff your 1.5.55 to cvs's. diff your 1.4.8 to cvs's. and if they are all correct, see if exactly one commit (patchset) and no more has the same of a delta. if "they match" and "one changeset per delta" are all true, it's close enough, and we can run with it [10:33:16] likewise, the ticket in RT where i explain all this is probably a good source of information [10:35:04] but not i need to go out for a while [10:35:07] "now" [10:39:04] i do some check [10:39:16] cu thanks [11:18:50] --- SecureEndpoints has left: Disconnected [11:26:29] --- SecureEndpoints has become available [11:36:05] --- reuteras has become available [11:43:36] --- reuteras has left [12:19:59] --- SecureEndpoints has left: Disconnected [12:24:44] --- SecureEndpoints has become available [12:38:07] --- dragos.tatulea has become available [13:07:33] --- SecureEndpoints has left: Replaced by new connection [13:14:55] --- edgester has left: Replaced by new connection [14:04:42] --- manfred furuholmen has left [14:25:35] --- dragos.tatulea has left [14:57:20] --- SecureEndpoints has become available [15:01:21] --- SecureEndpoints has left [15:57:55] --- edgester has become available [16:05:59] --- SecureEndpoints has become available [16:45:14] --- SecureEndpoints has left: Replaced by new connection [16:51:36] --- Simon Wilkinson has become available [16:59:04] --- SecureEndpoints has become available [17:21:47] --- manfred furuholmen has become available [18:17:20] --- manfred furuholmen has left [20:02:24] --- stevegt has left: Replaced by new connection [20:02:24] --- stevegt has become available [20:06:32] --- edgester has left [22:17:45] --- stevegt has left [22:18:58] --- stevegt has become available