{"thread":{"id":"32481","subject":"Heads up, an emergency fix for git-cvsimport is coming shortly","startedAt":"2012-12-30T19:21:16Z","lastAt":"2012-12-31T01:23:44Z","messageCount":3,"participants":["Eric S. Raymond","Chris Rorvick"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"205683","messageId":"20121230192116.C2A2444143@snark.thyrsus.com","threadId":"32481","inReplyTo":null,"subject":"Heads up, an emergency fix for git-cvsimport is coming shortly","fromName":"Eric S. Raymond","fromEmail":"esr@thyrsus.com","sentAt":"2012-12-30T19:21:16Z","receivedAt":"2012-12-30T19:21:16Z","isPatch":false,"sender":{"key":"esr@thyrsus.com","avatar":"https://avatars.githubusercontent.com/u/727961?v=4"},"body":"Bad news: the combination of cvsps and the existing git-cvsimport\nscript is seriously broken in both places. This morning I fixed a\nnasty bug in cvsps's branch detection and shipped 3.3. This is a\ndifferent bug from the broken (and now removed) ancestry-branch\ntracking.\n\nGood news: I have fixed all the urgent bugs (and now you know how I\nspent my holidays). Somewhat to my surprise, half the problems listed\non the git-cvsimport manual page turned out to be problems in\ngit-cvsimport itself, not more cvsps lossage. Those bugs are dead.\n\ncvsps is now much better about warning when it cannot translate a tag\nor sees a dubious branch structure. I've also enhanced git-cvsimport\nto have an engine switch so it can optionally use cvs2git as its \nconversion engine. If and when I can get parsecvs back into working\nshape, I will add it to the set of supported engines.\n\nI have a test suite that proves fixes for all the urgent problems, but\nthat needs a bit more work before I'm willing to call it done.\n\nIn a few days I will ship a patch that replaces git-cvsimport with a\nworking version and removes the t960[123] tests from the git tree.\nThose are not actually tests of git-cvsimport itself but of the\nunderlying conversion engine, and now form about half of cvsps's own\nregression-test suite.\n-- \n\t\tEric S. Raymond\n\nIt is proper to take alarm at the first experiment on our\nliberties. We hold this prudent jealousy to be the first duty of\ncitizens and one of the noblest characteristics of the late\nRevolution. The freemen of America did not wait till usurped power had\nstrengthened itself by exercise and entangled the question in\nprecedents. They saw all the consequences in the principle, and they\navoided the consequences by denying the principle. We revere this\nlesson too much ... to forget it\t-- James Madison.\n"},{"id":"205688","messageId":"CAEUsAPZ7kzc4qYSvD-YCk9sqQOuW219gOWyxpGqfkxmF2VC-PQ@mail.gmail.com","threadId":"32481","inReplyTo":"20121230192116.C2A2444143@snark.thyrsus.com","subject":"Re: Heads up, an emergency fix for git-cvsimport is coming shortly","fromName":"Chris Rorvick","fromEmail":"chris@rorvick.com","sentAt":"2012-12-31T00:15:41Z","receivedAt":"2012-12-31T00:15:41Z","isPatch":false,"sender":{"key":"chris@rorvick.com","avatar":"https://avatars.githubusercontent.com/u/824726?v=4"},"body":"On Sun, Dec 30, 2012 at 1:21 PM, Eric S. Raymond wrote:\n> Bad news: the combination of cvsps and the existing git-cvsimport\n> script is seriously broken in both places. This morning I fixed a\n> nasty bug in cvsps's branch detection and shipped 3.3. This is a\n> different bug from the broken (and now removed) ancestry-branch\n> tracking.\n\nI tried the new version and found I'm unable to import via pserver:\n\n $ ./cvsps --root :pserver:me@localhost:/cvsroot module\n cvsps: connect error: Connection refused\n cvsps: can't get CVS log data: Connection refused\n\nRunning 2.2b1 (the version packaged w/ Fedora 17) with the same\narguments with the addition of --cvs-direct connects OK. I haven't\ntaken much time to look into this, so I might be doing something dumb.\n Thought I'd find out if this is a known issue before delving into it.\n\nAlso, I'm curious what impact removing the caching from cvsps will\nhave on incremental imports. Is there any?\n\nThanks,\n\nChris Rorvick\n"},{"id":"205689","messageId":"20121231012344.GA17361@thyrsus.com","threadId":"32481","inReplyTo":"CAEUsAPZ7kzc4qYSvD-YCk9sqQOuW219gOWyxpGqfkxmF2VC-PQ@mail.gmail.com","subject":"Re: Heads up, an emergency fix for git-cvsimport is coming shortly","fromName":"Eric S. Raymond","fromEmail":"esr@thyrsus.com","sentAt":"2012-12-31T01:23:44Z","receivedAt":"2012-12-31T01:23:44Z","isPatch":false,"sender":{"key":"esr@thyrsus.com","avatar":"https://avatars.githubusercontent.com/u/727961?v=4"},"body":"Chris Rorvick :\n> I tried the new version and found I'm unable to import via pserver:\n> \n> $ ./cvsps --root :pserver:me@localhost:/cvsroot module\n> cvsps: connect error: Connection refused\n> cvsps: can't get CVS log data: Connection refused\n> \n> Running 2.2b1 (the version packaged w/ Fedora 17) with the same\n> arguments with the addition of --cvs-direct connects OK. I haven't\n> taken much time to look into this, so I might be doing something dumb.\n> Thought I'd find out if this is a known issue before delving into it.\n\nYour problem does reproduce here. This paragraph from the output of \n'aptitude show cvs' may be relevant:\n\n This package contains a CVS binary which can act as both client and server,\n although there is no CVS dæmon; to access remote repositories, please use\n :extssh: not :pserver: any more.\n\nIt's therefore possible there's something slightly busted about the pserver \nmethod at the CVS end, and the 3.[23] code trips over it even though the old\ncode did not. Note that new cvsps uses cvs-direct mode all the time; the old\nsupport for fetching logs through local CVS commands is gone.\n\nI use \n\n cvsps --root :local:$PWD/repo module\n\nfor my testing, and that works. I'm up to my ears in finishing up the\ntest suite and tracking bugs in the repo-analysis code; if you want to\nspeed the process up, try running a :pserver: fetch with -v on under\nboth old and new code to see how the protocol transactions differ.\n\n> Also, I'm curious what impact removing the caching from cvsps will\n> have on incremental imports. Is there any?\n\nNot that I know of. The caching was a performance hack for human viewing\nof changesets.\n-- \n\t\tEric S. Raymond\n"}]}