{"thread":{"id":"37863","subject":"Re: [PATCH v2] git-svn: use SVN::Ra::get_dir2 when possible","startedAt":"2014-10-31T21:53:54Z","lastAt":"2014-11-01T06:13:14Z","messageCount":5,"participants":["Hin-Tak Leung","Eric Wong"],"isPatch":true,"patchVersion":2,"patchTotal":null},"messages":[{"id":"251277","messageId":"1414792434.54806.YahooMailBasic@web172304.mail.ir2.yahoo.com","threadId":"37863","inReplyTo":null,"subject":"Re: [PATCH v2] git-svn: use SVN::Ra::get_dir2 when possible","fromName":"Hin-Tak Leung","fromEmail":"htl10@users.sourceforge.net","sentAt":"2014-10-31T21:53:54Z","receivedAt":"2014-10-31T21:53:54Z","isPatch":true,"sender":{"key":"htl10@users.sourceforge.net","avatar":null},"body":"-------------------------\nOn Fri, Oct 31, 2014 10:34 GMT Eric Wong wrote:\n\n>This avoids the following failure with normal \"get_dir\" on newer\n>versions of SVN (tested with SVN 1.8.8-1ubuntu3.1):\n>\n>  Incorrect parameters given: Could not convert '%ld' into a number\n>\n>get_dir2 also has the potential to be more efficient by requesting\n>less data.\n>\n>ref: <1414636504.45506.YahooMailBasic@web172304.mail.ir2.yahoo.com>\n>ref: <1414722617.89476.YahooMailBasic@web172305.mail.ir2.yahoo.com>\n>\n>Signed-off-by: Eric Wong \n>Cc: Hin-Tak Leung \n\nTested-by: Hin-Tak Leung \n\nOkay, this one on top of my \"git 2.1.0 + 10 recent git svn improvement patches\"\nallow me to fetch further.\n\nI suspect the problem must be elsewhere though, and this just band-aided\nover it.\n\nFor me, reverting the additional patches and just use git 2.1.0 also allow\nme to fetch further, so the problem appear to me to be associated with\nthe additional patches. But you are saying that git-svn on debian as shipped\nexhibit this problem? \n\n>---\n> Sorry, waaay past my bed time.  This version doesn't infinite loop\n> on autoload or older SVN(*) (at least I hope it doesn't)\n>\n> perl/Git/SVN/Ra.pm | 12 +++++++++++-\n> 1 file changed, 11 insertions(+), 1 deletion(-)\n>\n>diff --git a/perl/Git/SVN/Ra.pm b/perl/Git/SVN/Ra.pm\n>index 82d6108..496f430 100644\n>--- a/perl/Git/SVN/Ra.pm\n>+++ b/perl/Git/SVN/Ra.pm\n>@@ -177,7 +177,17 @@ sub get_dir {\n>         }\n>     }\n>     my $pool = SVN::Pool->new;\n>-    my ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);\n>+    my ($d, undef, $props);\n>+\n>+    if (::compare_svn_version('1.4.0') >= 0) {\n>+        # n.b. in addition to being potentially more efficient,\n>+        # this works around what appears to be a bug in some\n>+        # SVN 1.8 versions\n>+        my $kind = 1; # SVN_DIRENT_KIND\n>+        ($d, undef, $props) = $self->get_dir2($dir, $r, $kind, $pool);\n>+    } else {\n>+        ($d, undef, $props) = $self->SUPER::get_dir($dir, $r, $pool);\n>+    }\n>     my %dirents = map { $_ => { kind => $d->{$_}->kind } } keys %$d;\n>     $pool->clear;\n>     if ($r != $cache->{r}) {\n>-- \n>EW\n>\n"},{"id":"251278","messageId":"20141031232222.GA30046@dcvr.yhbt.net","threadId":"37863","inReplyTo":"1414792434.54806.YahooMailBasic@web172304.mail.ir2.yahoo.com","subject":"Re: [PATCH v2] git-svn: use SVN::Ra::get_dir2 when possible","fromName":"Eric Wong","fromEmail":"normalperson@yhbt.net","sentAt":"2014-10-31T23:22:22Z","receivedAt":"2014-10-31T23:22:22Z","isPatch":true,"sender":{"key":"e@80x24.org","avatar":null},"body":"Hin-Tak Leung wrote:\n> Tested-by: Hin-Tak Leung \n> \n> Okay, this one on top of my \"git 2.1.0 + 10 recent git svn improvement patches\"\n> allow me to fetch further.\n> \n> I suspect the problem must be elsewhere though, and this just band-aided\n> over it.\n> \n> For me, reverting the additional patches and just use git 2.1.0 also allow\n> me to fetch further, so the problem appear to me to be associated with\n> the additional patches.\n\nJunio already pulled all of those git-svn patches into git 2.2.0-rc0:\n\nhttp://mid.gmane.org/xmqq7fzgt0q4.fsf@gitster.dls.corp.google.com\n\nCan you give that a try? I've been able to complete the vbox\nfetch without problems.\n\n> But you are saying that git-svn on debian as shipped\n> exhibit this problem? \n\nNo, I'm saying unrelated, (old) functionality such as\n\"git svn show-ignore\" also got broken by this SVN bug.\n"},{"id":"251280","messageId":"1414808346.73568.YahooMailBasic@web172303.mail.ir2.yahoo.com","threadId":"37863","inReplyTo":"20141031232222.GA30046@dcvr.yhbt.net","subject":"Re: [PATCH v2] git-svn: use SVN::Ra::get_dir2 when possible","fromName":"Hin-Tak Leung","fromEmail":"htl10@users.sourceforge.net","sentAt":"2014-11-01T02:19:06Z","receivedAt":"2014-11-01T02:19:06Z","isPatch":true,"sender":{"key":"htl10@users.sourceforge.net","avatar":null},"body":"\n--------------------------------------------\nOn Sat, 1/11/14, Eric Wong wrote:\n \n Hin-Tak Leung \n wrote:\n> > Tested-by: Hin-Tak Leung \n > \n > Okay, this one on\n top of my \"git 2.1.0 + 10 recent git svn improvement\n patches\"\n > allow me to fetch\n further.\n > \n > I\n suspect the problem must be elsewhere though, and this just\n band-aided\n > over it.\n > \n > For me, reverting\n the additional patches and just use git 2.1.0 also allow\n > me to fetch further, so the problem appear\n to me to be associated with\n > the\n additional patches.\n \n > Junio\n already pulled all of those git-svn patches into git\n 2.2.0-rc0:\n \n> http://mid.gmane.org/xmqq7fzgt0q4.fsf@gitster.dls.corp.google.com\n \n> Can you give that a try? \n I've been able to complete the vbox\n fetch without problems.\n\nWhile my 2.10 + 11 patches continue to fetch, where it was stuck, now\nit does \"Couldn't find revmap...\" - also, the single branch clone is doing\nthe 'trunk/branches/... thing - are these supposed to happen?\n\n--stdlayout:\n\n$ git svn fetch --all\nIndex mismatch: d6c75bc195b1daad647322e2cc025bd31265c6b9 != 3927d05f6ab037fcf2b4d964c9633efade037d1b\nrereading a65b5fc0077c2fa80a344833b65ac19ff4ae88b6\n\tM\tsrc/VBox/Main/HostImpl.cpp\nChecking svn:mergeinfo changes since r28923: 1 sources, 1 changed\nCouldn't find revmap for http://www.virtualbox.org/svn/vbox/branches/VBox-3.0\nr28924 = 4e01065451e00c725c88113584461ee3284e3952 (refs/remotes/origin/trunk)\n\nwithout --stdlayout:\n\n$ git svn fetch --all\nIndex mismatch: d6c75bc195b1daad647322e2cc025bd31265c6b9 != 3927d05f6ab037fcf2b4d964c9633efade037d1b\nrereading a65b5fc0077c2fa80a344833b65ac19ff4ae88b6\n\tM\tsrc/VBox/Main/HostImpl.cpp\nChecking svn:mergeinfo changes since r28923: 1 sources, 1 changed\nCouldn't find revmap for http://www.virtualbox.org/svn/vbox/trunk/branches/VBox-3.0\nr28924 = 4e01065451e00c725c88113584461ee3284e3952 (refs/remotes/git-svn)\n \n"},{"id":"251281","messageId":"20141101031352.GA17306@dcvr.yhbt.net","threadId":"37863","inReplyTo":"1414808346.73568.YahooMailBasic@web172303.mail.ir2.yahoo.com","subject":"Re: [PATCH v2] git-svn: use SVN::Ra::get_dir2 when possible","fromName":"Eric Wong","fromEmail":"normalperson@yhbt.net","sentAt":"2014-11-01T03:13:52Z","receivedAt":"2014-11-01T03:13:52Z","isPatch":true,"sender":{"key":"e@80x24.org","avatar":null},"body":"Hin-Tak Leung wrote:\n> While my 2.10 + 11 patches continue to fetch, where it was stuck, now\n> it does \"Couldn't find revmap...\" - also, the single branch clone is doing\n> the 'trunk/branches/... thing - are these supposed to happen?\n\nI'm afraid this is a problem with the vbox repo not publishing anything\nbut trunk, but their svn:mergeinfo refers to non-public branches.\n\nThis shows only changes to trunk:\n\n\tsvn log -v http://www.virtualbox.org/svn/vbox\n\nI do not know what their development process is like. From the looks of\nit, every change is made by a \"vboxsync\" user. I suspect that's driven\nby an automated export script, and the real development happens in a\ndifferent (private) repo...\n"},{"id":"251285","messageId":"1414822394.95382.YahooMailBasic@web172305.mail.ir2.yahoo.com","threadId":"37863","inReplyTo":"20141101031352.GA17306@dcvr.yhbt.net","subject":"Re: [PATCH v2] git-svn: use SVN::Ra::get_dir2 when possible","fromName":"Hin-Tak Leung","fromEmail":"htl10@users.sourceforge.net","sentAt":"2014-11-01T06:13:14Z","receivedAt":"2014-11-01T06:13:14Z","isPatch":true,"sender":{"key":"htl10@users.sourceforge.net","avatar":null},"body":"--------------------------------------------\nOn Sat, 1/11/14, Eric Wong wrote:\n\n Hin-Tak Leung \n wrote:\n> > While my 2.10 + 11 patches\n continue to fetch, where it was stuck, now\n > it does \"Couldn't find\n revmap...\" - also, the single branch clone is doing\n > the 'trunk/branches/... thing - are\n these supposed to happen?\n \n> I'm afraid this is a problem with the vbox\n repo not publishing anything\n but trunk, but\n their svn:mergeinfo refers to non-public branches.\n \n> This shows only changes to\n trunk:\n \n>     svn log -v http://www.virtualbox.org/svn/vbox\n \n> I do not know what their\n development process is like.  From the looks of\n it, every change is made by a\n \"vboxsync\" user.  I suspect that's driven\n by an automated export script, and the real\n development happens in a\n different (private)\n repo...\n \nThey are Sun/Oracle technology so have that sort of style, I think.\n(Think Java and everything to do with Java).\n\nI did get a bit confused about that - their web code broswing interface\nshows a number of merges and deleted branches.\n"}]}