{"thread":{"id":"10886","subject":"[BUG] t9101 (master) busted on Leopard","startedAt":"2007-11-15T13:46:14Z","lastAt":"2007-11-17T13:28:02Z","messageCount":16,"participants":["Wincent Colaiuta","Brian Gernhardt","Benoit Sigoure","Väinö Järvelä","Björn Steinbrink","Junio C Hamano"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"59984","messageId":"D68F81D3-5833-460B-BC7A-98C7E1D8B3E4@wincent.com","threadId":"10886","inReplyTo":null,"subject":"[BUG] t9101 (master) busted on Leopard","fromName":"Wincent Colaiuta","fromEmail":"win@wincent.com","sentAt":"2007-11-15T13:46:14Z","receivedAt":"2007-11-15T13:46:14Z","isPatch":false,"sender":{"key":"greg@hurrell.net","avatar":"https://avatars.githubusercontent.com/u/7074?v=4"},"body":"Was just running the test suite against the master branch and saw that  \nt9101 is currently failing on Leopard, and a review with git-bisect  \nindicates that it has been ever since it was first introduced (in  \ncommit 15153451). Not sure if this problem is Leopard-specific or not  \nas I only have one machine.\n\nThis is the specific test that's failing:\n\n* FAIL 25: test propget\ngit-svn propget svn:ignore . | cmp - prop.expect &&\ncd deeply &&\ngit-svn propget svn:ignore . | cmp - ../prop.expect &&\ngit-svn propget svn:entry:committed-rev nested/directory/.keep \t  |  \ncmp - ../prop2.expect &&\ngit-svn propget svn:ignore .. | cmp - ../prop.expect &&\ngit-svn propget svn:ignore nested/ | cmp - ../prop.expect &&\ngit-svn propget svn:ignore ./nested | cmp - ../prop.expect &&\ngit-svn propget svn:ignore .././deeply/nested | cmp - ../prop.expect\n\nThe problem is that the for line:\n\ngit-svn propget svn:entry:committed-rev nested/directory/.keep\n\nThe test expects the \"8\", but it actually yields \"7\".\n\nI'm not a git-svn user myself, but if there's anything I can do to  \nhelp diagnose this problem further on Leopard please let me know.\n\nCheers,\nWincent\n"},{"id":"59994","messageId":"BB9A8E3F-DC19-4844-80E1-6AEAADF926CD@silverinsanity.com","threadId":"10886","inReplyTo":"D68F81D3-5833-460B-BC7A-98C7E1D8B3E4@wincent.com","subject":"Re: [BUG] t9101 (master) busted on Leopard","fromName":"Brian Gernhardt","fromEmail":"benji@silverinsanity.com","sentAt":"2007-11-15T16:04:41Z","receivedAt":"2007-11-15T16:04:41Z","isPatch":false,"sender":{"key":"benji@silverinsanity.com","avatar":"https://gravatar.com/avatar/e06c101dbc25c68114d859b4a9ec7cf8a2c52fd2b0270ef0eac0e2e63ff22311?d=mp&s=160"},"body":"\nOn Nov 15, 2007, at 8:46 AM, Wincent Colaiuta wrote:\n\n> Was just running the test suite against the master branch and saw  \n> that t9101 is currently failing on Leopard, and a review with git- \n> bisect indicates that it has been ever since it was first introduced  \n> (in commit 15153451). Not sure if this problem is Leopard-specific  \n> or not as I only have one machine.\n\nIt is not a Leopard specific problem, as far as I can tell.  I just  \nran the test and had no errors on my Leopard machine.  So perhaps it's  \nsome other detail of your setup?\n\n> I'm not a git-svn user myself, but if there's anything I can do to  \n> help diagnose this problem further on Leopard please let me know.\n\nI just tested it using svn from fink and (after discovering it exists)  \nfrom Leopard.  No problems.  Do you have an old svn package (client,  \nadmin, or perl binding) installed from Darwin Ports or Fink perhaps?\n\n~~ Brian\n"},{"id":"59995","messageId":"041C0054-5E50-483C-9779-B2FE1AE6947C@wincent.com","threadId":"10886","inReplyTo":"BB9A8E3F-DC19-4844-80E1-6AEAADF926CD@silverinsanity.com","subject":"Re: [BUG] t9101 (master) busted on Leopard","fromName":"Wincent Colaiuta","fromEmail":"win@wincent.com","sentAt":"2007-11-15T16:11:41Z","receivedAt":"2007-11-15T16:11:41Z","isPatch":false,"sender":{"key":"greg@hurrell.net","avatar":"https://avatars.githubusercontent.com/u/7074?v=4"},"body":"El 15/11/2007, a las 17:04, Brian Gernhardt escribió:\n\n> On Nov 15, 2007, at 8:46 AM, Wincent Colaiuta wrote:\n>\n>> Was just running the test suite against the master branch and saw  \n>> that t9101 is currently failing on Leopard, and a review with git- \n>> bisect indicates that it has been ever since it was first  \n>> introduced (in commit 15153451). Not sure if this problem is  \n>> Leopard-specific or not as I only have one machine.\n>\n> It is not a Leopard specific problem, as far as I can tell.  I just  \n> ran the test and had no errors on my Leopard machine.  So perhaps  \n> it's some other detail of your setup?\n>\n>> I'm not a git-svn user myself, but if there's anything I can do to  \n>> help diagnose this problem further on Leopard please let me know.\n>\n> I just tested it using svn from fink and (after discovering it  \n> exists) from Leopard.  No problems.  Do you have an old svn package  \n> (client, admin, or perl binding) installed from Darwin Ports or Fink  \n> perhaps?\n\nI don't use Darwin Ports or Fink, and this is a clean Leopard install  \n(ie. nothing installed in /usr/local apart from git and a very small  \nnumber of other tools that aren't related to Subversion).\n\nThis is the output of \"/usr/bin/svn --version\":\n\nsvn, version 1.4.4 (r25188)\n    compiled Sep 23 2007, 22:32:34\n\nPerhaps then it is something in the environment.\n\nCheers,\nWincent\n"},{"id":"60023","messageId":"1D7CC3C0-46C1-40D9-AAD5-B9ADFF99B58A@lrde.epita.fr","threadId":"10886","inReplyTo":"041C0054-5E50-483C-9779-B2FE1AE6947C@wincent.com","subject":"Re: [BUG] t9101 (master) busted on Leopard","fromName":"Benoit Sigoure","fromEmail":"tsuna@lrde.epita.fr","sentAt":"2007-11-15T21:13:18Z","receivedAt":"2007-11-15T21:13:18Z","isPatch":false,"sender":{"key":"tsunanet@gmail.com","avatar":"https://avatars.githubusercontent.com/u/128281?v=4"},"body":"On Nov 15, 2007, at 5:11 PM, Wincent Colaiuta wrote:\n\n> El 15/11/2007, a las 17:04, Brian Gernhardt escribió:\n>\n>> On Nov 15, 2007, at 8:46 AM, Wincent Colaiuta wrote:\n>>\n>>> Was just running the test suite against the master branch and saw  \n>>> that t9101 is currently failing on Leopard, and a review with git- \n>>> bisect indicates that it has been ever since it was first  \n>>> introduced (in commit 15153451). Not sure if this problem is  \n>>> Leopard-specific or not as I only have one machine.\n>>\n>> It is not a Leopard specific problem, as far as I can tell.  I  \n>> just ran the test and had no errors on my Leopard machine.  So  \n>> perhaps it's some other detail of your setup?\n>>\n>>> I'm not a git-svn user myself, but if there's anything I can do  \n>>> to help diagnose this problem further on Leopard please let me know.\n>>\n>> I just tested it using svn from fink and (after discovering it  \n>> exists) from Leopard.  No problems.  Do you have an old svn  \n>> package (client, admin, or perl binding) installed from Darwin  \n>> Ports or Fink perhaps?\n>\n> I don't use Darwin Ports or Fink, and this is a clean Leopard  \n> install (ie. nothing installed in /usr/local apart from git and a  \n> very small number of other tools that aren't related to Subversion).\n>\n> This is the output of \"/usr/bin/svn --version\":\n>\n> svn, version 1.4.4 (r25188)\n>    compiled Sep 23 2007, 22:32:34\n>\n> Perhaps then it is something in the environment.\n\nHi Wincent,\nCan you reproduce this deterministically?  If yes, can you re-run the  \ntest with the --verbose flag and post the gzipped output (or send it  \nto me if the list doesn't like this sort of attachment).\n\nThanks.\n\n-- \nBenoit Sigoure aka Tsuna\nEPITA Research and Development Laboratory\n\n\n"},{"id":"60038","messageId":"9A0785C3-1150-458C-BC84-1E13CC039D28@wincent.com","threadId":"10886","inReplyTo":"1D7CC3C0-46C1-40D9-AAD5-B9ADFF99B58A@lrde.epita.fr","subject":"Re: [BUG] t9101 (master) busted on Leopard","fromName":"Wincent Colaiuta","fromEmail":"win@wincent.com","sentAt":"2007-11-15T22:22:07Z","receivedAt":"2007-11-15T22:22:07Z","isPatch":false,"sender":{"key":"greg@hurrell.net","avatar":"https://avatars.githubusercontent.com/u/7074?v=4"},"body":"El 15/11/2007, a las 22:13, Benoit Sigoure escribió:\n\n> On Nov 15, 2007, at 5:11 PM, Wincent Colaiuta wrote:\n>\n>> El 15/11/2007, a las 17:04, Brian Gernhardt escribió:\n>>\n>>> On Nov 15, 2007, at 8:46 AM, Wincent Colaiuta wrote:\n>>>\n>>>> Was just running the test suite against the master branch and saw  \n>>>> that t9101 is currently failing on Leopard, and a review with git- \n>>>> bisect indicates that it has been ever since it was first  \n>>>> introduced (in commit 15153451). Not sure if this problem is  \n>>>> Leopard-specific or not as I only have one machine.\n>>>\n>>> It is not a Leopard specific problem, as far as I can tell.  I  \n>>> just ran the test and had no errors on my Leopard machine.  So  \n>>> perhaps it's some other detail of your setup?\n>>>\n>>>> I'm not a git-svn user myself, but if there's anything I can do  \n>>>> to help diagnose this problem further on Leopard please let me  \n>>>> know.\n>>>\n>>> I just tested it using svn from fink and (after discovering it  \n>>> exists) from Leopard.  No problems.  Do you have an old svn  \n>>> package (client, admin, or perl binding) installed from Darwin  \n>>> Ports or Fink perhaps?\n>>\n>> I don't use Darwin Ports or Fink, and this is a clean Leopard  \n>> install (ie. nothing installed in /usr/local apart from git and a  \n>> very small number of other tools that aren't related to Subversion).\n>>\n>> This is the output of \"/usr/bin/svn --version\":\n>>\n>> svn, version 1.4.4 (r25188)\n>>   compiled Sep 23 2007, 22:32:34\n>>\n>> Perhaps then it is something in the environment.\n>\n> Hi Wincent,\n> Can you reproduce this deterministically?  If yes, can you re-run  \n> the test with the --verbose flag and post the gzipped output (or  \n> send it to me if the list doesn't like this sort of attachment).\n\nYes, have just sent you the output of \"--verbose\" and also running the  \nscript using \"sh -x\" (off list).\n\nCheers,\nWincent\n"},{"id":"60050","messageId":"8B2E66EB-22C7-4C08-9947-C0D6A755309F@pp.inet.fi","threadId":"10886","inReplyTo":"041C0054-5E50-483C-9779-B2FE1AE6947C@wincent.com","subject":"Re: [BUG] t9101 (master) busted on Leopard","fromName":"Väinö Järvelä","fromEmail":"v@pp.inet.fi","sentAt":"2007-11-16T04:25:09Z","receivedAt":"2007-11-16T04:25:09Z","isPatch":false,"sender":{"key":"v@pp.inet.fi","avatar":null},"body":"\nOn Nov 15, 2007, at 18:11, Wincent Colaiuta wrote:\n\n> El 15/11/2007, a las 17:04, Brian Gernhardt escribió:\n>\n>> On Nov 15, 2007, at 8:46 AM, Wincent Colaiuta wrote:\n>>\n>>> Was just running the test suite against the master branch and saw  \n>>> that t9101 is currently failing on Leopard, and a review with git- \n>>> bisect indicates that it has been ever since it was first  \n>>> introduced (in commit 15153451). Not sure if this problem is  \n>>> Leopard-specific or not as I only have one machine.\n>>\n>> It is not a Leopard specific problem, as far as I can tell.  I just  \n>> ran the test and had no errors on my Leopard machine.  So perhaps  \n>> it's some other detail of your setup?\n>>\n>>> I'm not a git-svn user myself, but if there's anything I can do to  \n>>> help diagnose this problem further on Leopard please let me know.\n>>\n>> I just tested it using svn from fink and (after discovering it  \n>> exists) from Leopard.  No problems.  Do you have an old svn package  \n>> (client, admin, or perl binding) installed from Darwin Ports or  \n>> Fink perhaps?\n>\n> I don't use Darwin Ports or Fink, and this is a clean Leopard  \n> install (ie. nothing installed in /usr/local apart from git and a  \n> very small number of other tools that aren't related to Subversion).\n>\n> This is the output of \"/usr/bin/svn --version\":\n>\n> svn, version 1.4.4 (r25188)\n>   compiled Sep 23 2007, 22:32:34\n>\n> Perhaps then it is something in the environment.\n\nI cannot reproduce this on Leopard (not yet updated to 10.5.1). I  \ntested this with self-compiled GIT from commit id ca2b71c\n\nsvn version is the same as yours, I do use Fink, and I have used Fink  \nto install Perl SVN bindings.\n\n--\nVäinö"},{"id":"60067","messageId":"E20C625B-100C-44B0-9AF3-7342DAF9B2A7@wincent.com","threadId":"10886","inReplyTo":"8B2E66EB-22C7-4C08-9947-C0D6A755309F@pp.inet.fi","subject":"Re: [BUG] t9101 (master) busted on Leopard","fromName":"Wincent Colaiuta","fromEmail":"win@wincent.com","sentAt":"2007-11-16T08:37:55Z","receivedAt":"2007-11-16T08:37:55Z","isPatch":false,"sender":{"key":"greg@hurrell.net","avatar":"https://avatars.githubusercontent.com/u/7074?v=4"},"body":"El 16/11/2007, a las 5:25, Väinö Järvelä escribió:\n\n> On Nov 15, 2007, at 18:11, Wincent Colaiuta wrote:\n>\n>> I don't use Darwin Ports or Fink, and this is a clean Leopard  \n>> install (ie. nothing installed in /usr/local apart from git and a  \n>> very small number of other tools that aren't related to Subversion).\n>>\n>> This is the output of \"/usr/bin/svn --version\":\n>>\n>> svn, version 1.4.4 (r25188)\n>>  compiled Sep 23 2007, 22:32:34\n>>\n>> Perhaps then it is something in the environment.\n>\n> I cannot reproduce this on Leopard (not yet updated to 10.5.1). I  \n> tested this with self-compiled GIT from commit id ca2b71c\n>\n> svn version is the same as yours, I do use Fink, and I have used  \n> Fink to install Perl SVN bindings.\n\nStrange. I've looked in the environment and there is nothing  \nsuspicious; in fact, running the tests with a totally clean  \nenvironment (env -i ./t9101-git-svn-props.sh) produces exactly the  \nsame result.\n\nThis was with commit 039bc64e (HEAD of master yesterday). I just  \ntested the commit you mention (HEAD of next) and get the same failure;  \nthis is the procedure I'm using to test:\n\ngit fetch\ngit checkout -b next_test origin/next\ngit describe # (v1.5.3.5-1780-gca2b71c)\ngit clean\nmake clean && make\ncd t\nenv -i ./t9101-git-svn-props.sh\n\nSo really, not sure what could be causing this.\n\nCheers,\nWincent\n"},{"id":"60071","messageId":"4207CF02-36BD-4733-9861-17B0C4DB9D94@pp.inet.fi","threadId":"10886","inReplyTo":"E20C625B-100C-44B0-9AF3-7342DAF9B2A7@wincent.com","subject":"Re: [BUG] t9101 (master) busted on Leopard","fromName":"Väinö Järvelä","fromEmail":"v@pp.inet.fi","sentAt":"2007-11-16T08:59:54Z","receivedAt":"2007-11-16T08:59:54Z","isPatch":false,"sender":{"key":"v@pp.inet.fi","avatar":null},"body":"\nOn Nov 16, 2007, at 10:37, Wincent Colaiuta wrote:\n\n> El 16/11/2007, a las 5:25, Väinö Järvelä escribió:\n>\n>> On Nov 15, 2007, at 18:11, Wincent Colaiuta wrote:\n>>\n>>> I don't use Darwin Ports or Fink, and this is a clean Leopard  \n>>> install (ie. nothing installed in /usr/local apart from git and a  \n>>> very small number of other tools that aren't related to Subversion).\n>>>\n>>> This is the output of \"/usr/bin/svn --version\":\n>>>\n>>> svn, version 1.4.4 (r25188)\n>>> compiled Sep 23 2007, 22:32:34\n>>>\n>>> Perhaps then it is something in the environment.\n>>\n>> I cannot reproduce this on Leopard (not yet updated to 10.5.1). I  \n>> tested this with self-compiled GIT from commit id ca2b71c\n>>\n>> svn version is the same as yours, I do use Fink, and I have used  \n>> Fink to install Perl SVN bindings.\n>\n> Strange. I've looked in the environment and there is nothing  \n> suspicious; in fact, running the tests with a totally clean  \n> environment (env -i ./t9101-git-svn-props.sh) produces exactly the  \n> same result.\n>\n> This was with commit 039bc64e (HEAD of master yesterday). I just  \n> tested the commit you mention (HEAD of next) and get the same  \n> failure; this is the procedure I'm using to test:\n>\n> git fetch\n> git checkout -b next_test origin/next\n> git describe # (v1.5.3.5-1780-gca2b71c)\n> git clean\n> make clean && make\n> cd t\n> env -i ./t9101-git-svn-props.sh\n>\n> So really, not sure what could be causing this.\n\nDid you forget to install the newly compiled version? Or does the test  \nsystem use git from the source tree?\n\n--\nVäinö\n"},{"id":"60074","messageId":"EBFD3ABD-CD7F-4B9C-AB9D-F440C368AB61@wincent.com","threadId":"10886","inReplyTo":"4207CF02-36BD-4733-9861-17B0C4DB9D94@pp.inet.fi","subject":"Re: [BUG] t9101 (master) busted on Leopard","fromName":"Wincent Colaiuta","fromEmail":"win@wincent.com","sentAt":"2007-11-16T10:17:03Z","receivedAt":"2007-11-16T10:17:03Z","isPatch":false,"sender":{"key":"greg@hurrell.net","avatar":"https://avatars.githubusercontent.com/u/7074?v=4"},"body":"El 16/11/2007, a las 9:59, Väinö Järvelä escribió:\n\n> On Nov 16, 2007, at 10:37, Wincent Colaiuta wrote:\n>\n>> Strange. I've looked in the environment and there is nothing  \n>> suspicious; in fact, running the tests with a totally clean  \n>> environment (env -i ./t9101-git-svn-props.sh) produces exactly the  \n>> same result.\n>>\n>> This was with commit 039bc64e (HEAD of master yesterday). I just  \n>> tested the commit you mention (HEAD of next) and get the same  \n>> failure; this is the procedure I'm using to test:\n>>\n>> git fetch\n>> git checkout -b next_test origin/next\n>> git describe # (v1.5.3.5-1780-gca2b71c)\n>> git clean\n>> make clean && make\n>> cd t\n>> env -i ./t9101-git-svn-props.sh\n>>\n>> So really, not sure what could be causing this.\n>\n> Did you forget to install the newly compiled version? Or does the  \n> test system use git from the source tree?\n\nI didn't install the newly compiled version because, as far as know,  \nthe test system uses the built copy in the source tree (I can't  \nactually recall *any* Makefile-based open source projects that didn't  \nwork like that).\n\n$ git --version\ngit version 1.5.3.5\n$ /usr/local/bin/git --version\ngit version 1.5.3.5\n$ ./git --version # (from inside the source tree)\ngit version 1.5.3.5.1780.gca2b\n$ env -i git --version\nenv: git: No such file or directory\n\nSo the test run above *must* be using the just-built copy in the  \nsource tree otherwise *none* of the tests would pass.\n\nCheers,\nWincent\n"},{"id":"60078","messageId":"73246E38-9C22-4279-A53E-678434238E5C@wincent.com","threadId":"10886","inReplyTo":"1D7CC3C0-46C1-40D9-AAD5-B9ADFF99B58A@lrde.epita.fr","subject":"Re: [BUG] t9101 (master) busted on Leopard","fromName":"Wincent Colaiuta","fromEmail":"win@wincent.com","sentAt":"2007-11-16T11:00:02Z","receivedAt":"2007-11-16T11:00:02Z","isPatch":false,"sender":{"key":"greg@hurrell.net","avatar":"https://avatars.githubusercontent.com/u/7074?v=4"},"body":"El 15/11/2007, a las 22:13, Benoit Sigoure escribió:\n\n> On Nov 15, 2007, at 5:11 PM, Wincent Colaiuta wrote:\n>\n>> El 15/11/2007, a las 17:04, Brian Gernhardt escribió:\n>>\n>>> On Nov 15, 2007, at 8:46 AM, Wincent Colaiuta wrote:\n>>>\n>>>> Was just running the test suite against the master branch and saw  \n>>>> that t9101 is currently failing on Leopard, and a review with git- \n>>>> bisect indicates that it has been ever since it was first  \n>>>> introduced (in commit 15153451). Not sure if this problem is  \n>>>> Leopard-specific or not as I only have one machine.\n>>>\n>>> It is not a Leopard specific problem, as far as I can tell.  I  \n>>> just ran the test and had no errors on my Leopard machine.  So  \n>>> perhaps it's some other detail of your setup?\n>>>\n>>>> I'm not a git-svn user myself, but if there's anything I can do  \n>>>> to help diagnose this problem further on Leopard please let me  \n>>>> know.\n>>>\n>>> I just tested it using svn from fink and (after discovering it  \n>>> exists) from Leopard.  No problems.  Do you have an old svn  \n>>> package (client, admin, or perl binding) installed from Darwin  \n>>> Ports or Fink perhaps?\n>>\n>> I don't use Darwin Ports or Fink, and this is a clean Leopard  \n>> install (ie. nothing installed in /usr/local apart from git and a  \n>> very small number of other tools that aren't related to Subversion).\n>>\n>> This is the output of \"/usr/bin/svn --version\":\n>>\n>> svn, version 1.4.4 (r25188)\n>>   compiled Sep 23 2007, 22:32:34\n>>\n>> Perhaps then it is something in the environment.\n>\n> Hi Wincent,\n> Can you reproduce this deterministically?  If yes, can you re-run  \n> the test with the --verbose flag and post the gzipped output (or  \n> send it to me if the list doesn't like this sort of attachment).\n\nInspecting the output of --verbose has allowed me to the point where  \nthings are failing; look at the very first test in t9101-git-svn- \nprops.sh and witness the following:\n\n- before the test we are at revision 1\n- there are 3 \"svn commits\" in the test\n- but it looks like only 2 of the commits proceed\n- so the revision number goes up to 2 then 3, when it should be 4\n- as a result, in one of the later tests we fail because the test  \nexpects revision 8 but we only have revision 7\n\nChecked out revision 1.\n*   ok 1: checkout working copy from svn\n\n* expecting success: cd test_wc &&\n\t\techo Greetings >> kw.c &&\n\t\tpoke kw.c &&\n\t\tsvn commit -m \"Not yet an Id\" &&\n\t\techo Hello world >> kw.c &&\n\t\tpoke kw.c &&\n\t\tsvn commit -m \"Modified file, but still not yet an Id\" &&\n\t\tsvn propset svn:keywords Id kw.c &&\n\t\tpoke kw.c &&\n\t\tsvn commit -m \"Propset Id\" &&\n\tcd ..\nSending        kw.c\nTransmitting file data .\nCommitted revision 2.\nSending        kw.c\nTransmitting file data .\nCommitted revision 3.\nproperty 'svn:keywords' set on 'kw.c'\n*   ok 2: setup some commits to svn\n\nThat last commit is a no-op because, for some reason, the svn propset  \nbefore it is also a no-op:\n\nsvn propset svn:keywords Id kw.c\n\nIn other words, it echos \"property 'svn:keyword' set on 'kw.c'\" but if  \nI insert an \"svn status\" as the next command then I get no output at  \nall. If I run the exact same commands manually from the terminal then  \nthey work (ie. it is not a no-op and \"svn status\" shows that the file  \nis modified).\n\nActually, it's not really a no-op, because if I insert an \"svn  \nproplist -v kw.c\" I get:\n\nProperties on 'kw.c':\n   svn:keywords : Id\n\nSo the property is being set, it's just that \"svn commit\" and \"svn  \nstatus\" don't recognize the file as being modified. The \"poke\" in the  \ntest has no effect (file still shows up as unmodified) and only  \nmodifying the actual content (ie. appending to it by inserting another  \n\"echo\" statement) is enough to make that commit actually happen.\n\nSo now we know a bit more about *what* is happening, but I still don't  \nknow *why*. I'd start to suspect some kind of weird filesystem issue  \nif it weren't for the fact that these same commands work when executed  \nby hand outside of the test script.\n\nCheers,\nWincent\n"},{"id":"60083","messageId":"20071116124850.GA14473@atjola.homenet","threadId":"10886","inReplyTo":"73246E38-9C22-4279-A53E-678434238E5C@wincent.com","subject":"Re: [BUG] t9101 (master) busted on Leopard","fromName":"Björn Steinbrink","fromEmail":"b.steinbrink@gmx.de","sentAt":"2007-11-16T12:48:50Z","receivedAt":"2007-11-16T12:48:50Z","isPatch":false,"sender":{"key":"b.steinbrink@gmx.de","avatar":"https://avatars.githubusercontent.com/u/230962?v=4"},"body":"On 2007.11.16 12:00:02 +0100, Wincent Colaiuta wrote:\n> El 15/11/2007, a las 22:13, Benoit Sigoure escribió:\n>> Hi Wincent,\n>> Can you reproduce this deterministically?  If yes, can you re-run the test \n>> with the --verbose flag and post the gzipped output (or send it to me if \n>> the list doesn't like this sort of attachment).\n>\n> Inspecting the output of --verbose has allowed me to the point where things \n> are failing; look at the very first test in t9101-git-svn-props.sh and \n> witness the following:\n>\n> - before the test we are at revision 1\n> - there are 3 \"svn commits\" in the test\n> - but it looks like only 2 of the commits proceed\n> - so the revision number goes up to 2 then 3, when it should be 4\n> - as a result, in one of the later tests we fail because the test expects \n> revision 8 but we only have revision 7\n>\n> Checked out revision 1.\n> *   ok 1: checkout working copy from svn\n>\n> * expecting success: cd test_wc &&\n> \t\techo Greetings >> kw.c &&\n> \t\tpoke kw.c &&\n> \t\tsvn commit -m \"Not yet an Id\" &&\n> \t\techo Hello world >> kw.c &&\n> \t\tpoke kw.c &&\n> \t\tsvn commit -m \"Modified file, but still not yet an Id\" &&\n> \t\tsvn propset svn:keywords Id kw.c &&\n> \t\tpoke kw.c &&\n> \t\tsvn commit -m \"Propset Id\" &&\n> \tcd ..\n> Sending        kw.c\n> Transmitting file data .\n> Committed revision 2.\n> Sending        kw.c\n> Transmitting file data .\n> Committed revision 3.\n> property 'svn:keywords' set on 'kw.c'\n> *   ok 2: setup some commits to svn\n>\n> That last commit is a no-op because, for some reason, the svn propset \n> before it is also a no-op:\n>\n> svn propset svn:keywords Id kw.c\n>\n> In other words, it echos \"property 'svn:keyword' set on 'kw.c'\" but if I \n> insert an \"svn status\" as the next command then I get no output at all. If \n> I run the exact same commands manually from the terminal then they work \n> (ie. it is not a no-op and \"svn status\" shows that the file is modified).\n>\n> Actually, it's not really a no-op, because if I insert an \"svn proplist -v \n> kw.c\" I get:\n>\n> Properties on 'kw.c':\n>   svn:keywords : Id\n\nIt _is_ a no-op. At least here. Because I got an auto-props setting in\nmy ~/.subversion/config to automatically add Id for *.c files. So that\nproperty was already there before we explicitly ask for it and the\npropset turns into a no-op. If I remove that line from the subversion\nconfiguration, the test succeeds. Same for you I guess.\n\nThat said, I had a quick glance over the subversion CLI help, but it\ndidn't tell me how to ignore/override ~/.subversion/config. Anyone less\nclueless than me around, having a smart idea how to work around that\n\"bug\"?\n\nBjörn\n"},{"id":"60086","messageId":"406D1E35-2C24-4D47-9F3E-793EAFA39B0E@wincent.com","threadId":"10886","inReplyTo":"20071116124850.GA14473@atjola.homenet","subject":"Re: [PATCH] t9101 (master) busted on Leopard","fromName":"Wincent Colaiuta","fromEmail":"win@wincent.com","sentAt":"2007-11-16T13:14:24Z","receivedAt":"2007-11-16T13:14:24Z","isPatch":true,"sender":{"key":"greg@hurrell.net","avatar":"https://avatars.githubusercontent.com/u/7074?v=4"},"body":"El 16/11/2007, a las 13:48, Björn Steinbrink escribió:\n> On 2007.11.16 12:00:02 +0100, Wincent Colaiuta wrote:\n>> That last commit is a no-op because, for some reason, the svn propset\n>> before it is also a no-op:\n>>\n>> svn propset svn:keywords Id kw.c\n>\n> It _is_ a no-op. At least here. Because I got an auto-props setting in\n> my ~/.subversion/config to automatically add Id for *.c files. So that\n> property was already there before we explicitly ask for it and the\n> propset turns into a no-op. If I remove that line from the subversion\n> configuration, the test succeeds. Same for you I guess.\n>\n> That said, I had a quick glance over the subversion CLI help, but it\n> didn't tell me how to ignore/override ~/.subversion/config. Anyone  \n> less\n> clueless than me around, having a smart idea how to work around that\n> \"bug\"?\n\nAh, excellent catch, Björn. That's it. The --no-auto-props switch will  \nfix this then; will post a patch to the list in a few minutes.\n\nCheers,\nWincent\n"},{"id":"60087","messageId":"2F7DFDC9-D4E2-42D0-9E48-E51E7905FF42@wincent.com","threadId":"10886","inReplyTo":"20071116124850.GA14473@atjola.homenet","subject":"[PATCH] Fix t9101 test failure caused by Subversion \"auto-props\"","fromName":"Wincent Colaiuta","fromEmail":"win@wincent.com","sentAt":"2007-11-16T13:25:10Z","receivedAt":"2007-11-16T13:25:10Z","isPatch":true,"sender":{"key":"greg@hurrell.net","avatar":"https://avatars.githubusercontent.com/u/7074?v=4"},"body":"If a user has an \"auto-prop\" in his/her ~/.subversion/config file for\nautomatically setting the svn:keyword Id property on all \".c\" files\n(a reasonably common configuration in the Subversion world) then one\nof the \"svn propset\" operations in the very first test would become a\nno-op, which in turn would make the next commit a no-op.\n\nThis then caused the 25th test ('test propget') to fail because it\nexpects a certain number of commits to have taken place but the actual\nnumber of commits was off by one.\n\nBjörn Steinbrink identified the \"auto-prop\" feature as the cause\nof the failure. This patch avoids it by passing the \"--no-auto-prop\"\nflag to \"svn import\" when setting up the test repository, thus ensuring\nthat the \"svn propset\" operation is no longer a no-op, regardless of the\nusers' settings in their config.\n\nSigned-off-by: Wincent Colaiuta <win@wincent.com>\n---\n  t/t9101-git-svn-props.sh |    2 +-\n  1 files changed, 1 insertions(+), 1 deletions(-)\n\ndiff --git a/t/t9101-git-svn-props.sh b/t/t9101-git-svn-props.sh\nindex 3c83127..d7a7047 100755\n--- a/t/t9101-git-svn-props.sh\n+++ b/t/t9101-git-svn-props.sh\n@@ -48,7 +48,7 @@ EOF\n         printf \"\\r\\n\" > empty_crlf\n         a_empty_crlf=`git-hash-object -w empty_crlf`\n\n-       svn import -m 'import for git-svn' . \"$svnrepo\" >/dev/null\n+       svn import --no-auto-props -m 'import for git-svn' .  \n\"$svnrepo\" >/dev/null\n  cd ..\n\n  rm -rf import\n-- \n1.5.3.5\n"},{"id":"60097","messageId":"75D25F7C-EAB4-49ED-80FF-2BC535915585@lrde.epita.fr","threadId":"10886","inReplyTo":"2F7DFDC9-D4E2-42D0-9E48-E51E7905FF42@wincent.com","subject":"Re: [PATCH] Fix t9101 test failure caused by Subversion \"auto-props\"","fromName":"Benoit Sigoure","fromEmail":"tsuna@lrde.epita.fr","sentAt":"2007-11-16T17:56:17Z","receivedAt":"2007-11-16T17:56:17Z","isPatch":true,"sender":{"key":"tsunanet@gmail.com","avatar":"https://avatars.githubusercontent.com/u/128281?v=4"},"body":"On Nov 16, 2007, at 2:25 PM, Wincent Colaiuta wrote:\n\n> If a user has an \"auto-prop\" in his/her ~/.subversion/config file for\n> automatically setting the svn:keyword Id property on all \".c\" files\n> (a reasonably common configuration in the Subversion world) then one\n> of the \"svn propset\" operations in the very first test would become a\n> no-op, which in turn would make the next commit a no-op.\n>\n> This then caused the 25th test ('test propget') to fail because it\n> expects a certain number of commits to have taken place but the actual\n> number of commits was off by one.\n>\n> Björn Steinbrink identified the \"auto-prop\" feature as the cause\n> of the failure. This patch avoids it by passing the \"--no-auto-prop\"\n> flag to \"svn import\" when setting up the test repository, thus  \n> ensuring\n> that the \"svn propset\" operation is no longer a no-op, regardless  \n> of the\n> users' settings in their config.\n>\n> Signed-off-by: Wincent Colaiuta <win@wincent.com>\n> ---\n>  t/t9101-git-svn-props.sh |    2 +-\n>  1 files changed, 1 insertions(+), 1 deletions(-)\n>\n> diff --git a/t/t9101-git-svn-props.sh b/t/t9101-git-svn-props.sh\n> index 3c83127..d7a7047 100755\n> --- a/t/t9101-git-svn-props.sh\n> +++ b/t/t9101-git-svn-props.sh\n> @@ -48,7 +48,7 @@ EOF\n>         printf \"\\r\\n\" > empty_crlf\n>         a_empty_crlf=`git-hash-object -w empty_crlf`\n>\n> -       svn import -m 'import for git-svn' . \"$svnrepo\" >/dev/null\n> +       svn import --no-auto-props -m 'import for git-svn' .  \n> \"$svnrepo\" >/dev/null\n>  cd ..\n>\n>  rm -rf import\n\nGreat, thank you for tackling this issue.  It wasn't easy to find.\n\n-- \nBenoit Sigoure aka Tsuna\nEPITA Research and Development Laboratory\n\n\n"},{"id":"60119","messageId":"7vy7cxwwoo.fsf@gitster.siamese.dyndns.org","threadId":"10886","inReplyTo":"2F7DFDC9-D4E2-42D0-9E48-E51E7905FF42@wincent.com","subject":"Re: [PATCH] Fix t9101 test failure caused by Subversion \"auto-props\"","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2007-11-17T00:19:19Z","receivedAt":"2007-11-17T00:19:19Z","isPatch":true,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Wincent Colaiuta <win@wincent.com> writes:\n\n> If a user has an \"auto-prop\" in his/her ~/.subversion/config file for\n> automatically setting the svn:keyword Id property on all \".c\" files\n> (a reasonably common configuration in the Subversion world) then one\n> of the \"svn propset\" operations in the very first test would become a\n> no-op, which in turn would make the next commit a no-op.\n\nThanks for diagnosing and fixing.\n\nI presume this fix also applies to both 'maint' and 'master',\nright?\n"},{"id":"60139","messageId":"270B4D20-EC93-43C9-8BCD-EAE7018F2039@wincent.com","threadId":"10886","inReplyTo":"7vy7cxwwoo.fsf@gitster.siamese.dyndns.org","subject":"Re: [PATCH] Fix t9101 test failure caused by Subversion \"auto-props\"","fromName":"Wincent Colaiuta","fromEmail":"win@wincent.com","sentAt":"2007-11-17T13:28:02Z","receivedAt":"2007-11-17T13:28:02Z","isPatch":true,"sender":{"key":"greg@hurrell.net","avatar":"https://avatars.githubusercontent.com/u/7074?v=4"},"body":"El 17/11/2007, a las 1:19, Junio C Hamano escribió:\n\n> Wincent Colaiuta <win@wincent.com> writes:\n>\n>> If a user has an \"auto-prop\" in his/her ~/.subversion/config file for\n>> automatically setting the svn:keyword Id property on all \".c\" files\n>> (a reasonably common configuration in the Subversion world) then one\n>> of the \"svn propset\" operations in the very first test would become a\n>> no-op, which in turn would make the next commit a no-op.\n>\n> Thanks for diagnosing and fixing.\n>\n> I presume this fix also applies to both 'maint' and 'master',\n> right?\n\nI prepared it against master, but I believe it will apply cleanly to  \nmaint, where it's also needed.\n\nCheers,\nWincent\n"}]}