{"thread":{"id":"47736","subject":"t3404.6 breaks on master under GIT_FSMONITOR_TEST","startedAt":"2018-02-01T13:03:00Z","lastAt":"2018-10-08T18:27:33Z","messageCount":4,"participants":["Ævar Arnfjörð Bjarmason","Ben Peart"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"338029","messageId":"87k1vwn9qe.fsf@evledraar.gmail.com","threadId":"47736","inReplyTo":null,"subject":"t3404.6 breaks on master under GIT_FSMONITOR_TEST","fromName":"Ævar Arnfjörð Bjarmason","fromEmail":"avarab@gmail.com","sentAt":"2018-02-01T13:02:49Z","receivedAt":"2018-02-01T13:03:00Z","isPatch":false,"sender":{"key":"avarab@gmail.com","avatar":"https://avatars.githubusercontent.com/u/45301?v=4"},"body":"The GIT_FSMONITOR_TEST variable allows you to roundtrip the fsmonitor\ncodpath in the whole test suite. On both Debian & CentOS this breaks for\nme:\n\n    (cd t && GIT_FSMONITOR_TEST=$PWD/t7519/fsmonitor-all ./t3404-rebase-interactive.sh -i)\n\nWhereas this works:\n\n    (cd t && GIT_FSMONITOR_TEST=$PWD/t7519/fsmonitor-all GIT_SKIP_TESTS=t3404.6 ./t3404-rebase-interactive.sh -i)\n\nThe entirety of the rest of the test suite still passes with\nGIT_FSMONITOR_TEST.\n\nThis has been failing ever since GIT_FSMONITOR_TEST was introduced in\n883e248b8a (\"fsmonitor: teach git to optionally utilize a file system\nmonitor to speed up detecting new or changed files.\", 2017-09-22). Under\n-v -x -i:\n\n    + echo test_must_fail: command succeeded: env FAKE_LINES=exec_echo_foo_>file1 1 git rebase -i HEAD^\n    test_must_fail: command succeeded: env FAKE_LINES=exec_echo_foo_>file1 1 git rebase -i HEAD^\n    + return 1\n    error: last command exited with $?=1\n    not ok 6 - rebase -i with the exec command checks tree cleanness\n    #\n    #               git checkout master &&\n    #               set_fake_editor &&\n    #               test_must_fail env FAKE_LINES=\"exec_echo_foo_>file1 1\" git rebase -i HEAD^ &&\n\nMaybe once this is fixed running the test suite under GIT_FSMONITOR_TEST\nwould be a useful Travis target, but I don't know the current status of\nadding new options to Travis.\n"},{"id":"357528","messageId":"87sh2mrefa.fsf@evledraar.gmail.com","threadId":"47736","inReplyTo":"87k1vwn9qe.fsf@evledraar.gmail.com","subject":"Re: t3404.6 breaks on master under GIT_FSMONITOR_TEST","fromName":"Ævar Arnfjörð Bjarmason","fromEmail":"avarab@gmail.com","sentAt":"2018-09-06T12:36:09Z","receivedAt":"2018-09-06T12:36:14Z","isPatch":false,"sender":{"key":"avarab@gmail.com","avatar":"https://avatars.githubusercontent.com/u/45301?v=4"},"body":"\nOn Thu, Feb 01 2018, Ævar Arnfjörð Bjarmason wrote:\n\n> The GIT_FSMONITOR_TEST variable allows you to roundtrip the fsmonitor\n> codpath in the whole test suite. On both Debian & CentOS this breaks for\n> me:\n>\n>     (cd t && GIT_FSMONITOR_TEST=$PWD/t7519/fsmonitor-all ./t3404-rebase-interactive.sh -i)\n>\n> Whereas this works:\n>\n>     (cd t && GIT_FSMONITOR_TEST=$PWD/t7519/fsmonitor-all GIT_SKIP_TESTS=t3404.6 ./t3404-rebase-interactive.sh -i)\n>\n> The entirety of the rest of the test suite still passes with\n> GIT_FSMONITOR_TEST.\n>\n> This has been failing ever since GIT_FSMONITOR_TEST was introduced in\n> 883e248b8a (\"fsmonitor: teach git to optionally utilize a file system\n> monitor to speed up detecting new or changed files.\", 2017-09-22). Under\n> -v -x -i:\n>\n>     + echo test_must_fail: command succeeded: env FAKE_LINES=exec_echo_foo_>file1 1 git rebase -i HEAD^\n>     test_must_fail: command succeeded: env FAKE_LINES=exec_echo_foo_>file1 1 git rebase -i HEAD^\n>     + return 1\n>     error: last command exited with $?=1\n>     not ok 6 - rebase -i with the exec command checks tree cleanness\n>     #\n>     #               git checkout master &&\n>     #               set_fake_editor &&\n>     #               test_must_fail env FAKE_LINES=\"exec_echo_foo_>file1 1\" git rebase -i HEAD^ &&\n>\n> Maybe once this is fixed running the test suite under GIT_FSMONITOR_TEST\n> would be a useful Travis target, but I don't know the current status of\n> adding new options to Travis.\n\n*Poke* at this again. Ben, or anyone else with knowledge of fsmonitor:\nCan you reproduce this?\n\nThis failure along with the one I noted in\nhttps://public-inbox.org/git/87tvn2remn.fsf@evledraar.gmail.com/ is\nfailing the tests on Linux when run with GIT_FSMONITOR_TEST.\n\nI'm looking at this again because SZEDER's patches to the split index\nreminded me again that we have these long-standing failures in rare test\nmodes (see\nhttps://public-inbox.org/git/87va7ireuu.fsf@evledraar.gmail.com/ for the\nsplit index discussion).\n"},{"id":"359824","messageId":"87a7noa5ch.fsf@evledraar.gmail.com","threadId":"47736","inReplyTo":"87sh2mrefa.fsf@evledraar.gmail.com","subject":"Re: t3404.6 breaks on master under GIT_FSMONITOR_TEST","fromName":"Ævar Arnfjörð Bjarmason","fromEmail":"avarab@gmail.com","sentAt":"2018-10-08T14:19:26Z","receivedAt":"2018-10-08T14:19:32Z","isPatch":false,"sender":{"key":"avarab@gmail.com","avatar":"https://avatars.githubusercontent.com/u/45301?v=4"},"body":"\nOn Thu, Sep 06 2018, Ævar Arnfjörð Bjarmason wrote:\n\n> On Thu, Feb 01 2018, Ævar Arnfjörð Bjarmason wrote:\n>\n>> The GIT_FSMONITOR_TEST variable allows you to roundtrip the fsmonitor\n>> codpath in the whole test suite. On both Debian & CentOS this breaks for\n>> me:\n>>\n>>     (cd t && GIT_FSMONITOR_TEST=$PWD/t7519/fsmonitor-all ./t3404-rebase-interactive.sh -i)\n>>\n>> Whereas this works:\n>>\n>>     (cd t && GIT_FSMONITOR_TEST=$PWD/t7519/fsmonitor-all GIT_SKIP_TESTS=t3404.6 ./t3404-rebase-interactive.sh -i)\n>>\n>> The entirety of the rest of the test suite still passes with\n>> GIT_FSMONITOR_TEST.\n>>\n>> This has been failing ever since GIT_FSMONITOR_TEST was introduced in\n>> 883e248b8a (\"fsmonitor: teach git to optionally utilize a file system\n>> monitor to speed up detecting new or changed files.\", 2017-09-22). Under\n>> -v -x -i:\n>>\n>>     + echo test_must_fail: command succeeded: env FAKE_LINES=exec_echo_foo_>file1 1 git rebase -i HEAD^\n>>     test_must_fail: command succeeded: env FAKE_LINES=exec_echo_foo_>file1 1 git rebase -i HEAD^\n>>     + return 1\n>>     error: last command exited with $?=1\n>>     not ok 6 - rebase -i with the exec command checks tree cleanness\n>>     #\n>>     #               git checkout master &&\n>>     #               set_fake_editor &&\n>>     #               test_must_fail env FAKE_LINES=\"exec_echo_foo_>file1 1\" git rebase -i HEAD^ &&\n>>\n>> Maybe once this is fixed running the test suite under GIT_FSMONITOR_TEST\n>> would be a useful Travis target, but I don't know the current status of\n>> adding new options to Travis.\n>\n> *Poke* at this again. Ben, or anyone else with knowledge of fsmonitor:\n> Can you reproduce this?\n>\n> This failure along with the one I noted in\n> https://public-inbox.org/git/87tvn2remn.fsf@evledraar.gmail.com/ is\n> failing the tests on Linux when run with GIT_FSMONITOR_TEST.\n>\n> I'm looking at this again because SZEDER's patches to the split index\n> reminded me again that we have these long-standing failures in rare test\n> modes (see\n> https://public-inbox.org/git/87va7ireuu.fsf@evledraar.gmail.com/ for the\n> split index discussion).\n\nFor what it's worth this is still broken, but more importantly (I'm not\njust keeping bumping the same thing) the only thing that's now broken\nunder fsmonitor. I.e. my skip config is now GIT_SKIP_TESTS=\"t3404.7\"\nwhereas before 43f1180814 (\"git-mv: allow submodules and fsmonitor to\nwork together\", 2018-09-10) I needed to add \"t7411.3 t7411.4\" to that.\n"},{"id":"359847","messageId":"8d486342-e62f-c02c-9d4b-f5e911a12a2f@gmail.com","threadId":"47736","inReplyTo":"87a7noa5ch.fsf@evledraar.gmail.com","subject":"Re: t3404.6 breaks on master under GIT_FSMONITOR_TEST","fromName":"Ben Peart","fromEmail":"peartben@gmail.com","sentAt":"2018-10-08T18:27:26Z","receivedAt":"2018-10-08T18:27:33Z","isPatch":false,"sender":{"key":"benpeart@microsoft.com","avatar":"https://avatars.githubusercontent.com/u/15252029?v=4"},"body":"\n\nOn 10/8/2018 10:19 AM, Ævar Arnfjörð Bjarmason wrote:\n> \n> On Thu, Sep 06 2018, Ævar Arnfjörð Bjarmason wrote:\n> \n>> On Thu, Feb 01 2018, Ævar Arnfjörð Bjarmason wrote:\n>>\n>>> The GIT_FSMONITOR_TEST variable allows you to roundtrip the fsmonitor\n>>> codpath in the whole test suite. On both Debian & CentOS this breaks for\n>>> me:\n>>>\n>>>      (cd t && GIT_FSMONITOR_TEST=$PWD/t7519/fsmonitor-all ./t3404-rebase-interactive.sh -i)\n>>>\n>>> Whereas this works:\n>>>\n>>>      (cd t && GIT_FSMONITOR_TEST=$PWD/t7519/fsmonitor-all GIT_SKIP_TESTS=t3404.6 ./t3404-rebase-interactive.sh -i)\n>>>\n>>> The entirety of the rest of the test suite still passes with\n>>> GIT_FSMONITOR_TEST.\n>>>\n>>> This has been failing ever since GIT_FSMONITOR_TEST was introduced in\n>>> 883e248b8a (\"fsmonitor: teach git to optionally utilize a file system\n>>> monitor to speed up detecting new or changed files.\", 2017-09-22). Under\n>>> -v -x -i:\n>>>\n>>>      + echo test_must_fail: command succeeded: env FAKE_LINES=exec_echo_foo_>file1 1 git rebase -i HEAD^\n>>>      test_must_fail: command succeeded: env FAKE_LINES=exec_echo_foo_>file1 1 git rebase -i HEAD^\n>>>      + return 1\n>>>      error: last command exited with $?=1\n>>>      not ok 6 - rebase -i with the exec command checks tree cleanness\n>>>      #\n>>>      #               git checkout master &&\n>>>      #               set_fake_editor &&\n>>>      #               test_must_fail env FAKE_LINES=\"exec_echo_foo_>file1 1\" git rebase -i HEAD^ &&\n>>>\n>>> Maybe once this is fixed running the test suite under GIT_FSMONITOR_TEST\n>>> would be a useful Travis target, but I don't know the current status of\n>>> adding new options to Travis.\n>>\n>> *Poke* at this again. Ben, or anyone else with knowledge of fsmonitor:\n>> Can you reproduce this?\n>>\n>> This failure along with the one I noted in\n>> https://public-inbox.org/git/87tvn2remn.fsf@evledraar.gmail.com/ is\n>> failing the tests on Linux when run with GIT_FSMONITOR_TEST.\n>>\n>> I'm looking at this again because SZEDER's patches to the split index\n>> reminded me again that we have these long-standing failures in rare test\n>> modes (see\n>> https://public-inbox.org/git/87va7ireuu.fsf@evledraar.gmail.com/ for the\n>> split index discussion).\n> \n> For what it's worth this is still broken, but more importantly (I'm not\n> just keeping bumping the same thing) the only thing that's now broken\n> under fsmonitor. I.e. my skip config is now GIT_SKIP_TESTS=\"t3404.7\"\n> whereas before 43f1180814 (\"git-mv: allow submodules and fsmonitor to\n> work together\", 2018-09-10) I needed to add \"t7411.3 t7411.4\" to that.\n> \n\nI glanced at this for a few minutes but it wasn't obvious what was \nhappening.  It will take some additional effort to dig into and figure \nout the underlying issue.  I haven't forgotten about this - it's still \non my list, just below some other things I need to get finished up first.\n"}]}