{"thread":{"id":"50526","subject":"RE: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","startedAt":"2019-02-16T16:51:43Z","lastAt":"2019-02-20T20:16:33Z","messageCount":16,"participants":["Randall S. Becker","Junio C Hamano","Johannes Schindelin","Max Kirillov"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"369469","messageId":"001201d4c617$de429540$9ac7bfc0$@nexbridge.com","threadId":"50526","inReplyTo":null,"subject":"RE: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","fromName":"Randall S. Becker","fromEmail":"rsbecker@nexbridge.com","sentAt":"2019-02-16T16:51:25Z","receivedAt":"2019-02-16T16:51:43Z","isPatch":false,"sender":{"key":"randall.becker@nexbridge.ca","avatar":"https://avatars.githubusercontent.com/u/28956764?v=4"},"body":"On February 16, 2019 3:27, Max Kirillov wrote:\n> On Fri, Feb 15, 2019 at 04:13:15PM -0500, Randall S. Becker wrote:\n> > Sadly, the fix does not change the results. In fact, it makes the hang\n> > far more likely. Subtest 6,7,8 fails here, at close()\n> \n> Correct, I did not expect it to help, it was for the other issue.\n> \n> As for the hang issue, from your another message it seems to me that perl\n> waiting correctly, there are really child process which do not exit.\n> \n> What you could try is\n> https://public-inbox.org/git/20181124093719.10705-1-max@max630.net/\n> (I'm not sure it would not conflict by now), this would remove dependency\n> between tests. If it helps it would be very valuable information.\n\nGood news. This patch does seem to do the trick. I wonder whether this fixes\nthe Azure build also.\n\nI have run the test under the following conditions:\nRun 1 (system idle): Pass\nRun 2 (system idle): Pass\nRun 3 (system idle): Pass\nRun 4 (system idle): Pass\nRun 5 (system idle): Pass\nRun 6 (system mild load, heavy file system): Pass\nRun 7 (system mild load, moderate file system load - git fetch): Pass\nRun 8 (heavy system load, heavy file system load): Pass\nRun 9 (--verbose, heavy system load, heavy file system load): Pass\nRun 10 (GIT_TRACE=true, --verbose, heavy system load, heavy file system\nload): Pass\nRun 11 (very heavy system load, very heavy file system load): Pass\n\nThe current condition of the code is (the generate_zero_bytes delete was\npreviously removed so can be ignored for the patch):\n\ndiff --git a/t/t5562-http-backend-content-length.sh\nb/t/t5562-http-backend-content-length.sh\nindex bbadde2c6e..f0f425b2cf 100755\n--- a/t/t5562-http-backend-content-length.sh\n+++ b/t/t5562-http-backend-content-length.sh\n@@ -8,12 +8,12 @@ test_lazy_prereq GZIP 'gzip --version'\n verify_http_result() {\n        # some fatal errors still produce status 200\n        # so check if there is the error message\n-       if grep 'fatal:' act.err\n+       if grep 'fatal:' act.err.$test_count\n        then\n                return 1\n        fi\n\n-       if ! grep \"Status\" act.out >act\n+       if ! grep \"Status\" act.out.$test_count >act\n        then\n                printf \"Status: 200 OK\\r\\n\" >act\n        fi\n@@ -33,7 +33,7 @@ test_http_env() {\n                REQUEST_METHOD=POST \\\n                \"$PERL_PATH\" \\\n                \"$TEST_DIRECTORY\"/t5562/invoke-with-content-length.pl \\\n-                   \"$request_body\" git http-backend >act.out 2>act.err\n+                   \"$request_body\" git http-backend >act.out.$test_count\n2>act.err.$test_count\n }\n\n ssize_b100dots() {\n@@ -143,14 +143,14 @@ test_expect_success GZIP 'push gzipped empty' '\n\n test_expect_success 'CONTENT_LENGTH overflow ssite_t' '\n        NOT_FIT_IN_SSIZE=$(ssize_b100dots) &&\n-       generate_zero_bytes infinity  | env \\\n+       env \\\n                CONTENT_TYPE=application/x-git-upload-pack-request \\\n                QUERY_STRING=/repo.git/git-upload-pack \\\n                PATH_TRANSLATED=\"$PWD\"/.git/git-upload-pack \\\n                GIT_HTTP_EXPORT_ALL=TRUE \\\n                REQUEST_METHOD=POST \\\n                CONTENT_LENGTH=\"$NOT_FIT_IN_SSIZE\" \\\n-               git http-backend >/dev/null 2>err &&\n+               git http-backend </dev/null >/dev/null 2>err &&\n        grep \"fatal:.*CONTENT_LENGTH\" err\n '\n\n@@ -161,7 +161,7 @@ test_expect_success 'empty CONTENT_LENGTH' '\n                GIT_HTTP_EXPORT_ALL=TRUE \\\n                REQUEST_METHOD=GET \\\n                CONTENT_LENGTH=\"\" \\\n-               git http-backend <empty_body >act.out 2>act.err &&\n+               git http-backend <empty_body >act.out.$test_count\n2>act.err.$test_count &&\n        verify_http_result \"200 OK\"\n '\n\n\n\n"},{"id":"369470","messageId":"001801d4c619$68ae9160$3a0bb420$@nexbridge.com","threadId":"50526","inReplyTo":"001201d4c617$de429540$9ac7bfc0$@nexbridge.com","subject":"RE: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","fromName":"Randall S. Becker","fromEmail":"rsbecker@nexbridge.com","sentAt":"2019-02-16T17:02:27Z","receivedAt":"2019-02-16T17:02:41Z","isPatch":false,"sender":{"key":"randall.becker@nexbridge.ca","avatar":"https://avatars.githubusercontent.com/u/28956764?v=4"},"body":"On February 16, 2019 11:51, I wrote:\n> To: 'Max Kirillov' <max@max630.net>\n> Cc: 'SZEDER Gábor' <szeder.dev@gmail.com>; 'Johannes Schindelin'\n> <Johannes.Schindelin@gmx.de>; 'Junio C Hamano' <gitster@pobox.com>;\n> git@vger.kernel.org\n> Subject: RE: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News\n> \n> On February 16, 2019 3:27, Max Kirillov wrote:\n> > On Fri, Feb 15, 2019 at 04:13:15PM -0500, Randall S. Becker wrote:\n> > > Sadly, the fix does not change the results. In fact, it makes the\n> > > hang far more likely. Subtest 6,7,8 fails here, at close()\n> >\n> > Correct, I did not expect it to help, it was for the other issue.\n> >\n> > As for the hang issue, from your another message it seems to me that\n> > perl waiting correctly, there are really child process which do not\nexit.\n> >\n> > What you could try is\n> > https://public-inbox.org/git/20181124093719.10705-1-max@max630.net/\n> > (I'm not sure it would not conflict by now), this would remove\n> > dependency between tests. If it helps it would be very valuable\n> information.\n> \n> Good news. This patch does seem to do the trick. I wonder whether this\nfixes\n> the Azure build also.\n\nAnd apologies for the stupid wrapping done by my mail client when I pasted\nthe diff in. The patch is available from the link above anyway so you don't\nneed the junk from my mailer anyway.\n\nCheers,\nRandall\n\n"},{"id":"369471","messageId":"xmqqftsn4nik.fsf@gitster-ct.c.googlers.com","threadId":"50526","inReplyTo":"001201d4c617$de429540$9ac7bfc0$@nexbridge.com","subject":"Re: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2019-02-16T18:05:39Z","receivedAt":"2019-02-16T18:05:44Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"\"Randall S. Becker\" <rsbecker@nexbridge.com> writes:\n\n> On February 16, 2019 3:27, Max Kirillov wrote:\n>\n>> What you could try is\n>> https://public-inbox.org/git/20181124093719.10705-1-max@max630.net/\n>> (I'm not sure it would not conflict by now), this would remove dependency\n>> between tests. If it helps it would be very valuable information.\n>\n> Good news. This patch does seem to do the trick. I wonder whether this fixes\n> the Azure build also.\n>\n> I have run the test under the following conditions:\n> Run 1 (system idle): Pass\n> Run 2 (system idle): Pass\n> Run 3 (system idle): Pass\n> Run 4 (system idle): Pass\n> Run 5 (system idle): Pass\n> Run 6 (system mild load, heavy file system): Pass\n> Run 7 (system mild load, moderate file system load - git fetch): Pass\n> Run 8 (heavy system load, heavy file system load): Pass\n> Run 9 (--verbose, heavy system load, heavy file system load): Pass\n> Run 10 (GIT_TRACE=true, --verbose, heavy system load, heavy file system\n> load): Pass\n> Run 11 (very heavy system load, very heavy file system load): Pass\n\nThat indeed is a good news.\n\n> The current condition of the code is (the generate_zero_bytes delete was\n> previously removed so can be ignored for the patch):\n\nJust to make sure I do not misunderstand, this result is with Max's\npatch but without the generate_zero_bytes stuff?\n\nThanks, all.  Hopefully we can get this test failures behind us\nbefore -rc2; knock, knock...\n"},{"id":"369472","messageId":"000001d4c624$da8e05d0$8faa1170$@nexbridge.com","threadId":"50526","inReplyTo":"xmqqftsn4nik.fsf@gitster-ct.c.googlers.com","subject":"RE: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","fromName":"Randall S. Becker","fromEmail":"rsbecker@nexbridge.com","sentAt":"2019-02-16T18:24:23Z","receivedAt":"2019-02-16T18:24:39Z","isPatch":false,"sender":{"key":"randall.becker@nexbridge.ca","avatar":"https://avatars.githubusercontent.com/u/28956764?v=4"},"body":"On February 16, 2019 13:06, Junio C Hamano wrote:\n> \"Randall S. Becker\" <rsbecker@nexbridge.com> writes:\n> > On February 16, 2019 3:27, Max Kirillov wrote:\n> >> What you could try is\n> >> https://public-inbox.org/git/20181124093719.10705-1-\n> max@max630.net/\n> >> (I'm not sure it would not conflict by now), this would remove\n> >> dependency between tests. If it helps it would be very valuable\n> information.\n> >\n> > Good news. This patch does seem to do the trick. I wonder whether this\n> > fixes the Azure build also.\n> >\n> > I have run the test under the following conditions:\n> > Run 1 (system idle): Pass\n> > Run 2 (system idle): Pass\n> > Run 3 (system idle): Pass\n> > Run 4 (system idle): Pass\n> > Run 5 (system idle): Pass\n> > Run 6 (system mild load, heavy file system): Pass Run 7 (system mild\n> > load, moderate file system load - git fetch): Pass Run 8 (heavy system\n> > load, heavy file system load): Pass Run 9 (--verbose, heavy system\n> > load, heavy file system load): Pass Run 10 (GIT_TRACE=true, --verbose,\n> > heavy system load, heavy file system\n> > load): Pass\n> > Run 11 (very heavy system load, very heavy file system load): Pass\n> \n> That indeed is a good news.\n> \n> > The current condition of the code is (the generate_zero_bytes delete\n> > was previously removed so can be ignored for the patch):\n> \n> Just to make sure I do not misunderstand, this result is with Max's patch but\n> without the generate_zero_bytes stuff?\n\nCorrect.\n\n> Thanks, all.  Hopefully we can get this test failures behind us before -rc2;\n> knock, knock...\n\nOnce the fix is integrated and in the usual spots, I can verify with haste. The full test cycle is now at 50 hours (argh), which I will rerun in full at rc2, but this one is fast.\n\n"},{"id":"369473","messageId":"xmqq5ztj4l3j.fsf@gitster-ct.c.googlers.com","threadId":"50526","inReplyTo":"000001d4c624$da8e05d0$8faa1170$@nexbridge.com","subject":"Re: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2019-02-16T18:57:52Z","receivedAt":"2019-02-16T18:58:42Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"\"Randall S. Becker\" <rsbecker@nexbridge.com> writes:\n\n>> > The current condition of the code is (the generate_zero_bytes delete\n>> > was previously removed so can be ignored for the patch):\n>> \n>> Just to make sure I do not misunderstand, this result is with Max's patch but\n>> without the generate_zero_bytes stuff?\n>\n> Correct.\n\nThanks for a quick response.  I've been staring at b46221ff (\"Merge\nbranch 'rb/no-dev-zero-in-test'\", 2019-02-13).  IIUC, t5562 wouldn't\nhave passed if it still fed http-backend from /dev/zero, no?  The\nshell redirection would have failed, so we do need to keep that part\nof the change---i.e. in order to pass, we do need cc95bc20 (\"t5562:\nreplace /dev/zero with a pipe from generate_zero_bytes\", 2019-02-09)\nand Max's \"t5562: do not reuse output files\", right?\n\nI have been wondering about the whole /dev/zero business.  Although\nwe have b46221ff (\"Merge branch 'rb/no-dev-zero-in-test'\",\n2019-02-13) in 'master', \"git grep /dev/zero t\" has hits in\nt/helper/test-sha1.sh and t/t4152-am-resume-override-opts.sh, so it\nmust have been somewhat incomplete to help platforms that lack\n/dev/zero in the first place.\n\nWe haven't heard from Dscho in European timezone, but I'm\ninclined to\n\n - keep b46221ff in 'master', not reverted.\n - apply Max's \"t5562: do not reuse output files\"\n\nto 'master' and hope that we can declare victory in this part of the\ncode ;-).  There may be fix-ups for other topics before -rc2 on top\nof that, though.\n\n>> Thanks, all.  Hopefully we can get this test failures behind us before -rc2;\n>> knock, knock...\n>\n> Once the fix is integrated and in the usual spots, I can verify\n> with haste. The full test cycle is now at 50 hours (argh), which I\n> will rerun in full at rc2, but this one is fast.\n"},{"id":"369544","messageId":"xmqqef852awh.fsf@gitster-ct.c.googlers.com","threadId":"50526","inReplyTo":"xmqq5ztj4l3j.fsf@gitster-ct.c.googlers.com","subject":"Re: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2019-02-18T18:45:34Z","receivedAt":"2019-02-18T18:45:39Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Junio C Hamano <gitster@pobox.com> writes:\n\n> I have been wondering about the whole /dev/zero business.  Although\n> we have b46221ff (\"Merge branch 'rb/no-dev-zero-in-test'\",\n> 2019-02-13) in 'master', \"git grep /dev/zero t\" has hits in\n> t/helper/test-sha1.sh and t/t4152-am-resume-override-opts.sh, so it\n> must have been somewhat incomplete to help platforms that lack\n> /dev/zero in the first place.\n>\n> We haven't heard from Dscho in European timezone, but I'm\n> inclined to\n>\n>  - keep b46221ff in 'master', not reverted.\n>  - apply Max's \"t5562: do not reuse output files\"\n>\n> to 'master' and hope that we can declare victory in this part of the\n> code ;-).  There may be fix-ups for other topics before -rc2 on top\n> of that, though.\n\nJust to make sure you didn't miss the exchange, the above plan is\nstill the latest (\"Max's patch is [1]).  As I'll be offline most of\nthe day today, you still have time to stop or make me change course\nto avoid disaster by interjecting (which would be very much\nappreciated).\n\nThanks.\n\n\n[Reference]\n\n*1* https://public-inbox.org/git/20181124093719.10705-1-max@max630.net/\n"},{"id":"369545","messageId":"002401d4c7ba$d95a1130$8c0e3390$@nexbridge.com","threadId":"50526","inReplyTo":"xmqqef852awh.fsf@gitster-ct.c.googlers.com","subject":"RE: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","fromName":"Randall S. Becker","fromEmail":"rsbecker@nexbridge.com","sentAt":"2019-02-18T18:50:36Z","receivedAt":"2019-02-18T18:50:56Z","isPatch":false,"sender":{"key":"randall.becker@nexbridge.ca","avatar":"https://avatars.githubusercontent.com/u/28956764?v=4"},"body":"On February 18, 2019 13:46, Junio C Hamano wrote:\n> Junio C Hamano <gitster@pobox.com> writes:\n> \n> > I have been wondering about the whole /dev/zero business.  Although we\n> > have b46221ff (\"Merge branch 'rb/no-dev-zero-in-test'\",\n> > 2019-02-13) in 'master', \"git grep /dev/zero t\" has hits in\n> > t/helper/test-sha1.sh and t/t4152-am-resume-override-opts.sh, so it\n> > must have been somewhat incomplete to help platforms that lack\n> > /dev/zero in the first place.\n> >\n> > We haven't heard from Dscho in European timezone, but I'm inclined to\n> >\n> >  - keep b46221ff in 'master', not reverted.\n> >  - apply Max's \"t5562: do not reuse output files\"\n> >\n> > to 'master' and hope that we can declare victory in this part of the\n> > code ;-).  There may be fix-ups for other topics before -rc2 on top of\n> > that, though.\n> \n> Just to make sure you didn't miss the exchange, the above plan is still the\n> latest (\"Max's patch is [1]).  As I'll be offline most of the day today, you still\n> have time to stop or make me change course to avoid disaster by interjecting\n> (which would be very much appreciated).\n\nNo interjection expected. That is what I have been testing, so anticipate that this is correct.\n\n"},{"id":"369548","messageId":"nycvar.QRO.7.76.6.1902182008270.45@tvgsbejvaqbjf.bet","threadId":"50526","inReplyTo":"xmqq5ztj4l3j.fsf@gitster-ct.c.googlers.com","subject":"Re: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","fromName":"Johannes Schindelin","fromEmail":"johannes.schindelin@gmx.de","sentAt":"2019-02-18T19:37:43Z","receivedAt":"2019-02-18T19:38:00Z","isPatch":false,"sender":{"key":"johannes.schindelin@gmx.de","avatar":"https://avatars.githubusercontent.com/u/127790?v=4"},"body":"Hi Junio,\n\nOn Sat, 16 Feb 2019, Junio C Hamano wrote:\n\n> \"Randall S. Becker\" <rsbecker@nexbridge.com> writes:\n> \n> >> > The current condition of the code is (the generate_zero_bytes delete\n> >> > was previously removed so can be ignored for the patch):\n> >> \n> >> Just to make sure I do not misunderstand, this result is with Max's patch but\n> >> without the generate_zero_bytes stuff?\n> >\n> > Correct.\n> \n> Thanks for a quick response.  I've been staring at b46221ff (\"Merge\n> branch 'rb/no-dev-zero-in-test'\", 2019-02-13).  IIUC, t5562 wouldn't\n> have passed if it still fed http-backend from /dev/zero, no?  The\n> shell redirection would have failed, so we do need to keep that part\n> of the change---i.e. in order to pass, we do need cc95bc20 (\"t5562:\n> replace /dev/zero with a pipe from generate_zero_bytes\", 2019-02-09)\n> and Max's \"t5562: do not reuse output files\", right?\n> \n> I have been wondering about the whole /dev/zero business.  Although\n> we have b46221ff (\"Merge branch 'rb/no-dev-zero-in-test'\",\n> 2019-02-13) in 'master', \"git grep /dev/zero t\" has hits in\n> t/helper/test-sha1.sh and t/t4152-am-resume-override-opts.sh, so it\n> must have been somewhat incomplete to help platforms that lack\n> /dev/zero in the first place.\n> \n> We haven't heard from Dscho in European timezone,\n\nSome bacteria redirected me to /dev/zero for a few days. That hung my\ninbox.\n\n> but I'm inclined to\n> \n>  - keep b46221ff in 'master', not reverted.\n>  - apply Max's \"t5562: do not reuse output files\"\n> \n> to 'master' and hope that we can declare victory in this part of the\n> code ;-).  There may be fix-ups for other topics before -rc2 on top\n> of that, though.\n\nFor the record, I did not set up that fully automated PR build & test at\nhttps://github.com/gitgitgadget/git just so people would still wait for me\nto run the test; a simple PR would have tested this without waiting for\nme.\n\nAnyway, in the meantime, I tested it, and Max' test seems to work:\n\nhttps://dev.azure.com/git-for-windows/git/_build/results?buildId=31274\n\nCiao,\nDscho\n\n> >> Thanks, all.  Hopefully we can get this test failures behind us before -rc2;\n> >> knock, knock...\n> >\n> > Once the fix is integrated and in the usual spots, I can verify\n> > with haste. The full test cycle is now at 50 hours (argh), which I\n> > will rerun in full at rc2, but this one is fast.\n> \n"},{"id":"369552","messageId":"20190218202440.GA3373@jessie.local","threadId":"50526","inReplyTo":"xmqq5ztj4l3j.fsf@gitster-ct.c.googlers.com","subject":"Re: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","fromName":"Max Kirillov","fromEmail":"max@max630.net","sentAt":"2019-02-18T20:24:40Z","receivedAt":"2019-02-18T20:24:50Z","isPatch":false,"sender":{"key":"max@max630.net","avatar":"https://avatars.githubusercontent.com/u/381560?v=4"},"body":"On Sat, Feb 16, 2019 at 10:57:52AM -0800, Junio C Hamano wrote:\n> I'm inclined to\n> \n>  - keep b46221ff in 'master', not reverted.\n\nfrom the branch, cc95bc2025 \"t5562: replace /dev/zero with a\npipe from generate_zero_bytes\" could be replaced with [1]\n\"t5562: do not depend on /dev/zero\", for simplicity. I\ncannot however test the issue as I have no access to the\nplatform in question. As far as I understand, both should\nwork.\n\n[1] https://public-inbox.org/git/20190215164237.12250-1-max@max630.net/\n"},{"id":"369553","messageId":"004901d4c7c8$f6e5d0f0$e4b172d0$@nexbridge.com","threadId":"50526","inReplyTo":"20190218202440.GA3373@jessie.local","subject":"RE: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","fromName":"Randall S. Becker","fromEmail":"rsbecker@nexbridge.com","sentAt":"2019-02-18T20:31:39Z","receivedAt":"2019-02-18T20:31:53Z","isPatch":false,"sender":{"key":"randall.becker@nexbridge.ca","avatar":"https://avatars.githubusercontent.com/u/28956764?v=4"},"body":"On February 18, 2019 15:25, Max Kirillov wrote:\n> On Sat, Feb 16, 2019 at 10:57:52AM -0800, Junio C Hamano wrote:\n> > I'm inclined to\n> >\n> >  - keep b46221ff in 'master', not reverted.\n> \n> from the branch, cc95bc2025 \"t5562: replace /dev/zero with a pipe from\n> generate_zero_bytes\" could be replaced with [1]\n> \"t5562: do not depend on /dev/zero\", for simplicity. I cannot however test\n> the issue as I have no access to the platform in question. As far as I\n> understand, both should work.\n> \n> [1] https://public-inbox.org/git/20190215164237.12250-1-\n> max@max630.net/\n\nAFAIK, b46221ff and [1] should be included. I have no preference on the\ntitle. As soon as (dT based $DAYJOB and $SLEEP) the merge is available (or\n-rc2) comes out, I will test t5562 then run a full regression on NonStop and\nreport back. t5562 takes under a minute. The full suite is 50 hours. How's\nthat?\n\nCheers,\nRandall\n\n"},{"id":"369571","messageId":"nycvar.QRO.7.76.6.1902182249140.45@tvgsbejvaqbjf.bet","threadId":"50526","inReplyTo":"nycvar.QRO.7.76.6.1902182008270.45@tvgsbejvaqbjf.bet","subject":"Re: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","fromName":"Johannes Schindelin","fromEmail":"johannes.schindelin@gmx.de","sentAt":"2019-02-18T21:57:13Z","receivedAt":"2019-02-18T21:57:32Z","isPatch":false,"sender":{"key":"johannes.schindelin@gmx.de","avatar":"https://avatars.githubusercontent.com/u/127790?v=4"},"body":"Hi,\n\nOn Mon, 18 Feb 2019, Johannes Schindelin wrote:\n\n> On Sat, 16 Feb 2019, Junio C Hamano wrote:\n> \n> > [...] but I'm inclined to\n> > \n> >  - keep b46221ff in 'master', not reverted.\n> >  - apply Max's \"t5562: do not reuse output files\"\n> > \n> > to 'master' and hope that we can declare victory in this part of the\n> > code ;-).  There may be fix-ups for other topics before -rc2 on top\n> > of that, though.\n> \n> [...] in the meantime, I tested it, and Max' test seems to work:\n> \n> https://dev.azure.com/git-for-windows/git/_build/results?buildId=31274\n\nI have to take that assessment back. So sad.\n\nAfter that build, I cherry-picked the commit on top of shears/pu (which is\nGit for Windows' ever-green branch that continuously rebases Git for\nWindows' `master` onto git.git's `pu`), and the build seems to hang again:\n\nhttps://dev.azure.com/git-for-windows/git/_build/results?buildId=31291\n\nI kicked off another build with Max' patch reverted, and my `test-tool\ngenzeros` patch applied on top:\n\nhttps://dev.azure.com/git-for-windows/git/_build/results?buildId=31293\n\nI won't be able to wait for its outcome before logging off for the night\n(read: feel free to check its progress yourself, if the linux-clang job is\nstill running 10 minutes after I send this mail, it is probably safe to\nassume that it hangs, otherwise it is fine), but I am fairly certain that\nit will work around the issue again, as it did exactly that before in\nseveral tests (and also in v2.21.0-rc1.windows.1).\n\nCiao,\nDscho\n"},{"id":"369624","messageId":"nycvar.QRO.7.76.6.1902191103050.41@tvgsbejvaqbjf.bet","threadId":"50526","inReplyTo":"002401d4c7ba$d95a1130$8c0e3390$@nexbridge.com","subject":"RE: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","fromName":"Johannes Schindelin","fromEmail":"johannes.schindelin@gmx.de","sentAt":"2019-02-19T10:05:01Z","receivedAt":"2019-02-19T10:05:43Z","isPatch":false,"sender":{"key":"johannes.schindelin@gmx.de","avatar":"https://avatars.githubusercontent.com/u/127790?v=4"},"body":"Hi,\n\nOn Mon, 18 Feb 2019, Randall S. Becker wrote:\n\n> On February 18, 2019 13:46, Junio C Hamano wrote:\n> > Junio C Hamano <gitster@pobox.com> writes:\n> > \n> > > I have been wondering about the whole /dev/zero business.  Although we\n> > > have b46221ff (\"Merge branch 'rb/no-dev-zero-in-test'\",\n> > > 2019-02-13) in 'master', \"git grep /dev/zero t\" has hits in\n> > > t/helper/test-sha1.sh and t/t4152-am-resume-override-opts.sh, so it\n> > > must have been somewhat incomplete to help platforms that lack\n> > > /dev/zero in the first place.\n> > >\n> > > We haven't heard from Dscho in European timezone, but I'm inclined to\n> > >\n> > >  - keep b46221ff in 'master', not reverted.\n> > >  - apply Max's \"t5562: do not reuse output files\"\n> > >\n> > > to 'master' and hope that we can declare victory in this part of the\n> > > code ;-).  There may be fix-ups for other topics before -rc2 on top of\n> > > that, though.\n> > \n> > Just to make sure you didn't miss the exchange, the above plan is still the\n> > latest (\"Max's patch is [1]).  As I'll be offline most of the day today, you still\n> > have time to stop or make me change course to avoid disaster by interjecting\n> > (which would be very much appreciated).\n> \n> No interjection expected. That is what I have been testing, so anticipate that this is correct.\n\nAs I wrote in nycvar.QRO.7.76.6.1902182249140.45@tvgsbejvaqbjf.bet, Max'\npatch did not actually fix the hang on the Linux agents.\n\nSo unfortunately this plan won't work,\nJohannes\n"},{"id":"369665","messageId":"20190219201536.GA2354@jessie.local","threadId":"50526","inReplyTo":"nycvar.QRO.7.76.6.1902182249140.45@tvgsbejvaqbjf.bet","subject":"Re: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","fromName":"Max Kirillov","fromEmail":"max@max630.net","sentAt":"2019-02-19T20:15:36Z","receivedAt":"2019-02-19T20:15:46Z","isPatch":false,"sender":{"key":"max@max630.net","avatar":"https://avatars.githubusercontent.com/u/381560?v=4"},"body":"On Mon, Feb 18, 2019 at 10:57:13PM +0100, Johannes Schindelin wrote:\n> I have to take that assessment back. So sad.\n> \n> After that build, I cherry-picked the commit on top of shears/pu (which is\n> Git for Windows' ever-green branch that continuously rebases Git for\n> Windows' `master` onto git.git's `pu`), and the build seems to hang again:\n> \n> https://dev.azure.com/git-for-windows/git/_build/results?buildId=31291\n\nHi.\n\nYou seem to be talking about the hang like it's some old\nthing, I probably have missed some earlier discussion. I\nhave not heard before that it hangs on linux. The 60 seconds\nhang because of lost SIGCHILD is not it. Also the hang\nobserved at NonStop is not it as well since the no-reuse\nhack did not help (though numbered output files probably\nwould be more sure to avoid duplications expecially at\nWindows where you cannot just unlink busy file and reuse its\nplace in directory)\n\nFrom the tasks you have posted there seem to be no details\navailable. The test is not reported as completed, and the\noverall build fails, and there seem to no additional data\nexcept the log available.\n\nHave you or somebody else been investigating it or is there\notherwise any information about those hangs?\n\n-- \nMax\n"},{"id":"369667","messageId":"20190219203615.GB2354@jessie.local","threadId":"50526","inReplyTo":"20190219201536.GA2354@jessie.local","subject":"Re: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","fromName":"Max Kirillov","fromEmail":"max@max630.net","sentAt":"2019-02-19T20:36:15Z","receivedAt":"2019-02-19T20:36:21Z","isPatch":false,"sender":{"key":"max@max630.net","avatar":"https://avatars.githubusercontent.com/u/381560?v=4"},"body":"On Tue, Feb 19, 2019 at 10:15:36PM +0200, Max Kirillov wrote:\n> expecially at\n> Windows where you cannot just unlink busy file and reuse its\n> place in directory\n\nIt was at Linux actually. Well, whatever the rename feels more\nsure thing to me.\n"},{"id":"369668","messageId":"010e01d4c895$3c997800$b5cc6800$@nexbridge.com","threadId":"50526","inReplyTo":"20190219201536.GA2354@jessie.local","subject":"RE: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","fromName":"Randall S. Becker","fromEmail":"rsbecker@nexbridge.com","sentAt":"2019-02-19T20:53:53Z","receivedAt":"2019-02-19T20:54:14Z","isPatch":false,"sender":{"key":"randall.becker@nexbridge.ca","avatar":"https://avatars.githubusercontent.com/u/28956764?v=4"},"body":"On February 19, 2019 15:16, Max Kirillov <max@max630.net> wrote:\n> On Mon, Feb 18, 2019 at 10:57:13PM +0100, Johannes Schindelin wrote:\n> > I have to take that assessment back. So sad.\n> >\n> > After that build, I cherry-picked the commit on top of shears/pu\n> > (which is Git for Windows' ever-green branch that continuously rebases\n> > Git for Windows' `master` onto git.git's `pu`), and the build seems to\nhang\n> again:\n> >\n> > https://dev.azure.com/git-for-windows/git/_build/results?buildId=31291\n> \n> Hi.\n> \n> You seem to be talking about the hang like it's some old thing, I probably\n> have missed some earlier discussion. I have not heard before that it hangs\non\n> linux. The 60 seconds hang because of lost SIGCHILD is not it. Also the\nhang\n> observed at NonStop is not it as well since the no-reuse hack did not help\n> (though numbered output files probably would be more sure to avoid\n> duplications expecially at Windows where you cannot just unlink busy file\nand\n> reuse its place in directory)\n> \n> >From the tasks you have posted there seem to be no details\n> available. The test is not reported as completed, and the overall build\nfails,\n> and there seem to no additional data except the log available.\n> \n> Have you or somebody else been investigating it or is there otherwise any\n> information about those hangs?\n\nThe no-reuse hack made a big different\n(https://public-inbox.org/git/20181124093719.10705-1-max@max630.net/) in\nconjunction with replacing </dev/zero with </dev/null. It was the sleep\nhack, and generate_zero_bytes that had no effect. I previously supplied a\nmulti-run result. Please refer to that fix/test for reference, which I paste\nbelow:\n\nOn February 16, 2019 13:06, Junio C Hamano wrote:\n> \"Randall S. Becker\" <rsbecker@nexbridge.com> writes:\n> > On February 16, 2019 3:27, Max Kirillov wrote:\n> >> What you could try is\n> >> https://public-inbox.org/git/20181124093719.10705-1-\n> max@max630.net/\n> >> (I'm not sure it would not conflict by now), this would remove \n> >> dependency between tests. If it helps it would be very valuable\n> information.\n> >\n> > Good news. This patch does seem to do the trick. I wonder whether \n> > this fixes the Azure build also.\n> >\n> > I have run the test under the following conditions:\n> > Run 1 (system idle): Pass\n> > Run 2 (system idle): Pass\n> > Run 3 (system idle): Pass\n> > Run 4 (system idle): Pass\n> > Run 5 (system idle): Pass\n> > Run 6 (system mild load, heavy file system): Pass Run 7 (system mild \n> > load, moderate file system load - git fetch): Pass Run 8 (heavy \n> > system load, heavy file system load): Pass Run 9 (--verbose, heavy \n> > system load, heavy file system load): Pass Run 10 (GIT_TRACE=true, \n> > --verbose, heavy system load, heavy file system\n> > load): Pass\n> > Run 11 (very heavy system load, very heavy file system load): Pass\n> \n> That indeed is a good news.\n> \n> > The current condition of the code is (the generate_zero_bytes delete \n> > was previously removed so can be ignored for the patch):\n> \n> Just to make sure I do not misunderstand, this result is with Max's \n> patch but without the generate_zero_bytes stuff?\n\nCorrect.\n\n\n\n"},{"id":"369735","messageId":"nycvar.QRO.7.76.6.1902202112420.41@tvgsbejvaqbjf.bet","threadId":"50526","inReplyTo":"20190219201536.GA2354@jessie.local","subject":"Re: [ANNOUNCE] Git v2.21.0-rc1 (NonStop Results) - Good News","fromName":"Johannes Schindelin","fromEmail":"johannes.schindelin@gmx.de","sentAt":"2019-02-20T20:15:55Z","receivedAt":"2019-02-20T20:16:33Z","isPatch":false,"sender":{"key":"johannes.schindelin@gmx.de","avatar":"https://avatars.githubusercontent.com/u/127790?v=4"},"body":"Hi Max,\n\nOn Tue, 19 Feb 2019, Max Kirillov wrote:\n\n> On Mon, Feb 18, 2019 at 10:57:13PM +0100, Johannes Schindelin wrote:\n> > I have to take that assessment back. So sad.\n> > \n> > After that build, I cherry-picked the commit on top of shears/pu (which is\n> > Git for Windows' ever-green branch that continuously rebases Git for\n> > Windows' `master` onto git.git's `pu`), and the build seems to hang again:\n> > \n> > https://dev.azure.com/git-for-windows/git/_build/results?buildId=31291\n> \n> Hi.\n> \n> You seem to be talking about the hang like it's some old\n> thing, I probably have missed some earlier discussion. I\n> have not heard before that it hangs on linux. The 60 seconds\n> hang because of lost SIGCHILD is not it. Also the hang\n> observed at NonStop is not it as well since the no-reuse\n> hack did not help (though numbered output files probably\n> would be more sure to avoid duplications expecially at\n> Windows where you cannot just unlink busy file and reuse its\n> place in directory)\n> \n> From the tasks you have posted there seem to be no details\n> available. The test is not reported as completed, and the\n> overall build fails, and there seem to no additional data\n> except the log available.\n> \n> Have you or somebody else been investigating it or is there\n> otherwise any information about those hangs?\n\nSorry, there has been so much talk about \"hang\", and we got quite a bit of\nthings mixed up.\n\nWhat I was talking about, in the text you quoted, we an infinite hang in\nt5562.15 on Linux (but I could not reproduce locally, it only happened in\nthe CI, but there it happened *reliably*). All the builds timed out after\n60 *minutes*.\n\nThis has thankfully been addressed in the meantime, so \"my\" hang is gone.\n\nThanks,\nDscho\n\nP.S.: in some other thread, I was picking up on something you said about\nthat hang that *you* were talking about, the one timing out after 60\nseconds. I did not observe that here, but I had hoped (in vain) that a\nquick read over the code would turn up something useful (which was my\nsuggestion to close stdout in the gzip case, but it seemed to not make a\ndifference).\n"}]}