Teach bisect-runner.pl how to build a significant PERL_OBJECT commit.
authorNicholas Clark <nick@ccl4.org>
Wed, 7 Dec 2011 15:41:21 +0000 (16:41 +0100)
committerNicholas Clark <nick@ccl4.org>
Wed, 7 Dec 2011 21:16:20 +0000 (22:16 +0100)
Commit 22c35a8c2392967a is "phase 1 of somewhat major rearrangement of
PERL_OBJECT stuff" but doesn't build due to 2 simple errors. blead in this
broken state was merged to the cfgperl branch, and then these were
immediately corrected there. cfgperl (with the fixes) was merged back to
blead. The resultant rather twisty maze of commits completely confuses git
bisect (and at least me), causing it to the bisect run to confidently return
the wrong answer, an unrelated commit on the cfgperl branch.

Porting/bisect-runner.pl

index 3176bad..af37f77 100755 (executable)
@@ -2447,6 +2447,82 @@ EOPATCH
     }
 
     if ($major == 5
+        && `git rev-parse HEAD` eq "22c35a8c2392967a5ba6b5370695be464bd7012c\n") {
+        # Commit 22c35a8c2392967a is significant,
+        # "phase 1 of somewhat major rearrangement of PERL_OBJECT stuff"
+        # but doesn't build due to 2 simple errors. blead in this broken state
+        # was merged to the cfgperl branch, and then these were immediately
+        # corrected there. cfgperl (with the fixes) was merged back to blead.
+        # The resultant rather twisty maze of commits looks like this:
+
+=for comment
+
+* | |   commit 137225782c183172f360c827424b9b9f8adbef0e
+|\ \ \  Merge: 22c35a8 2a8ee23
+| |/ /  Author: Gurusamy Sarathy <gsar@cpan.org>
+| | |   Date:   Fri Oct 30 17:38:36 1998 +0000
+| | |
+| | |       integrate cfgperl tweaks into mainline
+| | |
+| | |       p4raw-id: //depot/perl@2144
+| | |
+| * | commit 2a8ee23279873759693fa83eca279355db2b665c
+| | | Author: Jarkko Hietaniemi <jhi@iki.fi>
+| | | Date:   Fri Oct 30 13:27:39 1998 +0000
+| | |
+| | |     There can be multiple yacc/bison errors.
+| | |
+| | |     p4raw-id: //depot/cfgperl@2143
+| | |
+| * | commit 93fb2ac393172fc3e2c14edb20b718309198abbc
+| | | Author: Jarkko Hietaniemi <jhi@iki.fi>
+| | | Date:   Fri Oct 30 13:18:43 1998 +0000
+| | |
+| | |     README.posix-bc update.
+| | |
+| | |     p4raw-id: //depot/cfgperl@2142
+| | |
+| * | commit 4ec43091e8e6657cb260b5e563df30aaa154effe
+| | | Author: Jarkko Hietaniemi <jhi@iki.fi>
+| | | Date:   Fri Oct 30 09:12:59 1998 +0000
+| | |
+| | |     #2133 fallout.
+| | |
+| | |     p4raw-id: //depot/cfgperl@2141
+| | |
+| * |   commit 134ca994cfefe0f613d43505a885e4fc2100b05c
+| |\ \  Merge: 7093112 22c35a8
+| |/ /  Author: Jarkko Hietaniemi <jhi@iki.fi>
+|/| |   Date:   Fri Oct 30 08:43:18 1998 +0000
+| | |
+| | |       Integrate from mainperl.
+| | |
+| | |       p4raw-id: //depot/cfgperl@2140
+| | |
+* | | commit 22c35a8c2392967a5ba6b5370695be464bd7012c
+| | | Author: Gurusamy Sarathy <gsar@cpan.org>
+| | | Date:   Fri Oct 30 02:51:39 1998 +0000
+| | |
+| | |     phase 1 of somewhat major rearrangement of PERL_OBJECT stuff
+| | |     (objpp.h is gone, embed.pl now does some of that); objXSUB.h
+| | |     should soon be automated also; the global variables that
+| | |     escaped the PL_foo conversion are now reined in; renamed
+| | |     MAGIC in regcomp.h to REG_MAGIC to avoid collision with the
+| | |     type of same name; duplicated lists of pp_things in various
+| | |     places is now gone; result has only been tested on win32
+| | |
+| | |     p4raw-id: //depot/perl@2133
+
+=cut
+
+        # and completely confuses git bisect (and at least me), causing it to
+        # the bisect run to confidently return the wrong answer, an unrelated
+        # commit on the cfgperl branch.
+
+        apply_commit('4ec43091e8e6657c');
+    }
+
+    if ($major == 5
         && extract_from_file('pp_sys.c', qr/PERL_EFF_ACCESS_R_OK/)
         && !extract_from_file('pp_sys.c', qr/XXX Configure test needed for eaccess/)) {
         # Between 5ff3f7a4e03a6b10 and c955f1177b2e311d^