3 # Script helping make fast-forwarding histories while still rebasing
4 # upstream deltas when working on Debian packaging
6 # Copyright (C)2017 Ian Jackson
8 # This program is free software: you can redistribute it and/or modify
9 # it under the terms of the GNU General Public License as published by
10 # the Free Software Foundation, either version 3 of the License, or
11 # (at your option) any later version.
13 # This program is distributed in the hope that it will be useful,
14 # but WITHOUT ANY WARRANTY; without even the implied warranty of
15 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 # GNU General Public License for more details.
18 # You should have received a copy of the GNU General Public License
19 # along with this program. If not, see <http://www.gnu.org/licenses/>.
22 # git-debrebase status
23 # git-debrebase start # like ffqrebase start + debrebase launder
24 # git-debrebase new-upstream [stuff] # see below
25 # git-debrebase <git-rebase options> # does debrebase start if necessary
27 # git-debrebase analyse
28 # git-debrebase launder # prints breakwater tip
29 # git-debrebase create-new-upstream-breakwater [-f] <upstreaminfo>...
32 # [,][<subdir>:][+]<commitid>[,...]
34 # if initial comma is supplied, entries are not positional. Unspecified
35 # <subdir> means root (and there may be only one).
36 # xxx want auto branch names
39 # [+]<commit> [<subdir/> [+]<commit>...]
41 # --new-upstream-different-subtrees
44 # git-debrebase new-upstream
45 # - previous breakwater merge must be gdr-generated
46 # - orig set is the same as before
47 # - implicitly uses upstream branches according to orig set
48 # - not all upstream branches need be updated
49 # - insists on fast-forward of each branch, unless
50 # --force (or --force=<subdir>[/])
51 # branch set adjustments
52 # git-debrebase new-upstream --add <subdir>/
53 # git-debrebase new-upstream --rm <subdir>/
54 # git-debrebase new-upstream / [<subdir>/ ...]
55 # - orig set is adjusted
56 # - otherwise like auto (--add is not checked for ffness, obv)
57 # - multiple --add and --rm may be specified
58 # - --add makes new upstream the last contributor
60 # git-debrebase / [<rootcommitid>] [<subdir>/ [<commitid>] ...]
61 # - orig set is precisely as specified now
62 # - previous breakwater merge is irrelevant
63 # - no fast forward checks
64 # for now only explicit with commitids
66 # implicitly uses `upstream'
67 # # (or multiple other branches)
68 # git-debrebase new-upstream \
69 # [<subdir>/]=<commitid>
71 # UPSTREAM[,[[SUBDIR:]SUBUPSTREAM]
72 # default for SUBDIR: is from previous upstream merge[xxx terminology]
76 # when starting must record original start (for ff)
77 # and new rebase basis
79 # git-ffqrebase start [BASE]
80 # # records previous HEAD so it can be overwritten
81 # # records base for future git-ffqrebase
82 # git-ffqrebase set-base BASE
83 # git-ffqrebase <git-rebase options>
84 # git-ffqrebase finish
85 # git-ffqrebase status [BRANCH]
87 # refs/ffqrebase-prev/BRANCH BRANCH may be refs/...; if not it means
88 # refs/ffqrebase-base/BRANCH refs/heads/BRANCH
89 # zero, one, or both of these may exist
91 # git-debrebase without start, if already started, is willing
92 # to strip pseudomerges provided that they overwrite exactly
94 # xxxx is this right ? what matters is have we pushed
95 # I think in fact the right answer is:
96 # git-debrebase always strips out pseudomerges from its branch
97 # a pseudomerge is put in at the time we want to push
98 # at that time, we make a pseudomerge of the remote tracking
99 # branch (if raw git) or the dgit view (if dgit)
100 # for raw git git-ffqrebase, do want preciseley to record
101 # value of remote tracking branch or our branch, on start, so we
102 # overwrite only things we intend to
103 # the previous pseudomerge check for tags and remote branches ?
107 use Debian::Dgit qw(:DEFAULT :playground);
114 use Getopt::Long qw(:config posix_default gnu_compat bundling);
118 die "bad usage: $m\n";
124 my @cmd = qw(git config -z);
125 push @cmd, qw(--get-all) if wantarray;
127 my $out = cmdoutput @cmd;
128 return split /\0/, $out;
135 my $data = git_cat_file $objid, 'commit';
136 $data =~ m/(?<=\n)\n/ or die "$objid ($data) ?";
140 sub D_UPS () { 0x02; } # upstream files
141 sub D_PAT_ADD () { 0x04; } # debian/patches/ extra patches at end
142 sub D_PAT_OTH () { 0x08; } # debian/patches other changes
143 sub D_DEB_CLOG () { 0x10; } # debian/ (not patches/ or changelog)
144 sub D_DEB_OTH () { 0x20; } # debian/changelog
145 sub DS_DEB () { D_DEB_CLOG | D_DEB_OTH; } # debian/ (not patches/)
147 our $playprefix = 'debrebase';
153 sub in_workarea ($) {
156 my $r = eval { $sub->(); };
157 { local $@; changedir $maindir; }
161 sub fresh_workarea () {
162 $workarea = fresh_playground "$playprefix/work";
163 in_workarea sub { playtree_setup };
166 sub get_differs ($$) {
168 # This resembles quiltify_trees_differ, in dgit, a bit.
169 # But we don't care about modes, or dpkg-source-unrepresentable
170 # changes, and we don't need the plethora of different modes.
171 # Conversely we need to distinguish different kinds of changes to
172 # debian/ and debian/patches/.
177 my ($opts, $limits, $fn) = @_;
178 my @cmd = (@git, qw(diff-tree -z --no-renames));
180 push @cmd, "$_:" foreach $x, $y;
182 my $diffs = cmdoutput @cmd;
183 foreach (split /\0/, $diffs) { $fn->(); }
186 $rundiff->([qw(--name-only)], [], sub {
187 $differs |= $_ eq 'debian' ? DS_DEB : D_UPS;
190 if ($differs & DS_DEB) {
192 $rundiff->([qw(--name-only -r)], [qw(debian)], sub {
194 m{^debian/patches/} ? D_PAT_OTH :
195 $_ eq 'debian/changelog' ? D_DEB_CLOG :
198 die "mysterious debian changes $x..$y"
199 unless $differs & (D_PAT_OTH|DS_DEB);
202 if ($differs & D_PAT_OTH) {
204 $differs &= ~D_PAT_OTH;
206 $differs |= D_PAT_OTH;
207 no warnings qw(exiting); last;
209 $rundiff->([qw(--name-status -r)], [qw(debian/patches/)], sub {
210 no warnings qw(exiting);
211 if (!defined $mode) {
214 die unless s{^debian/patches/}{};
216 if ($mode eq 'A' && !m/\.series$/s) {
218 } elsif ($mode eq 'M' && $_ eq 'series') {
219 my $x_s = git_cat_file "$x:debian/patches/series", 'blob';
220 my $y_s = git_cat_file "$y:debian/patches/series", 'blob';
221 chomp $x_s; $x_s .= "\n";
222 $ok = $x_s eq substr($y_s, 0, length $x_s);
227 $differs |= $ok ? D_PAT_ADD : D_PAT_OTH;
229 die "mysterious debian/patches changes $x..$y"
230 unless $differs & (D_PAT_ADD|D_PAT_OTH);
233 printdebug sprintf "get_differs %s, %s = %#x\n", $x, $y, $differs;
238 sub commit_pr_info ($) {
240 return Data::Dumper->dump([$r], [qw(commit)]);
243 sub calculate_committer_authline () {
244 my $c = cmdoutput @git, qw(commit-tree --no-gpg-sign -m),
245 'DUMMY COMMIT (git-debrebase)', "HEAD:";
246 my ($h,$m) = get_commit $c;
247 $h =~ m/^committer .*$/m or confess "($h) ?";
251 # classify returns an info hash like this
253 # Hdr => # commit headers, including 1 final newline
254 # Msg => # commit message (so one newline is dropped)
256 # Type => (see below)
258 # Ix => $index # ie 0, 1, 2, ...
260 # Differs => return value from get_differs
262 # IsDggitImport => 'orig' 'tarball' 'unpatched' 'package' (as from dgit)
264 # NewMsg => # commit message, but with any [dgit import ...] edited
265 # # to say "[was: ...]"
276 # has additional entres in classification result
277 # Overwritten = [ subset of Parents ]
278 # Contributor = $the_remaining_Parent
280 # DgitImportUnpatched
281 # has additional entry in classification result
282 # OrigParents = [ subset of Parents ]
284 # BreakwaterUpstreamMerge
285 # has additional entry in classification result
286 # OrigParents = [ subset of Parents ] # singleton list
288 sub parsecommit ($;$) {
289 my ($objid, $p_ref) = @_;
290 # => hash with CommitId Hdr Msg Tree Parents
291 # Parents entries have only Ix CommitId
292 # $p_ref, if provided, must be [] and is used as a base for Parents
297 my ($h,$m) = get_commit $objid;
299 my ($t) = $h =~ m/^tree (\w+)$/m or die $objid;
300 my (@ph) = $h =~ m/^parent (\w+)$/mg;
310 foreach my $ph (@ph) {
324 my $r = parsecommit($objid, \@p);
328 $p->{Differs} => (get_differs $p->{CommitId}, $t),
331 printdebug "classify $objid \$t=$t \@p",
332 (map { sprintf " %s/%#x", $_->{CommitId}, $_->{Differs} } @p),
336 my ($type, @rest) = @_;
337 $r = { %$r, Type => $type, @rest };
339 my $dd = new Data::Dumper [ $r ];
340 Terse $dd 1; Indent $dd 0; Useqq $dd 1;
341 printdebug " = $type ".(Dump $dd)."\n";
347 $r = { %$r, Type => qw(Unknown) };
348 printdebug " ** Unknown\n";
353 my $d = $r->{Parents}[0]{Differs};
354 if ($d == D_PAT_ADD) {
355 return $classify->(qw(AddPatches));
356 } elsif ($d & (D_PAT_ADD|D_PAT_OTH)) {
357 return $unknown->("edits debian/patches");
358 } elsif ($d & DS_DEB and !($d & ~DS_DEB)) {
359 my ($ty,$dummy) = git_cat_file "$ph[0]:debian";
361 if ($d == D_DEB_CLOG) {
362 return $classify->(qw(Changelog));
364 return $classify->(qw(Packaging));
366 } elsif ($ty eq 'missing') {
367 return $classify->(qw(BreakwaterStart));
369 return $unknown->("parent's debian is not a directory");
371 } elsif ($d == D_UPS) {
372 return $classify->(qw(Upstream));
373 } elsif ($d & DS_DEB and $d & D_UPS and !($d & ~(DS_DEB|D_UPS))) {
374 return $classify->(qw(Mixed));
376 return $unknown->("no changes");
378 confess "internal error $objid ?";
382 return $unknown->("origin commit");
385 my @identical = grep { !$_->{Differs} } @p;
386 if (@p == 2 && @identical == 1) {
387 my @overwritten = grep { $_->{Differs} } @p;
388 confess "internal error $objid ?" unless @overwritten==1;
389 return $classify->(qw(Pseudomerge),
390 Overwritten => $overwritten[0],
391 Contributor => $identical[0]);
393 if (@p == 2 && @identical == 2) {
394 my @bytime = nsort_by {
395 my ($ph,$pm) = get_commit $_->{CommitId};
396 $ph =~ m/^committer .* (\d+) [-+]\d+$/m or die "$_->{CommitId} ?";
399 return $classify->(qw(Pseudomerge),
400 SubType => qw(Ambiguous),
401 Overwritten => $bytime[0],
402 Contributor => $bytime[1]);
405 my ($p_h, $p_m) = get_commit $p->{CommitId};
406 $p->{IsOrigin} = $p_h !~ m/^parent \w+$/m;
407 ($p->{IsDgitImport},) = $p_m =~ m/^\[dgit import ([0-9a-z]+) .*\]$/m;
409 my @orig_ps = grep { ($_->{IsDgitImport}//'X') eq 'orig' } @p;
411 if (!(grep { !$_->{IsOrigin} } @p) and
412 (@orig_ps >= @p - 1) and
413 $m2 =~ s{^\[(dgit import unpatched .*)\]$}{[was: $1]}m) {
415 return $classify->(qw(DgitImportUnpatched),
416 OrigParents => \@orig_ps);
419 my ($stype, $series) = git_cat_file "$t:debian/patches/series";
420 my $haspatches = $stype ne 'missing' && $series =~ m/^\s*[^#\n\t ]/m;
422 # How to decide about l/r ordering of breakwater merges ? git
423 # --topo-order prefers to expand 2nd parent first. There's
424 # already an easy rune to look for debian/ history anyway (git log
425 # debian/) so debian breakwater branch should be 1st parent; that
426 # way also there's also an easy rune to look for the upstream
427 # patches (--topo-order).
429 # The above tells us which way *we* will generate them. But we
430 # might encounter ad-hoc breakwater merges generated manually,
431 # which might be the other way around. In principle, in some odd
432 # situations, a breakwater merge might have two identical parents.
433 # In that case we guess which way round it is (ie, which parent
434 # has the upstream history). The order of the 2-iteration loop
435 # controls which guess we make.
437 foreach my $prevbrw (qw(0 1)) {
440 !$p[$prevbrw]{IsOrigin} && # breakwater never starts with an origin
441 !($p[$prevbrw]{Differs} & ~DS_DEB) &&
442 !($p[!$prevbrw]{Differs} & ~D_UPS)) {
443 return $classify->(qw(BreakwaterUpstreamMerge),
444 OrigParents => [ $p[!$prevbrw] ]);
448 # multi-orig upstreams are represented with a breakwater merge
449 # from a single upstream commit which combines the orig tarballs
451 return $unknown->("complex merge");
457 $nogenerate,$report) = @_;
458 # => ($tip, $breakwater_tip)
459 # (or nothing, if $nogenerate)
461 # go through commits backwards
462 # we generate two lists of commits to apply:
463 # breakwater branch and upstream patches
464 my (@brw_cl, @upp_cl, @processed);
471 my ($appendinfo) = @_;
474 $ms .= "\n\n[git-debrebase $appendinfo]\n";
477 my $rewrite_from_here = sub {
478 my $sp_cl = { SpecialMethod => 'StartRewrite' };
479 push @brw_cl, $sp_cl;
480 push @processed, $sp_cl;
485 my $prprdelim = sub { print $report $prdelim if $report; $prdelim=""; };
488 return unless $report;
489 print $report $prdelim, @_;
493 my $bomb = sub { # usage: return $bomb->();
494 print $report " Unprocessable" if $report;
497 return (undef,undef);
499 die "commit $cur: Cannot cope with this commit (d.".
500 (join ' ', map { sprintf "%#x", $_->{Differs} }
501 @{ $cl->{Parents} }). ")";
507 my $build_start = sub {
508 my ($msg, $parent) = @_;
511 no warnings qw(exiting); last;
516 my $ty = $cl->{Type};
517 my $st = $cl->{SubType};
518 $prline->("$cl->{CommitId} $cl->{Type}");
519 $found{$ty. ( defined($st) ? "-$st" : '' )}++;
520 push @processed, $cl;
521 my $p0 = @{ $cl->{Parents} }==1 ? $cl->{Parents}[0]{CommitId} : undef;
522 if ($ty eq 'AddPatches') {
524 $rewrite_from_here->();
526 } elsif ($ty eq 'Packaging' or $ty eq 'Changelog') {
530 } elsif ($ty eq 'BreakwaterStart') {
531 $build_start->('FirstPackaging', $cur);
532 } elsif ($ty eq 'Upstream') {
536 } elsif ($ty eq 'Mixed') {
539 my $cls = { %$cl, $xmsg->("split mixed commit: $wh part") };
542 $queue->(\@brw_cl, "debian");
543 $queue->(\@upp_cl, "upstream");
544 $rewrite_from_here->();
547 } elsif ($ty eq 'Pseudomerge') {
548 my $contrib = $cl->{Contributor}{CommitId};
549 print $report " Contributor=$contrib" if $report;
550 push @pseudomerges, $cl;
551 $rewrite_from_here->();
554 } elsif ($ty eq 'BreakwaterUpstreamMerge') {
555 $build_start->("PreviousBreakwater", $cur);
556 } elsif ($ty eq 'DgitImportUnpatched') {
557 my $pm = $pseudomerges[-1];
559 # To an extent, this is heuristic. Imports don't have
560 # a useful history of the debian/ branch. We assume
561 # that the first pseudomerge after an import has a
562 # useful history of debian/, and ignore the histories
563 # from later pseudomerges. Often the first pseudomerge
564 # will be the dgit import of the upload to the actual
565 # suite intended by the non-dgit NMUer, and later
566 # pseudomerges may represent in-archive copies.
567 my $ovwrs = $pm->{Overwritten};
568 printf $report " PM=%s \@Overwr:%d", $pm, (scalar @$ovwrs)
573 my $ovwr = $ovwrs->[0]{CommitId};
574 printf $report " Overwr=%s", $ovwr if $report;
575 # This import has a tree which is just like a
576 # breakwater tree, but it has the wrong history. It
577 # ought to have the previous breakwater (which the
578 # pseudomerge overwrote) as an ancestor. That will
579 # make the history of the debian/ files correct. As
580 # for the upstream version: either it's the same as
581 # was ovewritten (ie, same as the previous
582 # breakwater), in which case that history is precisely
583 # right; or, otherwise, it was a non-gitish upload of a
584 # new upstream version. We can tell these apart by
585 # looking at the tree of the supposed upstream.
588 SpecialMethod => 'DgitImportDebianUpdate',
589 $xmsg->("convert dgit import: debian changes")
591 my $differs = (get_differs $ovwr, $cl->{Tree});
592 printf $report " Differs=%#x", $differs if $report;
593 if ($differs & D_UPS) {
594 printf $report " D_UPS" if $report;
595 # This will also trigger if a non-dgit git-based NMU
596 # deleted .gitignore (which is a thing that some of
597 # the existing git tools do if the user doesn't
598 # somehow tell them not to). Ah well.
601 SpecialMethod => 'DgitImportUpstreamUpdate',
602 $xmsg->("convert dgit import: upstream changes")
605 $prline->(" Import");
606 $rewrite_from_here->();
607 $upp_limit //= $#upp_cl; # further, deeper, patches discarded
611 # Everything is from this import. This kind of import
612 # is already in valid breakwater format, with the
613 # patches as commits.
614 printf $report " NoPM" if $report;
615 # last thing we processed will have been the first patch,
616 # if there is one; which is fine, so no need to rewrite
617 # on account of this import
618 $build_start->("ImportOrigin", $cur);
626 return if $nogenerate;
628 # Now we build it back up again
634 my $rm_tree_cached = sub {
636 runcmd @git, qw(rm --quiet -rf --cached --ignore-unmatch), $subdir;
638 my $read_tree_debian = sub {
640 $rm_tree_cached->(qw(debian));
641 runcmd @git, qw(read-tree --prefix=debian/), "$treeish:debian";
643 my $read_tree_upstream = sub {
645 runcmd @git, qw(read-tree), $treeish;
646 $read_tree_debian->($build);
649 my $committer_authline = calculate_committer_authline();
651 printdebug "WALK REBUILD $build ".(scalar @processed)."\n";
653 confess "internal error" unless $build eq (pop @processed)->{CommitId};
656 mkdir $rd or $!==EEXIST or die $!;
658 runcmd @git, qw(read-tree), $build;
659 foreach my $cl (qw(Debian), (reverse @brw_cl),
660 { SpecialMethod => 'RecordBreakwaterTip' },
661 qw(Upstream), (reverse @upp_cl)) {
663 $current_method = $cl;
666 my $method = $cl->{SpecialMethod} // $current_method;
667 my @parents = ($build);
668 my $cltree = $cl->{CommitId};
669 printdebug "WALK BUILD ".($cltree//'undef').
670 " $method (rewriting=$rewriting)\n";
671 if ($method eq 'Debian') {
672 $read_tree_debian->($cltree);
673 } elsif ($method eq 'Upstream') {
674 $read_tree_upstream->($cltree);
675 } elsif ($method eq 'StartRewrite') {
678 } elsif ($method eq 'RecordBreakwaterTip') {
679 $breakwater = $build;
681 } elsif ($method eq 'DgitImportDebianUpdate') {
682 $read_tree_debian->($cltree);
683 $rm_tree_cached->(qw(debian/patches));
684 } elsif ($method eq 'DgitImportUpstreamUpdate') {
685 $read_tree_upstream->($cltree);
686 push @parents, map { $_->{CommitId} } @{ $cl->{OrigParents} };
691 my $procd = (pop @processed) // 'UNDEF';
694 printdebug "WALK REWRITING NOW cl=$cl procd=$procd\n";
697 my $newtree = cmdoutput @git, qw(write-tree);
699 $ch =~ s{^tree .*}{tree $newtree}m or confess "$ch ?";
700 $ch =~ s{^parent .*\n}{}m;
701 $ch =~ s{(?=^author)}{
702 join '', map { "parent $_\n" } @parents
703 }me or confess "$ch ?";
705 $ch =~ s{^committer .*$}{$committer_authline}m
708 my $cf = "$rd/m$rewriting";
709 open CD, ">", $cf or die $!;
710 print CD $ch, "\n", $cl->{Msg} or die $!;
712 my @cmd = (@git, qw(hash-object));
713 push @cmd, qw(-w) if $rewriting;
714 push @cmd, qw(-t commit), $cf;
715 my $newcommit = cmdoutput @cmd;
716 confess "$ch ?" unless $rewriting or $newcommit eq $cl->{CommitId};
721 my $final_check = get_differs $build, $input;
722 die sprintf "internal error %#x %s %s", $final_check, $build, $input
723 if $final_check & ~D_PAT_ADD;
725 return ($build, $breakwater);
728 sub get_head () { return git_rev_parse qw(HEAD); }
730 sub update_head ($$$) {
731 my ($old, $new, $mrest) = @_;
732 runcmd @git, qw(update-ref -m), "debrebase: $mrest", 'HEAD', $new, $old;
735 sub update_head_checkout ($$$) {
736 my ($old, $new, $mrest) = @_;
737 my $symref = git_get_symref();
738 runcmd @git, qw(checkout), $new, qw(.);
739 update_head $old, $new, $mrest;
743 badusage "no arguments to launder allowed" if @ARGV;
744 my $old = get_head();
745 my ($tip,$breakwater) = walk $old;
746 update_head $old, $tip, 'launder';
747 # no tree changes except debian/patches
748 runcmd @git, qw(rm --quiet --ignore-unmatch -rf debian/patches);
749 printf "# breakwater tip\n%s\n", $breakwater;
750 printf "# working tip\n%s\n", $tip;
754 die if ($ARGV[0]//'') =~ m/^-/;
755 badusage "too many arguments to analyse" if @ARGV>1;
758 $old = git_rev_parse $old;
762 my ($dummy,$breakwater) = walk $old, 1,*STDOUT;
763 STDOUT->error and die $!;
766 sub cmd_downstream_rebase_launder_v0 () {
767 badusage "needs 1 argument, the baseline" unless @ARGV==1;
769 $base = git_rev_parse $base;
770 my $old_head = get_head();
771 my $current = $old_head;
774 if ($current eq $base) {
775 $topmost_keep //= $current;
776 print " $current BASE stop\n";
779 my $cl = classify $current;
780 print " $current $cl->{Type}";
782 my $p0 = $cl->{Parents}[0]{CommitId};
784 if ($cl->{Type} eq 'Pseudomerge') {
785 print " ^".($cl->{Contributor}{Ix}+1);
786 $next = $cl->{Contributor}{CommitId};
787 } elsif ($cl->{Type} eq 'AddPatches' or
788 $cl->{Type} eq 'Changelog') {
798 $topmost_keep //= $current;
800 die "to-be stripped changes not on top of the branch\n"
805 if ($topmost_keep eq $old_head) {
808 print "updating to $topmost_keep\n";
810 $old_head, $topmost_keep,
811 'downstream-rebase-launder-v0';
815 GetOptions("D+" => \$debuglevel) or die badusage "bad options\n";
816 initdebug('git-debrebase ');
817 enabledebug if $debuglevel;
819 my $toplevel = cmdoutput @git, qw(rev-parse --show-toplevel);
820 chdir $toplevel or die "chdir $toplevel: $!";
822 $rd = fresh_playground "$playprefix/misc";
824 my $cmd = shift @ARGV;
827 $cmdfn = ${*::}{"cmd_$cmdfn"};
829 $cmdfn or badusage "unknown git-debrebase sub-operation $cmd";