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 ?
111 use Getopt::Long qw(:config posix_default gnu_compat bundling);
113 use Debian::Dgit qw(:DEFAULT :playground);
117 die "bad usage: $m\n";
123 my @cmd = qw(git config -z);
124 push @cmd, qw(--get-all) if wantarray;
126 my $out = cmdoutput @cmd;
127 return split /\0/, $out;
134 my $data = git_cat_file $objid, 'commit';
135 $data =~ m/(?<=\n)\n/ or die "$objid ($data) ?";
139 sub D_UPS () { 0x02; } # upstream files
140 sub D_PAT_ADD () { 0x04; } # debian/patches/ extra patches at end
141 sub D_PAT_OTH () { 0x08; } # debian/patches other changes
142 sub D_DEB_CLOG () { 0x10; } # debian/ (not patches/ or changelog)
143 sub D_DEB_OTH () { 0x20; } # debian/changelog
144 sub DS_DEB () { D_DEB_CLOG | D_DEB_OTH; } # debian/ (not patches/)
146 our $playprefix = 'debrebase';
152 sub in_workarea ($) {
155 my $r = eval { $sub->(); };
159 sub fresh_workarea () {
160 $workarea = fresh_playground "$playprefix/work";
161 in_workarea sub { playtree_setup };
164 sub get_differs ($$) {
166 # This resembles quiltify_trees_differ, in dgit, a bit.
167 # But we don't care about modes, or dpkg-source-unrepresentable
168 # changes, and we don't need the plethora of different modes.
169 # Conversely we need to distinguish different kinds of changes to
170 # debian/ and debian/patches/.
175 my ($opts, $limits, $fn) = @_;
176 my @cmd = (@git, qw(diff-tree -z --no-renames));
178 push @cmd, "$_:" foreach $x, $y;
180 my $diffs = cmdoutput @cmd;
181 foreach (split /\0/, $diffs) { $fn->(); }
184 $rundiff->([qw(--name-only)], [], sub {
185 $differs |= $_ eq 'debian' ? DS_DEB : D_UPS;
188 if ($differs & DS_DEB) {
190 $rundiff->([qw(--name-only -r)], [qw(debian)], sub {
192 m{^debian/patches/} ? D_PAT_OTH :
193 $_ eq 'debian/changelog' ? D_DEB_CLOG :
196 die "mysterious debian changes $x..$y"
197 unless $differs & (D_PAT_OTH|DS_DEB);
200 if ($differs & D_PAT_OTH) {
202 $differs &= ~D_PAT_OTH;
204 $differs |= D_PAT_OTH;
205 no warnings qw(exiting); last;
207 $rundiff->([qw(--name-status -r)], [qw(debian/patches/)], sub {
208 no warnings qw(exiting);
209 if (!defined $mode) {
212 die unless s{^debian/patches/}{};
214 if ($mode eq 'A' && !m/(?:^|\.)series$/s) {
216 } elsif ($mode eq 'M' && $_ eq 'series') {
217 my $x_s = git_cat_file "$x:debian/patches/series", 'blob';
218 my $y_s = git_cat_file "$y:debian/patches/series", 'blob';
219 chomp $x_s; $x_s .= "\n";
220 $ok = $x_s eq substr($y_s, 0, length $x_s);
225 $differs |= $ok ? D_PAT_ADD : D_PAT_OTH;
227 die "mysterious debian/patches changes $x..$y"
228 unless $differs & (D_PAT_ADD|D_PAT_OTH);
231 printdebug sprintf "get_differs %s, %s = %#x\n", $x, $y, $differs;
236 sub commit_pr_info ($) {
238 return Data::Dumper->dump([$r], [qw(commit)]);
241 sub calculate_committer_authline () {
242 my $c = cmdoutput @git, qw(commit-tree --no-gpg-sign -m),
243 'DUMMY COMMIT (git-debrebase)', "HEAD:";
244 my ($h,$m) = get_commit $c;
245 $h =~ m/^committer .*$/m or confess "($h) ?";
249 # classify returns an info hash like this
251 # Hdr => # commit headers, including 1 final newline
252 # Msg => # commit message (so one newline is dropped)
254 # Type => (see below)
256 # Ix => $index # ie 0, 1, 2, ...
258 # Differs => return value from get_differs
260 # IsDggitImport => 'orig' 'tarball' 'unpatched' 'package' (as from dgit)
262 # NewMsg => # commit message, but with any [dgit import ...] edited
263 # # to say "[was: ...]"
274 # has additional entres in classification result
275 # Overwritten = [ subset of Parents ]
276 # Contributor = $the_remaining_Parent
278 # DgitImportUnpatched
279 # has additional entry in classification result
280 # OrigParents = [ subset of Parents ]
282 # BreakwaterUpstreamMerge
283 # has additional entry in classification result
284 # OrigParents = [ subset of Parents ]
289 my ($h,$m) = get_commit $objid;
291 my ($t) = $h =~ m/^tree (\w+)$/m or die $objid;
292 my (@ph) = $h =~ m/^parent (\w+)$/mg;
303 foreach my $ph (@ph) {
307 Differs => (get_differs $t, $ph),
311 printdebug "classify $objid \$t=$t \@p",
312 (map { sprintf " %s/%#x", $_->{CommitId}, $_->{Differs} } @p),
316 my ($type, @rest) = @_;
317 $r = { %$r, Type => $type, @rest };
319 my $dd = new Data::Dumper [ $r ];
320 Terse $dd 1; Indent $dd 0; Useqq $dd 1;
321 printdebug " = $type ".(Dump $dd)."\n";
327 $r = { %$r, Type => qw(Unknown) };
328 printdebug " ** Unknown\n";
333 my $d = $r->{Parents}[0]{Differs};
334 if ($d == D_PAT_ADD) {
335 return $classify->(qw(AddPatches));
336 } elsif ($d & (D_PAT_ADD|D_PAT_OTH)) {
337 return $unknown->("edits debian/patches");
338 } elsif ($d & DS_DEB and !($d & ~DS_DEB)) {
339 my ($ty,$dummy) = git_cat_file "$ph[0]:debian";
341 if ($d == D_DEB_CLOG) {
342 return $classify->(qw(Changelog));
344 return $classify->(qw(Packaging));
346 } elsif ($ty eq 'missing') {
347 return $classify->(qw(BreakwaterStart));
349 return $unknown->("parent's debian is not a directory");
351 } elsif ($d == D_UPS) {
352 return $classify->(qw(Upstream));
353 } elsif ($d & DS_DEB and $d & D_UPS and !($d & ~(DS_DEB|D_UPS))) {
354 return $classify->(qw(Mixed));
356 return $unknown->("no changes");
358 confess "internal error $objid ?";
362 return $unknown->("origin commit");
365 my @identical = grep { !$_->{Differs} } @p;
366 if (@p == 2 && @identical == 1) {
367 my @overwritten = grep { $_->{Differs} } @p;
368 confess "internal error $objid ?" unless @overwritten==1;
369 return $classify->(qw(Pseudomerge),
370 Overwritten => $overwritten[0],
371 Contributor => $identical[0]);
373 if (@p == 2 && @identical == 2) {
374 my @bytime = nsort_by {
375 my ($ph,$pm) = get_commit $_->{CommitId};
376 $ph =~ m/^committer .* (\d+) [-+]\d+$/m or die "$_->{CommitId} ?";
379 return $classify->(qw(Pseudomerge),
380 SubType => qw(Ambiguous),
381 Overwritten => $bytime[0],
382 Contributor => $bytime[1]);
385 my ($p_h, $p_m) = get_commit $p->{CommitId};
386 $p->{IsOrigin} = $p_h !~ m/^parent \w+$/m;
387 ($p->{IsDgitImport},) = $p_m =~ m/^\[dgit import ([0-9a-z]+) .*\]$/m;
389 my @orig_ps = grep { ($_->{IsDgitImport}//'X') eq 'orig' } @p;
391 if (!(grep { !$_->{IsOrigin} } @p) and
392 (@orig_ps >= @p - 1) and
393 $m2 =~ s{^\[(dgit import unpatched .*)\]$}{[was: $1]}m) {
395 return $classify->(qw(DgitImportUnpatched),
396 OrigParents => \@orig_ps);
399 my ($stype, $series) = git_cat_file "$t:debian/patches/series";
400 my $haspatches = $stype ne 'missing' && $series =~ m/^\s*[^#\n\t ]/m;
402 # How to decide about l/r ordering of breakwater merges ? git
403 # --topo-order prefers to expand 2nd parent first. There's
404 # already an easy rune to look for debian/ history anyway (git log
405 # debian/) so debian breakwater branch should be 1st parent; that
406 # way also there's also an easy rune to look for the upstream
407 # patches (--topo-order).
409 # The above tells us which way *we* will generate them. But we
410 # might encounter ad-hoc breakwater merges generated manually,
411 # which might be the other way around. In principle, in some odd
412 # situations, a breakwater merge might have two identical parents.
413 # In that case we guess which way round it is (ie, which parent
414 # has the upstream history). The order of the 2-iteration loop
415 # controls which guess we make.
417 foreach my $prevbrw (qw(0 1)) {
420 !$p[$prevbrw]{IsOrigin} && # breakwater never starts with an origin
421 !($p[$prevbrw]{Differs} & ~DS_DEB) &&
422 !($p[!$prevbrw]{Differs} & ~D_UPS)) {
423 return $classify->(qw(BreakwaterUpstreamMerge),
424 OrigParents => [ $p[!$prevbrw] ]);
426 # xxx multi-.orig upstreams
429 return $unknown->("complex merge");
435 $nogenerate,$report) = @_;
436 # => ($tip, $breakwater_tip)
437 # (or nothing, if $nogenerate)
439 # go through commits backwards
440 # we generate two lists of commits to apply:
441 # breakwater branch and upstream patches
442 my (@brw_cl, @upp_cl, @processed);
449 my ($appendinfo) = @_;
452 $ms .= "\n\n[git-debrebase $appendinfo]\n";
455 my $rewrite_from_here = sub {
456 my $sp_cl = { SpecialMethod => 'StartRewrite' };
457 push @brw_cl, $sp_cl;
458 push @processed, $sp_cl;
463 my $prprdelim = sub { print $report $prdelim if $report; $prdelim=""; };
466 return unless $report;
467 print $report $prdelim, @_;
471 my $bomb = sub { # usage: return $bomb->();
472 print $report " Unprocessable" if $report;
475 return (undef,undef);
477 die "commit $cur: Cannot cope with this commit";
483 my $build_start = sub {
484 my ($msg, $parent) = @_;
487 no warnings qw(exiting); last;
492 my $ty = $cl->{Type};
493 my $st = $cl->{SubType};
494 $prline->("$cl->{CommitId} $cl->{Type}");
495 $found{$ty. ( defined($st) ? "-$st" : '' )}++;
496 push @processed, $cl;
497 my $p0 = @{ $cl->{Parents} }==1 ? $cl->{Parents}[0]{CommitId} : undef;
498 if ($ty eq 'AddPatches') {
500 $rewrite_from_here->();
502 } elsif ($ty eq 'Packaging' or $ty eq 'Changelog') {
506 } elsif ($ty eq 'BreakwaterStart') {
507 $build_start->('FirstPackaging', $cur);
508 } elsif ($ty eq 'Upstream') {
512 } elsif ($ty eq 'Mixed') {
515 my $cls = { $cl, $xmsg->("split mixed commit: $wh part") };
518 $queue->(\@brw_cl, "debian");
519 $queue->(\@upp_cl, "upstream");
520 $rewrite_from_here->();
523 } elsif ($ty eq 'Pseudomerge') {
524 my $contrib = $cl->{Contributor}{CommitId};
525 print $report " Contributor=$contrib" if $report;
526 push @pseudomerges, $cl;
527 $rewrite_from_here->();
530 } elsif ($ty eq 'BreakwaterUpstreamMerge') {
531 $build_start->("PreviousBreakwater", $cur);
532 } elsif ($ty eq 'DgitImportUnpatched') {
533 my $pm = $pseudomerges[-1];
535 # To an extent, this is heuristic. Imports don't have
536 # a useful history of the debian/ branch. We assume
537 # that the first pseudomerge after an import has a
538 # useful history of debian/, and ignore the histories
539 # from later pseudomerges. Often the first pseudomerge
540 # will be the dgit import of the upload to the actual
541 # suite intended by the non-dgit NMUer, and later
542 # pseudomerges may represent in-archive copies.
543 my $ovwrs = $pm->{Overwritten};
544 printf $report " PM=%s \@Overwr:%d", $pm, (scalar @$ovwrs)
549 my $ovwr = $ovwrs->[0]{CommitId};
550 printf $report " Overwr=%s", $ovwr if $report;
551 # This import has a tree which is just like a
552 # breakwater tree, but it has the wrong history. It
553 # ought to have the previous breakwater (which the
554 # pseudomerge overwrote) as an ancestor. That will
555 # make the history of the debian/ files correct. As
556 # for the upstream version: either it's the same as
557 # was ovewritten (ie, same as the previous
558 # breakwater), in which case that history is precisely
559 # right; or, otherwise, it was a non-gitish upload of a
560 # new upstream version. We can tell these apart by
561 # looking at the tree of the supposed upstream.
564 SpecialMethod => 'DgitImportDebianUpdate',
565 $xmsg->("convert dgit import: debian changes")
567 my $differs = (get_differs $ovwr, $cl->{Tree});
568 printf $report " Differs=%#x", $differs if $report;
569 if ($differs & D_UPS) {
570 printf $report " D_UPS" if $report;
571 # This will also trigger if a non-dgit git-based NMU
572 # deleted .gitignore (which is a thing that some of
573 # the existing git tools do if the user doesn't
574 # somehow tell them not to). Ah well.
577 SpecialMethod => 'DgitImportUpstreamUpdate',
578 $xmsg->("convert dgit import: upstream changes")
581 $prline->(" Import");
582 $rewrite_from_here->();
583 $upp_limit //= $#upp_cl; # further, deeper, patches discarded
587 # Everything is from this import. This kind of import
588 # is already in valid breakwater format, with the
589 # patches as commits.
590 printf $report " NoPM" if $report;
591 # last thing we processed will have been the first patch,
592 # if there is one; which is fine, so no need to rewrite
593 # on account of this import
594 $build_start->("ImportOrigin", $cur);
602 return if $nogenerate;
604 # Now we build it back up again
610 my $rm_tree_cached = sub {
612 runcmd @git, qw(rm --quiet -rf --cached --ignore-unmatch), $subdir;
614 my $read_tree_debian = sub {
616 $rm_tree_cached->(qw(debian));
617 runcmd @git, qw(read-tree --prefix=debian/), "$treeish:debian";
619 my $read_tree_upstream = sub {
621 runcmd @git, qw(read-tree), $treeish;
622 $read_tree_debian->($build);
625 my $committer_authline = calculate_committer_authline();
628 mkdir $rd or $!==EEXIST or die $!;
630 foreach my $cl (qw(Debian), (reverse @brw_cl),
631 { SpecialMethod => 'RecordBreakwaterTip' },
632 qw(Upstream), (reverse @upp_cl)) {
634 $current_method = $cl;
637 my $method = $cl->{SpecialMethod} // $current_method;
638 my @parents = ($build);
639 my $cltree = $cl->{CommitId};
640 if ($method eq 'Debian') {
641 $read_tree_debian->($cltree);
642 } elsif ($method eq 'Upstream') {
643 $read_tree_upstream->($cltree);
644 } elsif ($method eq 'StartRewrite') {
647 } elsif ($method eq 'RecordBreakwaterTip') {
648 $breakwater = $build;
650 } elsif ($method eq 'DgitImportDebianUpdate') {
651 $read_tree_debian->($cltree);
652 $rm_tree_cached->(qw(debian/patches));
653 } elsif ($method eq 'DgitImportUpstreamUpdate') {
654 $read_tree_upstream->($cltree);
655 push @parents, map { $_->{CommitId} } @{ $cl->{OrigParents} };
659 $rewriting ||= $cl ne pop @processed;
660 my $newtree = cmdoutput @git, qw(write-tree);
662 $ch =~ s{^tree .*}{tree $newtree}m or confess "$ch ?";
663 $ch =~ s{^parent .*\n}{}m;
664 $ch =~ s{(?=^author)}{
665 map { "parent $_\n" } @parents
666 }me or confess "$ch ?";
668 $ch =~ s{^committer .*$}{$committer_authline}m
671 my $cf = "$rd/m$rewriting";
672 open CD, ">", $cf or die $!;
673 print CD $ch, "\n", $cl->{Msg} or die $!;
675 my @cmd = (@git, qw(hash-object));
676 push @cmd, qw(-w) if $rewriting;
677 push @cmd, qw(-t commit), $cf;
678 my $newcommit = cmdoutput @cmd;
679 confess "$ch ?" unless $rewriting or $newcommit eq $cl->{CommitId};
684 runcmd @git, qw(diff-tree --quiet), $input, $build;
686 return ($build, $breakwater);
689 sub get_head () { return git_rev_parse qw(HEAD); }
691 sub update_head ($$$) {
692 my ($old, $new, $mrest) = @_;
693 runcmd @git, qw(update-ref -m), "debrebase: $mrest", 'HEAD', $new, $old;
696 sub update_head_checkout ($$$) {
697 my ($old, $new, $mrest) = @_;
698 my $symref = git_get_symref();
699 runcmd @git, qw(checkout), $new, qw(.);
700 update_head $old, $new, $mrest;
704 badusage "no arguments to launder allowed" if @ARGV;
705 my $old = get_head();
706 my ($tip,$breakwater) = walk $old;
707 update_head $old, $tip, 'launder';
708 # no tree changes except debian/patches
709 runcmd @git, qw(rm --quiet -rf debian/patches);
710 printf "# breakwater tip\n%s\n", $breakwater;
714 die if ($ARGV[0]//'') =~ m/^-/;
715 badusage "too many arguments to analyse" if @ARGV>1;
718 $old = git_rev_parse $old;
722 my ($dummy,$breakwater) = walk $old, 1,*STDOUT;
723 STDOUT->error and die $!;
726 sub cmd_downstream_rebase_launder_v0 () {
727 badusage "needs 1 argument, the baseline" unless @ARGV=0;
729 $base = git_rev_parse $base;
730 my $old_head = get_head();
731 my $current = $old_head;
734 if ($current eq $base) {
735 $topmust_keep //= $current;
736 print "$current BASE: stopping\n";
739 my $cl = classify $current;
740 print " $current $cl->{Type}";
742 my $p0 = $cl->{Parents}[0]{CommitId};
743 if ($cl->{Type} eq 'Pseudomerge') {
744 $current = $cl->{Contributor}{CommitId};
745 print " ^".($cl->{Contributor}{Ix}+1);
746 } elsif ($cl->{Type} eq 'AddPatches' or
747 $cl->{Type} eq 'Changelog') {
757 $topmost_keep //= $current;
759 die "to-be stripped changes not on top of the branch\n"
760 if $topmost_unstripped;
763 if ($topmost_keep eq $old_head) {
766 print "updating to $topmost_keep\n";
767 update_head $old_head, $topmost_keep, 'downstream-rebase-launder-v0';
771 GetOptions("D+" => \$debuglevel) or die badusage "bad options\n";
772 initdebug('git-debrebase ');
773 enabledebug if $debuglevel;
775 my $toplevel = cmdoutput @git, qw(rev-parse --show-toplevel);
776 chdir $toplevel or die "chdir $toplevel: $!";
778 $rd = fresh_playground "$playprefix/misc";
780 my $cmd = shift @ARGV;
783 $cmdfn = ${*::}{"cmd_$cmdfn"};
785 $cmdfn or badusage "unknown git-debrebase sub-operation $cmd";