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 $wa);
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/, $out;
134 my ($type,$data) = git_cat_file $objid;
135 die unless $type eq 'commit';
136 $data =~ m/(?<=\n)\n/;
140 sub D_DEB () { return 0x1; } # debian/ (not including debian/patches/)
141 sub D_UPS () { return 0x2; } # upstream files
142 sub D_PAT_ADD () { return 0x4; } # debian/patches/ extra patches at end
143 sub D_PAT_OTH () { return 0x8; } # debian/patches other changes
145 our $rd = ".git/git-debrebase";
146 our $ud = "$rd/work";
149 sub get_differs ($$) {
151 # This resembles quiltify_trees_differ, in dgit, a bit.
152 # But we don't care about modes, or dpkg-source-unrepresentable
153 # changes, and we don't need the plethora of different modes.
154 # Conversely we need to distinguish different kinds of changes to
155 # debian/ and debian/patches/.
160 my ($opts, $sfx, $fn) = @_;
161 $opts //= [qw(--name-only)];
162 my @cmd = (@git, qw(diff-tree -z --no-renames));
164 push @cmd, "$_:$sfx" foreach $x, $y;
165 my $diffs = cmdoutput @cmd;
166 foreach (split /\0/, $diffs) { $fn->(); }
169 $rundiff->(undef, '', sub {
170 $differs |= $_ eq 'debian' ? D_DEB : D_UPS;
173 if ($differs & D_DEB) {
175 $rundiff->(undef, ':debian', sub {
176 $differs |= $_ eq 'patches' ? D_PAT_OTH : D_DEB;
178 die "mysterious debian changes $x..$y"
179 unless $differs & (D_PAT_OTH|D_DEB);
182 if ($differs & D_PAT_OTH) {
184 $differs &= ~D_PAT_OTH;
186 $differs |= D_PAT_OTH;
187 no warnings qw(exiting); last;
189 $rundiff->([qw(--name-status)], ':debian/patches', sub {
190 no warnings qw(exiting);
191 if (!defined $mode) {
195 if ($mode eq 'A' && !m/(?:^|\.)series$/s) {
197 } elsif ($mode eq 'M' && $_ eq 'series') {
198 my $x_s = git_cat_file "$x:debian/patches/series", 'blob';
199 my $y_s = git_cat_file "$y:debian/patches/series", 'blob';
200 chomp $x_s; $x_s .= "\n";
201 $ok = $x_s eq substr($y_s, 0, length $x_s);
206 $differs |= $ok ? D_PAT_ADD : D_PAT_OTH;
208 die "mysterious debian/patches changes $x..$y"
209 unless $differs & (D_PAT_ADD|D_PAT_OTH);
212 printdebug sprintf "get_differs %s, %s = %#x\n", $x, $y, $differs;
217 sub commit_pr_info ($) {
219 return Data::Dumper->dump([$r], [qw(commit)]);
222 sub calculate_committer_authline () {
223 my $c = cmdoutput @git, qw(commit-tree --no-gpg-sign -m),
224 'DUMMY COMMIT (git-debrebase)', "HEAD:";
225 my ($h,$m) = get_commit $c;
226 $h =~ m/^committer .*$/m or confess "($h) ?";
230 # classify returns an info hash like this
232 # Hdr => # commit headers, including 1 final newline
233 # Msg => # commit message (so one newline is dropped)
235 # Type => (see below)
237 # Ix => $index # ie 0, 1, 2, ...
239 # Differs => return value from get_differs
241 # IsDggitImport => 'orig' 'tarball' 'unpatched' 'package' (as from dgit)
243 # NewMsg => # commit message, but with any [dgit import ...] edited
244 # # to say "[was: ...]"
254 # has additional entres in classification result
255 # Overwritten = [ subset of Parents ]
256 # Contributor = $the_remaining_Parent
258 # DgitImportUnpatched
259 # has additional entry in classification result
260 # OrigParents = [ subset of Parents ]
262 # BreakwaterUpstreamMerge
263 # has additional entry in classification result
264 # OrigParents = [ subset of Parents ]
269 my ($h,$m) = get_commit $objid;
271 my ($t) = $h =~ m/^tree (\w+)$/m or die $objid;
272 my (@ph) = $h =~ m/^parent (\w+)$/mg;
283 foreach my $ph (@ph) {
287 Differs => (get_differs $t, $ph),
291 printdebug "classify $objid \$t=$t \@p",
292 (map { sprintf " %s/%#x", $_->{CommitId}, $_->{Differs} } @p),
296 my ($type, @rest) = @_;
297 $r = { %$r, Type => $type, @rest };
299 my $dd = new Data::Dumper [ $r ];
300 Terse $dd 1; Indent $dd 0; Useqq $dd 1;
301 printdebug " = $type ".(Dump $dd)."\n";
307 $r = { %$r, Type => qw(Unknown) };
308 printdebug " ** Unknown\n";
313 my $d = $r->{Parents}[0]{Differs};
314 if ($d == D_PAT_ADD) {
315 return $classify->(qw(AddPatches));
316 } elsif ($d & (D_PAT_ADD|D_PAT_OTH)) {
317 return $unknown->("edits debian/patches");
318 } elsif ($d == D_DEB) {
319 my ($ty,$dummy) = git_cat_file "$ph[0]:debian";
321 return $classify->(qw(Packaging));
322 } elsif ($ty eq 'missing') {
323 return $classify->(qw(BreakwaterStart));
325 return $unknown->("parent's debian is not a directory");
327 } elsif ($d == D_UPS) {
328 return $classify->(qw(Upstream));
329 } elsif ($d == (D_DEB|D_UPS)) {
330 return $classify->(qw(Mixed));
332 return $unknown->("no changes");
334 confess "internal error $objid ?";
338 return $unknown->("origin commit");
341 my @identical = grep { !$_->{Differs} } @p;
342 if (@p == 2 && @identical == 1) {
343 my @overwritten = grep { $_->{Differs} } @p;
344 confess "internal error $objid ?" unless @overwritten==1;
345 return $classify->(qw(Pseudomerge),
346 Overwritten => $overwritten[0],
347 Contributor => $identical[0]);
349 if (@p == 2 && @identical == 2) {
350 my @bytime = nsort_by {
351 my ($ph,$pm) = get_commit $_->{CommitId};
352 $ph =~ m/^committer .* (\d+) [-+]\d+$/m or die "$_->{CommitId} ?";
355 return $classify->(qw(Pseudomerge),
356 SubType => qw(Ambiguous),
357 Overwritten => $bytime[0],
358 Contributor => $bytime[1]);
361 my ($p_h, $p_m) = get_commit $p;
362 $p->{IsOrigin} = $p_h !~ m/^parent \w+$/m;
363 ($p->{IsDgitImport},) = $p_m =~ m/^\[dgit import ([0-9a-z]+) .*\]$/m;
365 my @orig_ps = grep { ($_->{IsDgitImport}//'X') eq 'orig' } @p;
367 if (!(grep { !$_->{IsOrigin} } @p) and
368 (@orig_ps >= @p - 1) and
369 $m2 =~ s{^\[(dgit import unpatched .*)\]$}{[was: $1]}m) {
371 return $classify->(qw(DgitImportUnpatched),
372 OrigParents => \@orig_ps);
375 my ($stype, $series) = git_cat_file "$t:debian/patches/series";
376 my $haspatches = $stype ne 'missing' && $series =~ m/^\s*[^#\n\t ]/m;
378 # How to decide about l/r ordering of breakwater merges ? git
379 # --topo-order prefers to expand 2nd parent first. There's
380 # already an easy rune to look for debian/ history anyway (git log
381 # debian/) so debian breakwater branch should be 1st parent; that
382 # way also there's also an easy rune to look for the upstream
383 # patches (--topo-order).
386 !$p[0]{IsOrigin} && # breakwater merge never starts with an origin
387 !($p[0]{Differs} & ~D_DEB) &&
388 !($p[1]{Differs} & ~D_UPS)) {
389 return $classify->(qw(BreakwaterUpstreamMerge),
390 OrigParents => [ $p[1] ]);
392 # xxx multi-.orig upstreams
394 return $unknown->("complex merge");
400 $nogenerate,$report) = @_;
401 # => ($tip, $breakwater_tip)
402 # (or nothing, if $nogenerate)
404 # go through commits backwards
405 # we generate two lists of commits to apply:
406 # breakwater branch and upstream patches
407 my (@brw_cl, @upp_cl, @processed);
414 my ($appendinfo) = @_;
417 $ms .= "\n\n[git-debrebase $appendinfo]\n";
420 my $rewrite_from_here = sub {
421 my $sp_cl = { SpecialMethod => 'StartRewrite' };
422 push @brw_cl, $sp_cl;
423 push @processed, $sp_cl;
428 my $prprdelim = sub { print $report $prdelim if $report; $prdelim=""; };
431 return unless $report;
432 print $report $prdelim, @_;
436 my $bomb = sub { # usage: return $bomb->();
437 print $report " Unprocessable" if $report;
440 return (undef,undef);
442 die "commit $cur: Cannot cope with this commit";
448 my $build_start = sub {
449 my ($msg, $parent) = @_;
452 no warnings qw(exiting); last;
457 my $ty = $cl->{Type};
458 my $st = $cl->{SubType};
459 $prline->("$cl->{CommitId} $cl->{Type}");
460 $found{$ty. ( defined($st) ? "-$st" : '' )}++;
461 push @processed, $cl;
462 my $p0 = @{ $cl->{Parents} }==1 ? $cl->{Parents}[0]{CommitId} : undef;
463 if ($ty eq 'AddPatches') {
465 $rewrite_from_here->();
467 } elsif ($ty eq 'Packaging') {
471 } elsif ($ty eq 'BreakwaterStart') {
472 $build_start->('FirstPackaging', $cur);
473 } elsif ($ty eq 'Upstream') {
477 } elsif ($ty eq 'Mixed') {
480 my $cls = { $cl, $xmsg->("split mixed commit: $wh part") };
483 $queue->(\@brw_cl, "debian");
484 $queue->(\@upp_cl, "upstream");
485 $rewrite_from_here->();
488 } elsif ($ty eq 'Pseudomerge') {
489 print $report " Contributor=$cl->{Contributor}" if $report;
490 push @pseudomerges, $cl;
491 $rewrite_from_here->();
492 $cur = $ty->{Contributor};
494 } elsif ($ty eq 'BreakwaterUpstreamMerge') {
495 $build_start->("PreviousBreakwater", $cur);
496 } elsif ($ty eq 'DgitImportUnpatched') {
497 my $pm = $pseudomerges[-1];
499 # To an extent, this is heuristic. Imports don't have
500 # a useful history of the debian/ branch. We assume
501 # that the first pseudomerge after an import has a
502 # useful history of debian/, and ignore the histories
503 # from later pseudomerges. Often the first pseudomerge
504 # will be the dgit import of the upload to the actual
505 # suite intended by the non-dgit NMUer, and later
506 # pseudomerges may represent in-archive copies.
507 my $ovwrs = $pm->{Overwritten};
508 printf $report " PM=%s \@Overwr:%d", $pm, (scalar @$ovwrs)
513 my $ovwr = $ovwrs->[0]{CommitId};
514 printf $report " Overwr=%s", $ovwr if $report;
515 # This import has a tree which is just like a
516 # breakwater tree, but it has the wrong history. It
517 # ought to have the previous breakwater (which the
518 # pseudomerge overwrote) as an ancestor. That will
519 # make the history of the debian/ files correct. As
520 # for the upstream version: either it's the same as
521 # was ovewritten (ie, same as the previous
522 # breakwater), in which case that history is precisely
523 # right; or, otherwise, it was a non-gitish upload of a
524 # new upstream version. We can tell these apart by
525 # looking at the tree of the supposed upstream.
528 SpecialMethod => 'DgitImportDebianUpdate',
529 $xmsg->("convert dgit import: debian changes")
531 my $differs = (get_differs $ovwr, $cl->{Tree});
532 printf $report " Differs=%#x", $differs if $report;
533 if ($differs & D_UPS) {
534 printf $report " D_UPS" if $report;
535 # This will also trigger if a non-dgit git-based NMU
536 # deleted .gitignore (which is a thing that some of
537 # the existing git tools do if the user doesn't
538 # somehow tell them not to). Ah well.
541 SpecialMethod => 'DgitImportUpstreamUpdate',
542 $xmsg->("convert dgit import: upstream changes")
545 $prline->(" Import");
546 $rewrite_from_here->();
547 $upp_limit //= $#upp_cl; # further, deeper, patches discarded
551 # Everything is from this import. This kind of import
552 # is already in valid breakwater format, with the
553 # patches as commits.
554 printf $report " NoPM" if $report;
555 # last thing we processed will have been the first patch,
556 # if there is one; which is fine, so no need to rewrite
557 # on account of this import
558 $build_start->("ImportOrigin", $cur);
566 return if $nogenerate;
568 # Now we build it back up again
574 my $rm_tree_cached = sub {
576 runcmd @git, qw(rm --quiet -rf --cached), $subdir;
578 my $read_tree_debian = sub {
580 $rm_tree_cached->(qw(debian));
581 runcmd @git, qw(read-tree --prefix=debian/), "$treeish:debian";
583 my $read_tree_upstream = sub {
585 runcmd @git, qw(read-tree), $treeish;
586 $read_tree_debian->($build);
589 my $committer_authline = calculate_committer_authline();
592 mkdir $rd or $!==EEXIST or die $!;
594 foreach my $cl (qw(Debian), (reverse @brw_cl),
595 { SpecialMethod => 'RecordBreakwaterTip' },
596 qw(Upstream), (reverse @upp_cl)) {
598 $current_method = $cl;
601 my $method = $cl->{SpecialMethod} // $current_method;
602 my @parents = ($build);
603 my $cltree = $cl->{CommitId};
604 if ($method eq 'Debian') {
605 $read_tree_debian->($cltree);
606 } elsif ($method eq 'Upstream') {
607 $read_tree_upstream->($cltree);
608 } elsif ($method eq 'StartRewrite') {
611 } elsif ($method eq 'RecordBreakwaterTip') {
612 $breakwater = $build;
614 } elsif ($method eq 'DgitImportDebianUpdate') {
615 $read_tree_debian->($cltree);
616 $rm_tree_cached->(qw(debian/patches));
617 } elsif ($method eq 'DgitImportUpstreamUpdate') {
618 $read_tree_upstream->($cltree);
619 push @parents, map { $_->{CommitId} } @{ $cl->{OrigParents} };
623 $rewriting ||= $cl ne pop @processed;
624 my $newtree = cmdoutput @git, qw(write-tree);
626 $ch =~ s{^tree .*}{tree $newtree}m or confess "$ch ?";
627 $ch =~ s{^parent .*\n}{}m;
628 $ch =~ s{(?=^author)}{
629 map { "parent $_\n" } @parents
630 }me or confess "$ch ?";
632 $ch =~ s{^committer .*$}{$committer_authline}m
635 my $cf = "$rd/m$rewriting";
636 open CD, ">", $cf or die $!;
637 print CD $ch, "\n", $cl->{Msg} or die $!;
639 my @cmd = (@git, qw(hash-object));
640 push @cmd, qw(-w) if $rewriting;
641 push @cmd, qw(-t commit), $cf;
642 my $newcommit = cmdoutput @cmd;
643 confess "$ch ?" unless $rewriting or $newcommit eq $cl->{CommitId};
648 runcmd @git, qw(diff-tree --quiet), $input, $build;
650 return ($build, $breakwater);
653 sub get_head () { return git_rev_parse qw(HEAD); }
655 sub update_head ($$$) {
656 my ($old, $new, $mrest) = @_;
657 runcmd @git, qw(update-ref -m), "git-debrebase $mrest", $new, $old;
661 badusage "no arguments to launder allowed";
662 my $old = get_head();
663 my ($tip,$breakwater) = walk $old;
664 update_head $old, $tip, 'launder';
665 # no tree changes except debian/patches
666 runcmd @git, qw(rm --quiet -rf debian/patches);
667 printf "# breakwater tip\n%s\n", $breakwater;
671 die if ($ARGV[0]//'') =~ m/^-/;
672 badusage "too many arguments to analyse" if @ARGV>1;
675 $old = git_rev_parse $old;
679 my ($dummy,$breakwater) = walk $old, 1,*STDOUT;
680 STDOUT->error and die $!;
683 GetOptions("D+" => \$debuglevel) or die badusage "bad options\n";
684 initdebug('git-debrebase ');
685 enabledebug if $debuglevel;
687 my $toplevel = cmdoutput @git, qw(rev-parse --show-toplevel);
688 chdir $toplevel or die "chdir $toplevel: $!";
690 my $cmd = shift @ARGV;
693 $cmdfn = ${*::}{"cmd_$cmdfn"};
695 $cmdfn or badusage "unknown git-debrebase sub-operation $cmd";