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 ?
112 use Debian::Dgit qw(:DEFAULT $wa);
116 die "bad usage: $m\n";
122 my @cmd = qw(git config -z);
123 push @cmd, qw(--get-all) if wantarray;
125 my $out = cmdoutput @cmd;
126 return split /\0/, $out;
133 my ($type,$data) = git_cat_file $objid;
134 die unless $type eq 'commit';
135 $data =~ m/(?<=\n)\n/;
139 sub D_DEB () { return 0x1; } # debian/ (not including debian/patches/)
140 sub D_UPS () { return 0x2; } # upstream files
141 sub D_PAT_ADD () { return 0x4; } # debian/patches/ extra patches at end
142 sub D_PAT_OTH () { return 0x8; } # debian/patches other changes
144 our $rd = ".git/git-debrebase";
145 our $ud = "$rd/work";
148 sub get_differs ($$) {
150 # This resembles quiltify_trees_differ, in dgit, a bit.
151 # But we don't care about modes, or dpkg-source-unrepresentable
152 # changes, and we don't need the plethora of different modes.
153 # Conversely we need to distinguish different kinds of changes to
154 # 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 @bcmd, $x, $y;
166 foreach $f (split /\0/, $diffs) { $fn->(); }
169 $rundiff(undef, '', sub {
170 $differs |= $f eq 'debian' ? D_DEB : D_UPS;
173 if ($differs & D_DEB) {
175 $rundiff(undef, ':debian', sub {
176 $differs |= $f 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 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' && $f !~ m/(?:^|\.)series$/s) {
197 } elsif ($mode eq 'M' && $f 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 $differs & (D_PAT_ADD|D_PAT_OTH);
215 sub commit_pr_info ($) {
217 return Data::Dumper->dump([$r], [qw(commit)]);
220 sub calculate_committer_authline () {
221 my $c = cmdoutput @git, qw(commit-tree --no-gpg-sign -m),
222 'DUMMY COMMIT (git-debrebase)', "HEAD:";
223 my ($h,$m) = get_commit $c;
224 $h =~ m/^committer .*$/m or confess "($h) ?";
228 # classify returns an info hash like this
230 # Hdr => # commit headers, including 1 final newline
231 # Msg => # commit message (so one newline is dropped)
233 # Type => (see below)
235 # Ix => $index # ie 0, 1, 2, ...
237 # Differs => return value from get_differs
239 # IsDggitImport => 'orig' 'tarball' 'unpatched' 'package' (as from dgit)
241 # NewMsg => # commit message, but with any [dgit import ...] edited
242 # # to say "[was: ...]"
252 # has additional entres in classification result
253 # Overwritten = [ subset of Parents ]
254 # Contributor = $the_remaining_Parent
256 # DgitImportUnpatched
257 # has additional entry in classification result
258 # OrigParents = [ subset of Parents ]
260 # BreakwaterUpstreamMerge
261 # has additional entry in classification result
262 # OrigParents = [ subset of Parents ]
267 my ($h,$m) = get_commit $objid;
269 my ($t) = $h =~ m/^tree (\w+)$/m or die $objid;
270 my (@ph) = $h =~ m/^parent (\w+)$/m;
281 foreach my $ph (@ph) {
285 Differs => (get_differs $t, $ph),
290 my ($type, @rest) = @_;
291 $r = { %$r, Type => $type, @rest };
296 $r = { %$r, Type => qw(Unknown) };
301 my $d = $r->{Parents}[0]{Differs};
302 if ($d == D_PAT_ADD) {
303 return $classify->(qw(AddPatches));
304 } elsif ($d & (D_PAT_ADD|D_PAT_OTH)) {
305 return $unknown->("edits debian/patches");
306 } elsif ($d == D_DEB) {
307 my ($ty,$dummy) = git_cat_file "$ph[0]:debian";
309 return $classify->(qw(Packaging));
310 } elsif ($ty eq 'missing') {
311 return $classify->(qw(BreakwaterStart));
313 return $unknown->("parent's debian is not a directory");
315 } elsif ($d == D_UPS) {
316 return $classify->(qw(Upstream));
317 } elsif ($d == (D_DEB|D_UPS)) {
318 return $classify->(qw(Mixed));
320 return $unknown->("no changes");
322 confess "internal error $objid ?";
326 return $unknown->("origin commit");
329 my @identical = grep { !$_->{Differs} } @p;
330 if (@p == 2 && @identical == 1) {
331 my @overwritten = grep { $_->{Differs} } @p;
332 confess "internal error $objid ?" unless @overwritten==1;
333 return $classify->(qw(Pseudomerge),
334 Overwritten => $overwritten[0],
335 Contributor => $identical[0]);
337 if (@p == 2 && @identical == 2) {
338 my @bytime = nsort_by {
339 my ($ph,$pm) = get_commit $_->{CommitId};
340 $ph =~ m/^committer .* (\d+) [-+]\d+$/m or die "$_->{CommitId} ?";
343 return $classify->(qw(Pseudomerge),
344 SubType => qw(Ambiguous),
345 Overwritten => $bytime[0],
346 Contributor => $bytime[1]);
349 my ($p_h, $p_m) = get_commit $p;
350 $p->{IsOrigin} = $p_h !~ m/^parent \w+$/m;
351 ($p->{IsDgitImport},) = $p_m =~ m/^\[dgit import ([0-9a-z]+) .*\]$/m;
353 my @orig_ps = grep { ($_->{IsDgitImport}//'X') eq 'orig' } @p;
355 if (!(grep { !$_->{IsOrigin} } @p) and
356 (@orig_ps >= @p - 1) and
357 $m2 =~ s{^\[(dgit import unpatched .*)\]$}{[was: $1]}m) {
359 return $classify->(qw(DgitImportUnpatched),
360 OrigParents => \@orig_ps);
363 my ($stype, $series) = git_cat_file "$t:debian/patches/series";
364 my $haspatches = $stype ne 'missing' && $series =~ m/^\s*[^#\n\t ]/m;
366 # How to decide about l/r ordering of breakwater merges ? git
367 # --topo-order prefers to expand 2nd parent first. There's
368 # already an easy rune to look for debian/ history anyway (git log
369 # debian/) so debian breakwater branch should be 1st parent; that
370 # way also there's also an easy rune to look for the upstream
371 # patches (--topo-order).
374 !$p[0]{IsOrigin} && # breakwater merge never starts with an origin
375 !($p[0]{Differs} & ~D_DEB) &&
376 !($p[1]{Differs} & ~D_UPS)) {
377 return $classify->(qw(BreakwaterUpstreamMerge),
378 OrigParents => [ $p[1] ]);
380 # xxx multi-.orig upstreams
382 return $unknown->("complex merge");
388 $nogenerate,$report) = @_;
389 # => ($tip, $breakwater_tip)
390 # (or nothing, if $nogenerate)
392 # go through commits backwards
393 # we generate two lists of commits to apply:
394 # breakwater branch and upstream patches
395 my (@brw_cl, @upp_cl, @processed);
402 my ($appendinfo) = @_;
405 $ms .= "\n\n[git-debrebase $appendinfo]\n";
408 my $rewrite_from_here = sub {
409 my $sp_cl = { SpecialMethod => 'StartRewrite' };
410 push @brw_cl, $sp_cl;
411 push @processed, $sp_cl;
416 my $prprdelim = sub { print $report $prdelim if $report; $prdelim=""; };
419 return unless $report;
420 print $report $prdelim, @_;
424 my $bomb = sub { # usage: return $bomb->();
425 print $report " Unprocessable" if $report;
428 return (undef,undef);
430 die "commit $cur: Cannot cope with this commit";
436 my $build_start = sub {
437 my ($msg, $parent) = @_;
440 no warnings qw(exiting); last;
445 my $ty = $cl->{Type};
446 my $st = $cl->{SubType};
447 $prline->("$cl->{CommitId} $cl->{Type}");
448 $found{$ty. ( defined($st) ? "-$st" : '' )}++;
449 push @processed, $cl;
450 my $p0 = @{ $cl->{Parents} }==1 ? $cl->{Parents}[0]{CommitId} : undef;
451 if ($ty eq 'AddPatches') {
453 $rewrite_from_here->();
455 } elsif ($ty eq 'Packaging') {
459 } elsif ($ty eq 'BreakwaterStart') {
460 $build_start->('FirstPackaging', $cur);
461 } elsif ($ty eq 'Upstream') {
465 } elsif ($ty eq 'Mixed') {
468 my $cls = { $cl, $xmsg->("split mixed commit: $wh part") };
471 $queue->(\@brw_cl, "debian");
472 $queue->(\@upp_cl, "upstream");
473 $rewrite_from_here->();
476 } elsif ($ty eq 'Pseudomerge') {
477 print $report " Contributor=$ty->{Contributor}" if $report;
478 push @pseudomerges, $cl;
479 $rewrite_from_here->();
480 $cur = $ty->{Contributor};
482 } elsif ($ty eq 'BreakwaterUpstreamMerge') {
483 $build_start->("PreviousBreakwater", $cur);
484 } elsif ($ty eq 'DgitImportUnpatched') {
485 my $pm = $pseudomerges[-1];
487 # To an extent, this is heuristic. Imports don't have
488 # a useful history of the debian/ branch. We assume
489 # that the first pseudomerge after an import has a
490 # useful history of debian/, and ignore the histories
491 # from later pseudomerges. Often the first pseudomerge
492 # will be the dgit import of the upload to the actual
493 # suite intended by the non-dgit NMUer, and later
494 # pseudomerges may represent in-archive copies.
495 my $ovwrs = $pm->{Overwritten};
496 printf $report " PM=%s \@Overwr:%d", $pm, (scalar @$ovwrs)
501 my $ovwr = $ovwrs->[0]{CommitId};
502 printf $report " Overwr=%s", $ovwr if $report;
503 # This import has a tree which is just like a
504 # breakwater tree, but it has the wrong history. It
505 # ought to have the previous breakwater (which the
506 # pseudomerge overwrote) as an ancestor. That will
507 # make the history of the debian/ files correct. As
508 # for the upstream version: either it's the same as
509 # was ovewritten (ie, same as the previous
510 # breakwater), in which case that history is precisely
511 # right; or, otherwise, it was a non-gitish upload of a
512 # new upstream version. We can tell these apart by
513 # looking at the tree of the supposed upstream.
516 SpecialMethod => 'DgitImportDebianUpdate',
517 $xmsg->("convert dgit import: debian changes")
519 my $differs = (get_differs $ovwr, $cl->{Tree});
520 printf $report " Differs=%#x", $differs if $report;
521 if ($differs & D_UPS) {
522 printf $report " D_UPS" if $report;
523 # This will also trigger if a non-dgit git-based NMU
524 # deleted .gitignore (which is a thing that some of
525 # the existing git tools do if the user doesn't
526 # somehow tell them not to). Ah well.
529 SpecialMethod => 'DgitImportUpstreamUpdate',
530 $xmsg->("convert dgit import: upstream changes")
533 $prline->(" Import");
534 $rewrite_from_here->();
535 $upp_limit //= $#upp_cl; # further, deeper, patches discarded
539 # Everything is from this import. This kind of import
540 # is already in valid breakwater format, with the
541 # patches as commits.
542 printf $report " NoPM" if $report;
543 # last thing we processed will have been the first patch,
544 # if there is one; which is fine, so no need to rewrite
545 # on account of this import
546 $build_start->("ImportOrigin", $cur);
554 return if $nogenerate;
556 # Now we build it back up again
562 my $rm_tree_cached = sub {
564 runcmd @git, qw(rm --quiet -rf --cached), $subdir;
566 my $read_tree_debian = sub {
568 $rm_tree_cached->(qw(debian));
569 runcmd @git, qw(read-tree --prefix=debian/), "$treeish:debian";
571 my $read_tree_upstream = sub {
573 runcmd @git, qw(read-tree), $treeish;
574 $read_tree_debian->($build);
577 my $committer_authline = calculate_committer_authline();
580 mkdir $rd or $!==EEXIST or die $!;
582 foreach my $cl (qw(Debian), (reverse @brw_cl),
583 { SpecialMethod => 'RecordBreakwaterTip' },
584 qw(Upstream), (reverse @upp_cl)) {
586 $current_method = $cl;
589 my $method = $cl->{SpecialMethod} // $current_method;
590 my @parents = ($build);
591 my $cltree = $cl->{CommitId};
592 if ($method eq 'Debian') {
593 $read_tree_debian->($cltree);
594 } elsif ($method eq 'Upstream') {
595 $read_tree_upstream->($cltree);
596 } elsif ($method eq 'StartRewrite') {
599 } elsif ($method eq 'RecordBreakwaterTip') {
600 $breakwater = $build;
602 } elsif ($method eq 'DgitImportDebianUpdate') {
603 $read_tree_debian->($cltree);
604 $rm_tree_cached->(qw(debian/patches));
605 } elsif ($method eq 'DgitImportUpstreamUpdate') {
606 $read_tree_upstream->($cltree);
607 push @parents, map { $_->{CommitId} } @{ $cl->{OrigParents} };
611 $rewriting ||= $cl ne pop @processed;
612 my $newtree = cmdoutput @git, qw(write-tree);
614 $ch =~ s{^tree .*}{tree $newtree}m or confess "$ch ?";
615 $ch =~ s{^parent .*\n}{}m;
616 $ch =~ s{(?=^author)}{
617 map { "parent $_\n" } @parents
618 }me or confess "$ch ?";
620 $ch =~ s{^committer .*$}{$committer_authline}m
623 my $cf = "$rd/m$rewriting";
624 open CD, ">", $cf or die $!;
625 print CD $ch, "\n", $cl->{Msg} or die $!;
627 my @cmd = (@git, qw(hash-object));
628 push @cmd, qw(-w) if $rewriting;
629 push @cmd, qw(-t commit), $cf;
630 my $newcommit = cmdoutput @cmd;
631 confess "$ch ?" unless $rewriting or $newcommit eq $cl->{CommitId};
636 runcmd @git, qw(diff-tree --quiet), $input, $build;
638 return ($build, $breakwater);
641 sub get_head () { return git_rev_parse qw(HEAD); }
643 sub update_head ($$$) {
644 my ($old, $new, $mrest) = @_;
645 runcmd @git, qw(update-ref -m), "git-debrebase $mrest", $new, $old;
649 badusage "no arguments to launder allowed";
650 my $old = get_head();
651 my ($tip,$breakwater) = walk $old;
652 update_head $old, $tip, 'launder';
653 # no tree changes except debian/patches
654 runcmd @git, qw(rm --quiet -rf debian/patches);
655 printf "# breakwater tip\n%s\n", $breakwater;
659 die if ($ARGV[0]//'') =~ m/^-/;
660 badusage "too many arguments to analyse" if @ARGV>1;
663 $old = git_rev_parse $old;
667 my ($dummy,$breakwater) = walk $old, 1,*STDOUT;
668 print "$breakwater BREAKWATER\n";
669 STDOUT->error and die $!;
672 my $toplevel = cmdoutput @git, qw(rev-parse --show-toplevel);
673 chdir $toplevel or die "chdir $toplevel: $!";
675 my $cmd = shift @ARGV;
678 $cmdfn = ${*::}{"cmd_$cmdfn"};
680 $cmdfn or badusage "unknown git-debrebase sub-operation $cmd";