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: ...]"
273 # has additional entres in classification result
274 # Overwritten = [ subset of Parents ]
275 # Contributor = $the_remaining_Parent
277 # DgitImportUnpatched
278 # has additional entry in classification result
279 # OrigParents = [ subset of Parents ]
281 # BreakwaterUpstreamMerge
282 # has additional entry in classification result
283 # OrigParents = [ subset of Parents ]
288 my ($h,$m) = get_commit $objid;
290 my ($t) = $h =~ m/^tree (\w+)$/m or die $objid;
291 my (@ph) = $h =~ m/^parent (\w+)$/mg;
302 foreach my $ph (@ph) {
306 Differs => (get_differs $t, $ph),
310 printdebug "classify $objid \$t=$t \@p",
311 (map { sprintf " %s/%#x", $_->{CommitId}, $_->{Differs} } @p),
315 my ($type, @rest) = @_;
316 $r = { %$r, Type => $type, @rest };
318 my $dd = new Data::Dumper [ $r ];
319 Terse $dd 1; Indent $dd 0; Useqq $dd 1;
320 printdebug " = $type ".(Dump $dd)."\n";
326 $r = { %$r, Type => qw(Unknown) };
327 printdebug " ** Unknown\n";
332 my $d = $r->{Parents}[0]{Differs};
333 if ($d == D_PAT_ADD) {
334 return $classify->(qw(AddPatches));
335 } elsif ($d & (D_PAT_ADD|D_PAT_OTH)) {
336 return $unknown->("edits debian/patches");
337 } elsif ($d & DS_DEB and !($d & ~DS_DEB)) {
338 my ($ty,$dummy) = git_cat_file "$ph[0]:debian";
340 return $classify->(qw(Packaging));
341 } elsif ($ty eq 'missing') {
342 return $classify->(qw(BreakwaterStart));
344 return $unknown->("parent's debian is not a directory");
346 } elsif ($d == D_UPS) {
347 return $classify->(qw(Upstream));
348 } elsif ($d & DS_DEB and $d & D_UPS and !($d & ~(DS_DEB|D_UPS))) {
349 return $classify->(qw(Mixed));
351 return $unknown->("no changes");
353 confess "internal error $objid ?";
357 return $unknown->("origin commit");
360 my @identical = grep { !$_->{Differs} } @p;
361 if (@p == 2 && @identical == 1) {
362 my @overwritten = grep { $_->{Differs} } @p;
363 confess "internal error $objid ?" unless @overwritten==1;
364 return $classify->(qw(Pseudomerge),
365 Overwritten => $overwritten[0],
366 Contributor => $identical[0]);
368 if (@p == 2 && @identical == 2) {
369 my @bytime = nsort_by {
370 my ($ph,$pm) = get_commit $_->{CommitId};
371 $ph =~ m/^committer .* (\d+) [-+]\d+$/m or die "$_->{CommitId} ?";
374 return $classify->(qw(Pseudomerge),
375 SubType => qw(Ambiguous),
376 Overwritten => $bytime[0],
377 Contributor => $bytime[1]);
380 my ($p_h, $p_m) = get_commit $p->{CommitId};
381 $p->{IsOrigin} = $p_h !~ m/^parent \w+$/m;
382 ($p->{IsDgitImport},) = $p_m =~ m/^\[dgit import ([0-9a-z]+) .*\]$/m;
384 my @orig_ps = grep { ($_->{IsDgitImport}//'X') eq 'orig' } @p;
386 if (!(grep { !$_->{IsOrigin} } @p) and
387 (@orig_ps >= @p - 1) and
388 $m2 =~ s{^\[(dgit import unpatched .*)\]$}{[was: $1]}m) {
390 return $classify->(qw(DgitImportUnpatched),
391 OrigParents => \@orig_ps);
394 my ($stype, $series) = git_cat_file "$t:debian/patches/series";
395 my $haspatches = $stype ne 'missing' && $series =~ m/^\s*[^#\n\t ]/m;
397 # How to decide about l/r ordering of breakwater merges ? git
398 # --topo-order prefers to expand 2nd parent first. There's
399 # already an easy rune to look for debian/ history anyway (git log
400 # debian/) so debian breakwater branch should be 1st parent; that
401 # way also there's also an easy rune to look for the upstream
402 # patches (--topo-order).
404 # The above tells us which way *we* will generate them. But we
405 # might encounter ad-hoc breakwater merges generated manually,
406 # which might be the other way around. In principle, in some odd
407 # situations, a breakwater merge might have two identical parents.
408 # In that case we guess which way round it is (ie, which parent
409 # has the upstream history). The order of the 2-iteration loop
410 # controls which guess we make.
412 foreach my $prevbrw (qw(0 1)) {
415 !$p[$prevbrw]{IsOrigin} && # breakwater never starts with an origin
416 !($p[$prevbrw]{Differs} & ~DS_DEB) &&
417 !($p[!$prevbrw]{Differs} & ~D_UPS)) {
418 return $classify->(qw(BreakwaterUpstreamMerge),
419 OrigParents => [ $p[!$prevbrw] ]);
421 # xxx multi-.orig upstreams
424 return $unknown->("complex merge");
430 $nogenerate,$report) = @_;
431 # => ($tip, $breakwater_tip)
432 # (or nothing, if $nogenerate)
434 # go through commits backwards
435 # we generate two lists of commits to apply:
436 # breakwater branch and upstream patches
437 my (@brw_cl, @upp_cl, @processed);
444 my ($appendinfo) = @_;
447 $ms .= "\n\n[git-debrebase $appendinfo]\n";
450 my $rewrite_from_here = sub {
451 my $sp_cl = { SpecialMethod => 'StartRewrite' };
452 push @brw_cl, $sp_cl;
453 push @processed, $sp_cl;
458 my $prprdelim = sub { print $report $prdelim if $report; $prdelim=""; };
461 return unless $report;
462 print $report $prdelim, @_;
466 my $bomb = sub { # usage: return $bomb->();
467 print $report " Unprocessable" if $report;
470 return (undef,undef);
472 die "commit $cur: Cannot cope with this commit";
478 my $build_start = sub {
479 my ($msg, $parent) = @_;
482 no warnings qw(exiting); last;
487 my $ty = $cl->{Type};
488 my $st = $cl->{SubType};
489 $prline->("$cl->{CommitId} $cl->{Type}");
490 $found{$ty. ( defined($st) ? "-$st" : '' )}++;
491 push @processed, $cl;
492 my $p0 = @{ $cl->{Parents} }==1 ? $cl->{Parents}[0]{CommitId} : undef;
493 if ($ty eq 'AddPatches') {
495 $rewrite_from_here->();
497 } elsif ($ty eq 'Packaging') {
501 } elsif ($ty eq 'BreakwaterStart') {
502 $build_start->('FirstPackaging', $cur);
503 } elsif ($ty eq 'Upstream') {
507 } elsif ($ty eq 'Mixed') {
510 my $cls = { $cl, $xmsg->("split mixed commit: $wh part") };
513 $queue->(\@brw_cl, "debian");
514 $queue->(\@upp_cl, "upstream");
515 $rewrite_from_here->();
518 } elsif ($ty eq 'Pseudomerge') {
519 my $contrib = $cl->{Contributor}{CommitId};
520 print $report " Contributor=$contrib" if $report;
521 push @pseudomerges, $cl;
522 $rewrite_from_here->();
525 } elsif ($ty eq 'BreakwaterUpstreamMerge') {
526 $build_start->("PreviousBreakwater", $cur);
527 } elsif ($ty eq 'DgitImportUnpatched') {
528 my $pm = $pseudomerges[-1];
530 # To an extent, this is heuristic. Imports don't have
531 # a useful history of the debian/ branch. We assume
532 # that the first pseudomerge after an import has a
533 # useful history of debian/, and ignore the histories
534 # from later pseudomerges. Often the first pseudomerge
535 # will be the dgit import of the upload to the actual
536 # suite intended by the non-dgit NMUer, and later
537 # pseudomerges may represent in-archive copies.
538 my $ovwrs = $pm->{Overwritten};
539 printf $report " PM=%s \@Overwr:%d", $pm, (scalar @$ovwrs)
544 my $ovwr = $ovwrs->[0]{CommitId};
545 printf $report " Overwr=%s", $ovwr if $report;
546 # This import has a tree which is just like a
547 # breakwater tree, but it has the wrong history. It
548 # ought to have the previous breakwater (which the
549 # pseudomerge overwrote) as an ancestor. That will
550 # make the history of the debian/ files correct. As
551 # for the upstream version: either it's the same as
552 # was ovewritten (ie, same as the previous
553 # breakwater), in which case that history is precisely
554 # right; or, otherwise, it was a non-gitish upload of a
555 # new upstream version. We can tell these apart by
556 # looking at the tree of the supposed upstream.
559 SpecialMethod => 'DgitImportDebianUpdate',
560 $xmsg->("convert dgit import: debian changes")
562 my $differs = (get_differs $ovwr, $cl->{Tree});
563 printf $report " Differs=%#x", $differs if $report;
564 if ($differs & D_UPS) {
565 printf $report " D_UPS" if $report;
566 # This will also trigger if a non-dgit git-based NMU
567 # deleted .gitignore (which is a thing that some of
568 # the existing git tools do if the user doesn't
569 # somehow tell them not to). Ah well.
572 SpecialMethod => 'DgitImportUpstreamUpdate',
573 $xmsg->("convert dgit import: upstream changes")
576 $prline->(" Import");
577 $rewrite_from_here->();
578 $upp_limit //= $#upp_cl; # further, deeper, patches discarded
582 # Everything is from this import. This kind of import
583 # is already in valid breakwater format, with the
584 # patches as commits.
585 printf $report " NoPM" if $report;
586 # last thing we processed will have been the first patch,
587 # if there is one; which is fine, so no need to rewrite
588 # on account of this import
589 $build_start->("ImportOrigin", $cur);
597 return if $nogenerate;
599 # Now we build it back up again
605 my $rm_tree_cached = sub {
607 runcmd @git, qw(rm --quiet -rf --cached --ignore-unmatch), $subdir;
609 my $read_tree_debian = sub {
611 $rm_tree_cached->(qw(debian));
612 runcmd @git, qw(read-tree --prefix=debian/), "$treeish:debian";
614 my $read_tree_upstream = sub {
616 runcmd @git, qw(read-tree), $treeish;
617 $read_tree_debian->($build);
620 my $committer_authline = calculate_committer_authline();
623 mkdir $rd or $!==EEXIST or die $!;
625 foreach my $cl (qw(Debian), (reverse @brw_cl),
626 { SpecialMethod => 'RecordBreakwaterTip' },
627 qw(Upstream), (reverse @upp_cl)) {
629 $current_method = $cl;
632 my $method = $cl->{SpecialMethod} // $current_method;
633 my @parents = ($build);
634 my $cltree = $cl->{CommitId};
635 if ($method eq 'Debian') {
636 $read_tree_debian->($cltree);
637 } elsif ($method eq 'Upstream') {
638 $read_tree_upstream->($cltree);
639 } elsif ($method eq 'StartRewrite') {
642 } elsif ($method eq 'RecordBreakwaterTip') {
643 $breakwater = $build;
645 } elsif ($method eq 'DgitImportDebianUpdate') {
646 $read_tree_debian->($cltree);
647 $rm_tree_cached->(qw(debian/patches));
648 } elsif ($method eq 'DgitImportUpstreamUpdate') {
649 $read_tree_upstream->($cltree);
650 push @parents, map { $_->{CommitId} } @{ $cl->{OrigParents} };
654 $rewriting ||= $cl ne pop @processed;
655 my $newtree = cmdoutput @git, qw(write-tree);
657 $ch =~ s{^tree .*}{tree $newtree}m or confess "$ch ?";
658 $ch =~ s{^parent .*\n}{}m;
659 $ch =~ s{(?=^author)}{
660 map { "parent $_\n" } @parents
661 }me or confess "$ch ?";
663 $ch =~ s{^committer .*$}{$committer_authline}m
666 my $cf = "$rd/m$rewriting";
667 open CD, ">", $cf or die $!;
668 print CD $ch, "\n", $cl->{Msg} or die $!;
670 my @cmd = (@git, qw(hash-object));
671 push @cmd, qw(-w) if $rewriting;
672 push @cmd, qw(-t commit), $cf;
673 my $newcommit = cmdoutput @cmd;
674 confess "$ch ?" unless $rewriting or $newcommit eq $cl->{CommitId};
679 runcmd @git, qw(diff-tree --quiet), $input, $build;
681 return ($build, $breakwater);
684 sub get_head () { return git_rev_parse qw(HEAD); }
686 sub update_head ($$$) {
687 my ($old, $new, $mrest) = @_;
688 runcmd @git, qw(update-ref -m), "git-debrebase $mrest", $new, $old;
692 badusage "no arguments to launder allowed" if @ARGV;
693 my $old = get_head();
694 my ($tip,$breakwater) = walk $old;
695 update_head $old, $tip, 'launder';
696 # no tree changes except debian/patches
697 runcmd @git, qw(rm --quiet -rf debian/patches);
698 printf "# breakwater tip\n%s\n", $breakwater;
702 die if ($ARGV[0]//'') =~ m/^-/;
703 badusage "too many arguments to analyse" if @ARGV>1;
706 $old = git_rev_parse $old;
710 my ($dummy,$breakwater) = walk $old, 1,*STDOUT;
711 STDOUT->error and die $!;
714 GetOptions("D+" => \$debuglevel) or die badusage "bad options\n";
715 initdebug('git-debrebase ');
716 enabledebug if $debuglevel;
718 my $toplevel = cmdoutput @git, qw(rev-parse --show-toplevel);
719 chdir $toplevel or die "chdir $toplevel: $!";
721 $rd = fresh_playground "$playprefix/misc";
723 my $cmd = shift @ARGV;
726 $cmdfn = ${*::}{"cmd_$cmdfn"};
728 $cmdfn or badusage "unknown git-debrebase sub-operation $cmd";