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_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
145 our $playprefix = 'debrebase';
151 sub in_workarea ($) {
154 my $r = eval { $sub->(); };
158 sub fresh_workarea () {
159 $workarea = fresh_playground "$playprefix/work";
160 in_workarea sub { playtree_setup };
163 sub get_differs ($$) {
165 # This resembles quiltify_trees_differ, in dgit, a bit.
166 # But we don't care about modes, or dpkg-source-unrepresentable
167 # changes, and we don't need the plethora of different modes.
168 # Conversely we need to distinguish different kinds of changes to
169 # debian/ and debian/patches/.
174 my ($opts, $limits, $fn) = @_;
175 my @cmd = (@git, qw(diff-tree -z --no-renames));
177 push @cmd, "$_:" foreach $x, $y;
179 my $diffs = cmdoutput @cmd;
180 foreach (split /\0/, $diffs) { $fn->(); }
183 $rundiff->([qw(--name-only)], [], sub {
184 $differs |= $_ eq 'debian' ? D_DEB : D_UPS;
187 if ($differs & D_DEB) {
189 $rundiff->([qw(--name-only -r)], [qw(debian)], sub {
190 $differs |= $_ eq m{^debian/patches/} ? D_PAT_OTH : D_DEB;
192 die "mysterious debian changes $x..$y"
193 unless $differs & (D_PAT_OTH|D_DEB);
196 if ($differs & D_PAT_OTH) {
198 $differs &= ~D_PAT_OTH;
200 $differs |= D_PAT_OTH;
201 no warnings qw(exiting); last;
203 $rundiff->([qw(--name-status -r)], [qw(debian/patches/)], sub {
204 no warnings qw(exiting);
205 if (!defined $mode) {
208 die unless s{^debian/patches/}{};
210 if ($mode eq 'A' && !m/(?:^|\.)series$/s) {
212 } elsif ($mode eq 'M' && $_ eq 'series') {
213 my $x_s = git_cat_file "$x:debian/patches/series", 'blob';
214 my $y_s = git_cat_file "$y:debian/patches/series", 'blob';
215 chomp $x_s; $x_s .= "\n";
216 $ok = $x_s eq substr($y_s, 0, length $x_s);
221 $differs |= $ok ? D_PAT_ADD : D_PAT_OTH;
223 die "mysterious debian/patches changes $x..$y"
224 unless $differs & (D_PAT_ADD|D_PAT_OTH);
227 printdebug sprintf "get_differs %s, %s = %#x\n", $x, $y, $differs;
232 sub commit_pr_info ($) {
234 return Data::Dumper->dump([$r], [qw(commit)]);
237 sub calculate_committer_authline () {
238 my $c = cmdoutput @git, qw(commit-tree --no-gpg-sign -m),
239 'DUMMY COMMIT (git-debrebase)', "HEAD:";
240 my ($h,$m) = get_commit $c;
241 $h =~ m/^committer .*$/m or confess "($h) ?";
245 # classify returns an info hash like this
247 # Hdr => # commit headers, including 1 final newline
248 # Msg => # commit message (so one newline is dropped)
250 # Type => (see below)
252 # Ix => $index # ie 0, 1, 2, ...
254 # Differs => return value from get_differs
256 # IsDggitImport => 'orig' 'tarball' 'unpatched' 'package' (as from dgit)
258 # NewMsg => # commit message, but with any [dgit import ...] edited
259 # # to say "[was: ...]"
269 # has additional entres in classification result
270 # Overwritten = [ subset of Parents ]
271 # Contributor = $the_remaining_Parent
273 # DgitImportUnpatched
274 # has additional entry in classification result
275 # OrigParents = [ subset of Parents ]
277 # BreakwaterUpstreamMerge
278 # has additional entry in classification result
279 # OrigParents = [ subset of Parents ]
284 my ($h,$m) = get_commit $objid;
286 my ($t) = $h =~ m/^tree (\w+)$/m or die $objid;
287 my (@ph) = $h =~ m/^parent (\w+)$/mg;
298 foreach my $ph (@ph) {
302 Differs => (get_differs $t, $ph),
306 printdebug "classify $objid \$t=$t \@p",
307 (map { sprintf " %s/%#x", $_->{CommitId}, $_->{Differs} } @p),
311 my ($type, @rest) = @_;
312 $r = { %$r, Type => $type, @rest };
314 my $dd = new Data::Dumper [ $r ];
315 Terse $dd 1; Indent $dd 0; Useqq $dd 1;
316 printdebug " = $type ".(Dump $dd)."\n";
322 $r = { %$r, Type => qw(Unknown) };
323 printdebug " ** Unknown\n";
328 my $d = $r->{Parents}[0]{Differs};
329 if ($d == D_PAT_ADD) {
330 return $classify->(qw(AddPatches));
331 } elsif ($d & (D_PAT_ADD|D_PAT_OTH)) {
332 return $unknown->("edits debian/patches");
333 } elsif ($d == D_DEB) {
334 my ($ty,$dummy) = git_cat_file "$ph[0]:debian";
336 return $classify->(qw(Packaging));
337 } elsif ($ty eq 'missing') {
338 return $classify->(qw(BreakwaterStart));
340 return $unknown->("parent's debian is not a directory");
342 } elsif ($d == D_UPS) {
343 return $classify->(qw(Upstream));
344 } elsif ($d == (D_DEB|D_UPS)) {
345 return $classify->(qw(Mixed));
347 return $unknown->("no changes");
349 confess "internal error $objid ?";
353 return $unknown->("origin commit");
356 my @identical = grep { !$_->{Differs} } @p;
357 if (@p == 2 && @identical == 1) {
358 my @overwritten = grep { $_->{Differs} } @p;
359 confess "internal error $objid ?" unless @overwritten==1;
360 return $classify->(qw(Pseudomerge),
361 Overwritten => $overwritten[0],
362 Contributor => $identical[0]);
364 if (@p == 2 && @identical == 2) {
365 my @bytime = nsort_by {
366 my ($ph,$pm) = get_commit $_->{CommitId};
367 $ph =~ m/^committer .* (\d+) [-+]\d+$/m or die "$_->{CommitId} ?";
370 return $classify->(qw(Pseudomerge),
371 SubType => qw(Ambiguous),
372 Overwritten => $bytime[0],
373 Contributor => $bytime[1]);
376 my ($p_h, $p_m) = get_commit $p->{CommitId};
377 $p->{IsOrigin} = $p_h !~ m/^parent \w+$/m;
378 ($p->{IsDgitImport},) = $p_m =~ m/^\[dgit import ([0-9a-z]+) .*\]$/m;
380 my @orig_ps = grep { ($_->{IsDgitImport}//'X') eq 'orig' } @p;
382 if (!(grep { !$_->{IsOrigin} } @p) and
383 (@orig_ps >= @p - 1) and
384 $m2 =~ s{^\[(dgit import unpatched .*)\]$}{[was: $1]}m) {
386 return $classify->(qw(DgitImportUnpatched),
387 OrigParents => \@orig_ps);
390 my ($stype, $series) = git_cat_file "$t:debian/patches/series";
391 my $haspatches = $stype ne 'missing' && $series =~ m/^\s*[^#\n\t ]/m;
393 # How to decide about l/r ordering of breakwater merges ? git
394 # --topo-order prefers to expand 2nd parent first. There's
395 # already an easy rune to look for debian/ history anyway (git log
396 # debian/) so debian breakwater branch should be 1st parent; that
397 # way also there's also an easy rune to look for the upstream
398 # patches (--topo-order).
400 # The above tells us which way *we* will generate them. But we
401 # might encounter ad-hoc breakwater merges generated manually,
402 # which might be the other way around. In principle, in some odd
403 # situations, a breakwater merge might have two identical parents.
404 # In that case we guess which way round it is (ie, which parent
405 # has the upstream history). The order of the 2-iteration loop
406 # controls which guess we make.
408 foreach my $prevbrw (qw(0 1)) {
411 !$p[$prevbrw]{IsOrigin} && # breakwater never starts with an origin
412 !($p[$prevbrw]{Differs} & ~D_DEB) &&
413 !($p[!$prevbrw]{Differs} & ~D_UPS)) {
414 return $classify->(qw(BreakwaterUpstreamMerge),
415 OrigParents => [ $p[!$prevbrw] ]);
417 # xxx multi-.orig upstreams
420 return $unknown->("complex merge");
426 $nogenerate,$report) = @_;
427 # => ($tip, $breakwater_tip)
428 # (or nothing, if $nogenerate)
430 # go through commits backwards
431 # we generate two lists of commits to apply:
432 # breakwater branch and upstream patches
433 my (@brw_cl, @upp_cl, @processed);
440 my ($appendinfo) = @_;
443 $ms .= "\n\n[git-debrebase $appendinfo]\n";
446 my $rewrite_from_here = sub {
447 my $sp_cl = { SpecialMethod => 'StartRewrite' };
448 push @brw_cl, $sp_cl;
449 push @processed, $sp_cl;
454 my $prprdelim = sub { print $report $prdelim if $report; $prdelim=""; };
457 return unless $report;
458 print $report $prdelim, @_;
462 my $bomb = sub { # usage: return $bomb->();
463 print $report " Unprocessable" if $report;
466 return (undef,undef);
468 die "commit $cur: Cannot cope with this commit";
474 my $build_start = sub {
475 my ($msg, $parent) = @_;
478 no warnings qw(exiting); last;
483 my $ty = $cl->{Type};
484 my $st = $cl->{SubType};
485 $prline->("$cl->{CommitId} $cl->{Type}");
486 $found{$ty. ( defined($st) ? "-$st" : '' )}++;
487 push @processed, $cl;
488 my $p0 = @{ $cl->{Parents} }==1 ? $cl->{Parents}[0]{CommitId} : undef;
489 if ($ty eq 'AddPatches') {
491 $rewrite_from_here->();
493 } elsif ($ty eq 'Packaging') {
497 } elsif ($ty eq 'BreakwaterStart') {
498 $build_start->('FirstPackaging', $cur);
499 } elsif ($ty eq 'Upstream') {
503 } elsif ($ty eq 'Mixed') {
506 my $cls = { $cl, $xmsg->("split mixed commit: $wh part") };
509 $queue->(\@brw_cl, "debian");
510 $queue->(\@upp_cl, "upstream");
511 $rewrite_from_here->();
514 } elsif ($ty eq 'Pseudomerge') {
515 my $contrib = $cl->{Contributor}{CommitId};
516 print $report " Contributor=$contrib" if $report;
517 push @pseudomerges, $cl;
518 $rewrite_from_here->();
521 } elsif ($ty eq 'BreakwaterUpstreamMerge') {
522 $build_start->("PreviousBreakwater", $cur);
523 } elsif ($ty eq 'DgitImportUnpatched') {
524 my $pm = $pseudomerges[-1];
526 # To an extent, this is heuristic. Imports don't have
527 # a useful history of the debian/ branch. We assume
528 # that the first pseudomerge after an import has a
529 # useful history of debian/, and ignore the histories
530 # from later pseudomerges. Often the first pseudomerge
531 # will be the dgit import of the upload to the actual
532 # suite intended by the non-dgit NMUer, and later
533 # pseudomerges may represent in-archive copies.
534 my $ovwrs = $pm->{Overwritten};
535 printf $report " PM=%s \@Overwr:%d", $pm, (scalar @$ovwrs)
540 my $ovwr = $ovwrs->[0]{CommitId};
541 printf $report " Overwr=%s", $ovwr if $report;
542 # This import has a tree which is just like a
543 # breakwater tree, but it has the wrong history. It
544 # ought to have the previous breakwater (which the
545 # pseudomerge overwrote) as an ancestor. That will
546 # make the history of the debian/ files correct. As
547 # for the upstream version: either it's the same as
548 # was ovewritten (ie, same as the previous
549 # breakwater), in which case that history is precisely
550 # right; or, otherwise, it was a non-gitish upload of a
551 # new upstream version. We can tell these apart by
552 # looking at the tree of the supposed upstream.
555 SpecialMethod => 'DgitImportDebianUpdate',
556 $xmsg->("convert dgit import: debian changes")
558 my $differs = (get_differs $ovwr, $cl->{Tree});
559 printf $report " Differs=%#x", $differs if $report;
560 if ($differs & D_UPS) {
561 printf $report " D_UPS" if $report;
562 # This will also trigger if a non-dgit git-based NMU
563 # deleted .gitignore (which is a thing that some of
564 # the existing git tools do if the user doesn't
565 # somehow tell them not to). Ah well.
568 SpecialMethod => 'DgitImportUpstreamUpdate',
569 $xmsg->("convert dgit import: upstream changes")
572 $prline->(" Import");
573 $rewrite_from_here->();
574 $upp_limit //= $#upp_cl; # further, deeper, patches discarded
578 # Everything is from this import. This kind of import
579 # is already in valid breakwater format, with the
580 # patches as commits.
581 printf $report " NoPM" if $report;
582 # last thing we processed will have been the first patch,
583 # if there is one; which is fine, so no need to rewrite
584 # on account of this import
585 $build_start->("ImportOrigin", $cur);
593 return if $nogenerate;
595 # Now we build it back up again
601 my $rm_tree_cached = sub {
603 runcmd @git, qw(rm --quiet -rf --cached --ignore-unmatch), $subdir;
605 my $read_tree_debian = sub {
607 $rm_tree_cached->(qw(debian));
608 runcmd @git, qw(read-tree --prefix=debian/), "$treeish:debian";
610 my $read_tree_upstream = sub {
612 runcmd @git, qw(read-tree), $treeish;
613 $read_tree_debian->($build);
616 my $committer_authline = calculate_committer_authline();
619 mkdir $rd or $!==EEXIST or die $!;
621 foreach my $cl (qw(Debian), (reverse @brw_cl),
622 { SpecialMethod => 'RecordBreakwaterTip' },
623 qw(Upstream), (reverse @upp_cl)) {
625 $current_method = $cl;
628 my $method = $cl->{SpecialMethod} // $current_method;
629 my @parents = ($build);
630 my $cltree = $cl->{CommitId};
631 if ($method eq 'Debian') {
632 $read_tree_debian->($cltree);
633 } elsif ($method eq 'Upstream') {
634 $read_tree_upstream->($cltree);
635 } elsif ($method eq 'StartRewrite') {
638 } elsif ($method eq 'RecordBreakwaterTip') {
639 $breakwater = $build;
641 } elsif ($method eq 'DgitImportDebianUpdate') {
642 $read_tree_debian->($cltree);
643 $rm_tree_cached->(qw(debian/patches));
644 } elsif ($method eq 'DgitImportUpstreamUpdate') {
645 $read_tree_upstream->($cltree);
646 push @parents, map { $_->{CommitId} } @{ $cl->{OrigParents} };
650 $rewriting ||= $cl ne pop @processed;
651 my $newtree = cmdoutput @git, qw(write-tree);
653 $ch =~ s{^tree .*}{tree $newtree}m or confess "$ch ?";
654 $ch =~ s{^parent .*\n}{}m;
655 $ch =~ s{(?=^author)}{
656 map { "parent $_\n" } @parents
657 }me or confess "$ch ?";
659 $ch =~ s{^committer .*$}{$committer_authline}m
662 my $cf = "$rd/m$rewriting";
663 open CD, ">", $cf or die $!;
664 print CD $ch, "\n", $cl->{Msg} or die $!;
666 my @cmd = (@git, qw(hash-object));
667 push @cmd, qw(-w) if $rewriting;
668 push @cmd, qw(-t commit), $cf;
669 my $newcommit = cmdoutput @cmd;
670 confess "$ch ?" unless $rewriting or $newcommit eq $cl->{CommitId};
675 runcmd @git, qw(diff-tree --quiet), $input, $build;
677 return ($build, $breakwater);
680 sub get_head () { return git_rev_parse qw(HEAD); }
682 sub update_head ($$$) {
683 my ($old, $new, $mrest) = @_;
684 runcmd @git, qw(update-ref -m), "git-debrebase $mrest", $new, $old;
688 badusage "no arguments to launder allowed" if @ARGV;
689 my $old = get_head();
690 my ($tip,$breakwater) = walk $old;
691 update_head $old, $tip, 'launder';
692 # no tree changes except debian/patches
693 runcmd @git, qw(rm --quiet -rf debian/patches);
694 printf "# breakwater tip\n%s\n", $breakwater;
698 die if ($ARGV[0]//'') =~ m/^-/;
699 badusage "too many arguments to analyse" if @ARGV>1;
702 $old = git_rev_parse $old;
706 my ($dummy,$breakwater) = walk $old, 1,*STDOUT;
707 STDOUT->error and die $!;
710 GetOptions("D+" => \$debuglevel) or die badusage "bad options\n";
711 initdebug('git-debrebase ');
712 enabledebug if $debuglevel;
714 my $toplevel = cmdoutput @git, qw(rev-parse --show-toplevel);
715 chdir $toplevel or die "chdir $toplevel: $!";
717 $rd = fresh_playground "$playprefix/misc";
719 my $cmd = shift @ARGV;
722 $cmdfn = ${*::}{"cmd_$cmdfn"};
724 $cmdfn or badusage "unknown git-debrebase sub-operation $cmd";