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);
115 $wa = '.git/debrebase/work';
119 die "bad usage: $m\n";
125 my @cmd = qw(git config -z);
126 push @cmd, qw(--get-all) if wantarray;
128 my $out = cmdoutput @cmd;
129 return split /\0/, $out;
136 my $data = git_cat_file $objid, 'commit';
137 $data =~ m/(?<=\n)\n/ or die "$objid ($data) ?";
141 sub D_DEB () { return 0x1; } # debian/ (not including debian/patches/)
142 sub D_UPS () { return 0x2; } # upstream files
143 sub D_PAT_ADD () { return 0x4; } # debian/patches/ extra patches at end
144 sub D_PAT_OTH () { return 0x8; } # debian/patches other changes
146 our $rd = ".git/git-debrebase";
147 our $ud = "$rd/work";
150 sub get_differs ($$) {
152 # This resembles quiltify_trees_differ, in dgit, a bit.
153 # But we don't care about modes, or dpkg-source-unrepresentable
154 # changes, and we don't need the plethora of different modes.
155 # Conversely we need to distinguish different kinds of changes to
156 # debian/ and debian/patches/.
161 my ($opts, $limits, $fn) = @_;
162 my @cmd = (@git, qw(diff-tree -z --no-renames));
164 push @cmd, "$_:" foreach $x, $y;
166 my $diffs = cmdoutput @cmd;
167 foreach (split /\0/, $diffs) { $fn->(); }
170 $rundiff->([qw(--name-only)], [], sub {
171 $differs |= $_ eq 'debian' ? D_DEB : D_UPS;
174 if ($differs & D_DEB) {
176 $rundiff->([qw(--name-only -r)], [qw(debian)], sub {
177 $differs |= $_ eq m{^debian/patches/} ? D_PAT_OTH : D_DEB;
179 die "mysterious debian changes $x..$y"
180 unless $differs & (D_PAT_OTH|D_DEB);
183 if ($differs & D_PAT_OTH) {
185 $differs &= ~D_PAT_OTH;
187 $differs |= D_PAT_OTH;
188 no warnings qw(exiting); last;
190 $rundiff->([qw(--name-status -r)], [qw(debian/patches/)], sub {
191 no warnings qw(exiting);
192 if (!defined $mode) {
195 die unless s{^debian/patches/}{};
197 if ($mode eq 'A' && !m/(?:^|\.)series$/s) {
199 } elsif ($mode eq 'M' && $_ eq 'series') {
200 my $x_s = git_cat_file "$x:debian/patches/series", 'blob';
201 my $y_s = git_cat_file "$y:debian/patches/series", 'blob';
202 chomp $x_s; $x_s .= "\n";
203 $ok = $x_s eq substr($y_s, 0, length $x_s);
208 $differs |= $ok ? D_PAT_ADD : D_PAT_OTH;
210 die "mysterious debian/patches changes $x..$y"
211 unless $differs & (D_PAT_ADD|D_PAT_OTH);
214 printdebug sprintf "get_differs %s, %s = %#x\n", $x, $y, $differs;
219 sub commit_pr_info ($) {
221 return Data::Dumper->dump([$r], [qw(commit)]);
224 sub calculate_committer_authline () {
225 my $c = cmdoutput @git, qw(commit-tree --no-gpg-sign -m),
226 'DUMMY COMMIT (git-debrebase)', "HEAD:";
227 my ($h,$m) = get_commit $c;
228 $h =~ m/^committer .*$/m or confess "($h) ?";
232 # classify returns an info hash like this
234 # Hdr => # commit headers, including 1 final newline
235 # Msg => # commit message (so one newline is dropped)
237 # Type => (see below)
239 # Ix => $index # ie 0, 1, 2, ...
241 # Differs => return value from get_differs
243 # IsDggitImport => 'orig' 'tarball' 'unpatched' 'package' (as from dgit)
245 # NewMsg => # commit message, but with any [dgit import ...] edited
246 # # to say "[was: ...]"
256 # has additional entres in classification result
257 # Overwritten = [ subset of Parents ]
258 # Contributor = $the_remaining_Parent
260 # DgitImportUnpatched
261 # has additional entry in classification result
262 # OrigParents = [ subset of Parents ]
264 # BreakwaterUpstreamMerge
265 # has additional entry in classification result
266 # OrigParents = [ subset of Parents ]
271 my ($h,$m) = get_commit $objid;
273 my ($t) = $h =~ m/^tree (\w+)$/m or die $objid;
274 my (@ph) = $h =~ m/^parent (\w+)$/mg;
285 foreach my $ph (@ph) {
289 Differs => (get_differs $t, $ph),
293 printdebug "classify $objid \$t=$t \@p",
294 (map { sprintf " %s/%#x", $_->{CommitId}, $_->{Differs} } @p),
298 my ($type, @rest) = @_;
299 $r = { %$r, Type => $type, @rest };
301 my $dd = new Data::Dumper [ $r ];
302 Terse $dd 1; Indent $dd 0; Useqq $dd 1;
303 printdebug " = $type ".(Dump $dd)."\n";
309 $r = { %$r, Type => qw(Unknown) };
310 printdebug " ** Unknown\n";
315 my $d = $r->{Parents}[0]{Differs};
316 if ($d == D_PAT_ADD) {
317 return $classify->(qw(AddPatches));
318 } elsif ($d & (D_PAT_ADD|D_PAT_OTH)) {
319 return $unknown->("edits debian/patches");
320 } elsif ($d == D_DEB) {
321 my ($ty,$dummy) = git_cat_file "$ph[0]:debian";
323 return $classify->(qw(Packaging));
324 } elsif ($ty eq 'missing') {
325 return $classify->(qw(BreakwaterStart));
327 return $unknown->("parent's debian is not a directory");
329 } elsif ($d == D_UPS) {
330 return $classify->(qw(Upstream));
331 } elsif ($d == (D_DEB|D_UPS)) {
332 return $classify->(qw(Mixed));
334 return $unknown->("no changes");
336 confess "internal error $objid ?";
340 return $unknown->("origin commit");
343 my @identical = grep { !$_->{Differs} } @p;
344 if (@p == 2 && @identical == 1) {
345 my @overwritten = grep { $_->{Differs} } @p;
346 confess "internal error $objid ?" unless @overwritten==1;
347 return $classify->(qw(Pseudomerge),
348 Overwritten => $overwritten[0],
349 Contributor => $identical[0]);
351 if (@p == 2 && @identical == 2) {
352 my @bytime = nsort_by {
353 my ($ph,$pm) = get_commit $_->{CommitId};
354 $ph =~ m/^committer .* (\d+) [-+]\d+$/m or die "$_->{CommitId} ?";
357 return $classify->(qw(Pseudomerge),
358 SubType => qw(Ambiguous),
359 Overwritten => $bytime[0],
360 Contributor => $bytime[1]);
363 my ($p_h, $p_m) = get_commit $p->{CommitId};
364 $p->{IsOrigin} = $p_h !~ m/^parent \w+$/m;
365 ($p->{IsDgitImport},) = $p_m =~ m/^\[dgit import ([0-9a-z]+) .*\]$/m;
367 my @orig_ps = grep { ($_->{IsDgitImport}//'X') eq 'orig' } @p;
369 if (!(grep { !$_->{IsOrigin} } @p) and
370 (@orig_ps >= @p - 1) and
371 $m2 =~ s{^\[(dgit import unpatched .*)\]$}{[was: $1]}m) {
373 return $classify->(qw(DgitImportUnpatched),
374 OrigParents => \@orig_ps);
377 my ($stype, $series) = git_cat_file "$t:debian/patches/series";
378 my $haspatches = $stype ne 'missing' && $series =~ m/^\s*[^#\n\t ]/m;
380 @p = reverse @p; #xxx
382 # How to decide about l/r ordering of breakwater merges ? git
383 # --topo-order prefers to expand 2nd parent first. There's
384 # already an easy rune to look for debian/ history anyway (git log
385 # debian/) so debian breakwater branch should be 1st parent; that
386 # way also there's also an easy rune to look for the upstream
387 # patches (--topo-order).
390 !$p[0]{IsOrigin} && # breakwater merge never starts with an origin
391 !($p[0]{Differs} & ~D_DEB) &&
392 !($p[1]{Differs} & ~D_UPS)) {
393 return $classify->(qw(BreakwaterUpstreamMerge),
394 OrigParents => [ $p[1] ]);
396 # xxx multi-.orig upstreams
398 return $unknown->("complex merge");
404 $nogenerate,$report) = @_;
405 # => ($tip, $breakwater_tip)
406 # (or nothing, if $nogenerate)
408 # go through commits backwards
409 # we generate two lists of commits to apply:
410 # breakwater branch and upstream patches
411 my (@brw_cl, @upp_cl, @processed);
418 my ($appendinfo) = @_;
421 $ms .= "\n\n[git-debrebase $appendinfo]\n";
424 my $rewrite_from_here = sub {
425 my $sp_cl = { SpecialMethod => 'StartRewrite' };
426 push @brw_cl, $sp_cl;
427 push @processed, $sp_cl;
432 my $prprdelim = sub { print $report $prdelim if $report; $prdelim=""; };
435 return unless $report;
436 print $report $prdelim, @_;
440 my $bomb = sub { # usage: return $bomb->();
441 print $report " Unprocessable" if $report;
444 return (undef,undef);
446 die "commit $cur: Cannot cope with this commit";
452 my $build_start = sub {
453 my ($msg, $parent) = @_;
456 no warnings qw(exiting); last;
461 my $ty = $cl->{Type};
462 my $st = $cl->{SubType};
463 $prline->("$cl->{CommitId} $cl->{Type}");
464 $found{$ty. ( defined($st) ? "-$st" : '' )}++;
465 push @processed, $cl;
466 my $p0 = @{ $cl->{Parents} }==1 ? $cl->{Parents}[0]{CommitId} : undef;
467 if ($ty eq 'AddPatches') {
469 $rewrite_from_here->();
471 } elsif ($ty eq 'Packaging') {
475 } elsif ($ty eq 'BreakwaterStart') {
476 $build_start->('FirstPackaging', $cur);
477 } elsif ($ty eq 'Upstream') {
481 } elsif ($ty eq 'Mixed') {
484 my $cls = { $cl, $xmsg->("split mixed commit: $wh part") };
487 $queue->(\@brw_cl, "debian");
488 $queue->(\@upp_cl, "upstream");
489 $rewrite_from_here->();
492 } elsif ($ty eq 'Pseudomerge') {
493 my $contrib = $cl->{Contributor}{CommitId};
494 print $report " Contributor=$contrib" if $report;
495 push @pseudomerges, $cl;
496 $rewrite_from_here->();
499 } elsif ($ty eq 'BreakwaterUpstreamMerge') {
500 $build_start->("PreviousBreakwater", $cur);
501 } elsif ($ty eq 'DgitImportUnpatched') {
502 my $pm = $pseudomerges[-1];
504 # To an extent, this is heuristic. Imports don't have
505 # a useful history of the debian/ branch. We assume
506 # that the first pseudomerge after an import has a
507 # useful history of debian/, and ignore the histories
508 # from later pseudomerges. Often the first pseudomerge
509 # will be the dgit import of the upload to the actual
510 # suite intended by the non-dgit NMUer, and later
511 # pseudomerges may represent in-archive copies.
512 my $ovwrs = $pm->{Overwritten};
513 printf $report " PM=%s \@Overwr:%d", $pm, (scalar @$ovwrs)
518 my $ovwr = $ovwrs->[0]{CommitId};
519 printf $report " Overwr=%s", $ovwr if $report;
520 # This import has a tree which is just like a
521 # breakwater tree, but it has the wrong history. It
522 # ought to have the previous breakwater (which the
523 # pseudomerge overwrote) as an ancestor. That will
524 # make the history of the debian/ files correct. As
525 # for the upstream version: either it's the same as
526 # was ovewritten (ie, same as the previous
527 # breakwater), in which case that history is precisely
528 # right; or, otherwise, it was a non-gitish upload of a
529 # new upstream version. We can tell these apart by
530 # looking at the tree of the supposed upstream.
533 SpecialMethod => 'DgitImportDebianUpdate',
534 $xmsg->("convert dgit import: debian changes")
536 my $differs = (get_differs $ovwr, $cl->{Tree});
537 printf $report " Differs=%#x", $differs if $report;
538 if ($differs & D_UPS) {
539 printf $report " D_UPS" if $report;
540 # This will also trigger if a non-dgit git-based NMU
541 # deleted .gitignore (which is a thing that some of
542 # the existing git tools do if the user doesn't
543 # somehow tell them not to). Ah well.
546 SpecialMethod => 'DgitImportUpstreamUpdate',
547 $xmsg->("convert dgit import: upstream changes")
550 $prline->(" Import");
551 $rewrite_from_here->();
552 $upp_limit //= $#upp_cl; # further, deeper, patches discarded
556 # Everything is from this import. This kind of import
557 # is already in valid breakwater format, with the
558 # patches as commits.
559 printf $report " NoPM" if $report;
560 # last thing we processed will have been the first patch,
561 # if there is one; which is fine, so no need to rewrite
562 # on account of this import
563 $build_start->("ImportOrigin", $cur);
571 return if $nogenerate;
573 # Now we build it back up again
579 my $rm_tree_cached = sub {
581 runcmd @git, qw(rm --quiet -rf --cached --ignore-unmatch), $subdir;
583 my $read_tree_debian = sub {
585 $rm_tree_cached->(qw(debian));
586 runcmd @git, qw(read-tree --prefix=debian/), "$treeish:debian";
588 my $read_tree_upstream = sub {
590 runcmd @git, qw(read-tree), $treeish;
591 $read_tree_debian->($build);
594 my $committer_authline = calculate_committer_authline();
597 mkdir $rd or $!==EEXIST or die $!;
599 foreach my $cl (qw(Debian), (reverse @brw_cl),
600 { SpecialMethod => 'RecordBreakwaterTip' },
601 qw(Upstream), (reverse @upp_cl)) {
603 $current_method = $cl;
606 my $method = $cl->{SpecialMethod} // $current_method;
607 my @parents = ($build);
608 my $cltree = $cl->{CommitId};
609 if ($method eq 'Debian') {
610 $read_tree_debian->($cltree);
611 } elsif ($method eq 'Upstream') {
612 $read_tree_upstream->($cltree);
613 } elsif ($method eq 'StartRewrite') {
616 } elsif ($method eq 'RecordBreakwaterTip') {
617 $breakwater = $build;
619 } elsif ($method eq 'DgitImportDebianUpdate') {
620 $read_tree_debian->($cltree);
621 $rm_tree_cached->(qw(debian/patches));
622 } elsif ($method eq 'DgitImportUpstreamUpdate') {
623 $read_tree_upstream->($cltree);
624 push @parents, map { $_->{CommitId} } @{ $cl->{OrigParents} };
628 $rewriting ||= $cl ne pop @processed;
629 my $newtree = cmdoutput @git, qw(write-tree);
631 $ch =~ s{^tree .*}{tree $newtree}m or confess "$ch ?";
632 $ch =~ s{^parent .*\n}{}m;
633 $ch =~ s{(?=^author)}{
634 map { "parent $_\n" } @parents
635 }me or confess "$ch ?";
637 $ch =~ s{^committer .*$}{$committer_authline}m
640 my $cf = "$rd/m$rewriting";
641 open CD, ">", $cf or die $!;
642 print CD $ch, "\n", $cl->{Msg} or die $!;
644 my @cmd = (@git, qw(hash-object));
645 push @cmd, qw(-w) if $rewriting;
646 push @cmd, qw(-t commit), $cf;
647 my $newcommit = cmdoutput @cmd;
648 confess "$ch ?" unless $rewriting or $newcommit eq $cl->{CommitId};
653 runcmd @git, qw(diff-tree --quiet), $input, $build;
655 return ($build, $breakwater);
658 sub get_head () { return git_rev_parse qw(HEAD); }
660 sub update_head ($$$) {
661 my ($old, $new, $mrest) = @_;
662 runcmd @git, qw(update-ref -m), "git-debrebase $mrest", $new, $old;
666 badusage "no arguments to launder allowed" if @ARGV;
667 my $old = get_head();
668 my ($tip,$breakwater) = walk $old;
669 update_head $old, $tip, 'launder';
670 # no tree changes except debian/patches
671 runcmd @git, qw(rm --quiet -rf debian/patches);
672 printf "# breakwater tip\n%s\n", $breakwater;
676 die if ($ARGV[0]//'') =~ m/^-/;
677 badusage "too many arguments to analyse" if @ARGV>1;
680 $old = git_rev_parse $old;
684 my ($dummy,$breakwater) = walk $old, 1,*STDOUT;
685 STDOUT->error and die $!;
688 GetOptions("D+" => \$debuglevel) or die badusage "bad options\n";
689 initdebug('git-debrebase ');
690 enabledebug if $debuglevel;
692 my $toplevel = cmdoutput @git, qw(rev-parse --show-toplevel);
693 chdir $toplevel or die "chdir $toplevel: $!";
695 my $cmd = shift @ARGV;
698 $cmdfn = ${*::}{"cmd_$cmdfn"};
700 $cmdfn or badusage "unknown git-debrebase sub-operation $cmd";