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 commit_pr_info ($) {
150 return Data::Dumper->dump([$r], [qw(commit)]);
153 sub calculate_committer_authline () {
154 my $c = cmdoutput @git, qw(commit-tree --no-gpg-sign -m),
155 'DUMMY COMMIT (git-debrebase)', "HEAD:";
156 my ($h,$m) = get_commit $c;
157 $h =~ m/^committer .*$/m or confess "($h) ?";
161 # classify returns an info hash like this
163 # Hdr => # commit headers, including 1 final newline
164 # Msg => # commit message (so one newline is dropped)
166 # Type => (see below)
168 # Ix => $index # ie 0, 1, 2, ...
170 # Differs => return value from get_differs
172 # IsDggitImport => 'orig' 'tarball' 'unpatched' 'package' (as from dgit)
174 # NewMsg => # commit message, but with any [dgit import ...] edited
175 # # to say "[was: ...]"
185 # has additional entres in classification result
186 # Overwritten = [ subset of Parents ]
187 # Contributor = $the_remaining_Parent
189 # DgitImportUnpatched
190 # has additional entry in classification result
191 # OrigParents = [ subset of Parents ]
193 # BreakwaterUpstreamMerge
194 # has additional entry in classification result
195 # OrigParents = [ subset of Parents ]
200 my ($h,$m) = get_commit $objid;
202 my ($t) = $h =~ m/^tree (\w+)$/m or die $objid;
203 my (@ph) = $h =~ m/^parent (\w+)$/m;
214 foreach my $ph (@ph) {
218 Differs => (get_differs $t, $ph),
223 my ($type, @rest) = @_;
224 $r = { %$r, Type => $type, @rest };
229 $r = { %$r, Type => qw(Unknown) };
234 my $d = $r->{Parents}[0]{Differs};
235 if ($d == D_PAT_ADD) {
236 return $classify->(qw(AddPatches));
237 } elsif ($d & (D_PAT_ADD|D_PAT_OTH)) {
238 return $unknown->("edits debian/patches");
239 } elsif ($d == D_DEB) {
240 my ($ty,$dummy) = git_cat_file "$ph[0]:debian";
242 return $classify->(qw(Packaging));
243 } elsif ($ty eq 'missing') {
244 return $classify->(qw(BreakwaterStart));
246 return $unknown->("parent's debian is not a directory");
248 } elsif ($d == D_UPS) {
249 return $classify->(qw(Upstream));
250 } elsif ($d == (D_DEB|D_UPS)) {
251 return $classify->(qw(Mixed));
253 return $unknown->("no changes");
255 confess "internal error $objid ?";
259 return $unknown->("origin commit");
262 my @identical = grep { !$_->{Differs} } @p;
263 if (@p == 2 && @identical == 1) {
264 my @overwritten = grep { $_->{Differs} } @p;
265 confess "internal error $objid ?" unless @overwritten==1;
266 return $classify->(qw(Pseudomerge),
267 Overwritten => $overwritten[0],
268 Contributor => $identical[0]);
270 if (@p == 2 && @identical == 2) {
271 my @bytime = nsort_by {
272 my ($ph,$pm) = get_commit $_->{CommitId};
273 $ph =~ m/^committer .* (\d+) [-+]\d+$/m or die "$_->{CommitId} ?";
276 return $classify->(qw(Pseudomerge),
277 SubType => qw(Ambiguous),
278 Overwritten => $bytime[0],
279 Contributor => $bytime[1]);
282 my ($p_h, $p_m) = get_commit $p;
283 $p->{IsOrigin} = $p_h !~ m/^parent \w+$/m;
284 ($p->{IsDgitImport},) = $p_m =~ m/^\[dgit import ([0-9a-z]+) .*\]$/m;
286 my @orig_ps = grep { ($_->{IsDgitImport}//'X') eq 'orig' } @p;
288 if (!(grep { !$_->{IsOrigin} } @p) and
289 (@orig_ps >= @p - 1) and
290 $m2 =~ s{^\[(dgit import unpatched .*)\]$}{[was: $1]}m) {
292 return $classify->(qw(DgitImportUnpatched),
293 OrigParents => \@orig_ps);
296 my ($stype, $series) = git_cat_file "$t:debian/patches/series";
297 my $haspatches = $stype ne 'missing' && $series =~ m/^\s*[^#\n\t ]/m;
299 # How to decide about l/r ordering of breakwater merges ? git
300 # --topo-order prefers to expand 2nd parent first. There's
301 # already an easy rune to look for debian/ history anyway (git log
302 # debian/) so debian breakwater branch should be 1st parent; that
303 # way also there's also an easy rune to look for the upstream
304 # patches (--topo-order).
307 !$p[0]{IsOrigin} && # breakwater merge never starts with an origin
308 !($p[0]{Differs} & ~D_DEB) &&
309 !($p[1]{Differs} & ~D_UPS)) {
310 return $classify->(qw(BreakwaterUpstreamMerge),
311 OrigParents => [ $p[1] ]);
313 # xxx multi-.orig upstreams
315 return $unknown->("complex merge");
321 $nogenerate,$report) = @_;
322 # => ($tip, $breakwater_tip)
323 # (or nothing, if $nogenerate)
325 # go through commits backwards
326 # we generate two lists of commits to apply:
327 # breakwater branch and upstream patches
328 my (@brw_cl, @upp_cl, @processed);
335 my ($appendinfo) = @_;
338 $ms .= "\n\n[git-debrebase $appendinfo]\n";
341 my $rewrite_from_here = sub {
342 my $sp_cl = { SpecialMethod => 'StartRewrite' };
343 push @brw_cl, $sp_cl;
344 push @processed, $sp_cl;
349 my $prprdelim = sub { print $report $prdelim if $report; $prdelim=""; };
352 return unless $report;
353 print $report $prdelim, @_;
357 my $bomb = sub { # usage: return $bomb->();
358 print $report " Unprocessable" if $report;
361 return (undef,undef);
363 die "commit $cur: Cannot cope with this commit";
369 my $build_start = sub {
370 my ($msg, $parent) = @_;
373 no warnings qw(exiting); last;
378 my $ty = $cl->{Type};
379 my $st = $cl->{SubType};
380 $prline->("$cl->{CommitId} $cl->{Type}");
381 $found{$ty. ( defined($st) ? "-$st" : '' )}++;
382 push @processed, $cl;
383 my $p0 = @{ $cl->{Parents} }==1 ? $cl->{Parents}[0]{CommitId} : undef;
384 if ($ty eq 'AddPatches') {
386 $rewrite_from_here->();
388 } elsif ($ty eq 'Packaging') {
392 } elsif ($ty eq 'BreakwaterStart') {
393 $build_start->('FirstPackaging', $cur);
394 } elsif ($ty eq 'Upstream') {
398 } elsif ($ty eq 'Mixed') {
401 my $cls = { $cl, $xmsg->("split mixed commit: $wh part") };
404 $queue->(\@brw_cl, "debian");
405 $queue->(\@upp_cl, "upstream");
406 $rewrite_from_here->();
409 } elsif ($ty eq 'Pseudomerge') {
410 print $report " Contributor=$ty->{Contributor}" if $report;
411 push @pseudomerges, $cl;
412 $rewrite_from_here->();
413 $cur = $ty->{Contributor};
415 } elsif ($ty eq 'BreakwaterUpstreamMerge') {
416 $build_start->("PreviousBreakwater", $cur);
417 } elsif ($ty eq 'DgitImportUnpatched') {
418 my $pm = $pseudomerges[-1];
420 # To an extent, this is heuristic. Imports don't have
421 # a useful history of the debian/ branch. We assume
422 # that the first pseudomerge after an import has a
423 # useful history of debian/, and ignore the histories
424 # from later pseudomerges. Often the first pseudomerge
425 # will be the dgit import of the upload to the actual
426 # suite intended by the non-dgit NMUer, and later
427 # pseudomerges may represent in-archive copies.
428 my $ovwrs = $pm->{Overwritten};
429 printf $report " PM=%s \@Overwr:%d", $pm, (scalar @$ovwrs)
434 my $ovwr = $ovwrs->[0]{CommitId};
435 printf $report " Overwr=%s", $ovwr if $report;
436 # This import has a tree which is just like a
437 # breakwater tree, but it has the wrong history. It
438 # ought to have the previous breakwater (which the
439 # pseudomerge overwrote) as an ancestor. That will
440 # make the history of the debian/ files correct. As
441 # for the upstream version: either it's the same as
442 # was ovewritten (ie, same as the previous
443 # breakwater), in which case that history is precisely
444 # right; or, otherwise, it was a non-gitish upload of a
445 # new upstream version. We can tell these apart by
446 # looking at the tree of the supposed upstream.
449 SpecialMethod => 'DgitImportDebianUpdate',
450 $xmsg->("convert dgit import: debian changes")
452 my $differs = (get_differs $ovwr, $cl->{Tree});
453 printf $report " Differs=%#x", $differs if $report;
454 if ($differs & D_UPS) {
455 printf $report " D_UPS" if $report;
456 # This will also trigger if a non-dgit git-based NMU
457 # deleted .gitignore (which is a thing that some of
458 # the existing git tools do if the user doesn't
459 # somehow tell them not to). Ah well.
462 SpecialMethod => 'DgitImportUpstreamUpdate',
463 $xmsg->("convert dgit import: upstream changes")
466 $prline->(" Import");
467 $rewrite_from_here->();
468 $upp_limit //= $#upp_cl; # further, deeper, patches discarded
472 # Everything is from this import. This kind of import
473 # is already in valid breakwater format, with the
474 # patches as commits.
475 printf $report " NoPM" if $report;
476 # last thing we processed will have been the first patch,
477 # if there is one; which is fine, so no need to rewrite
478 # on account of this import
479 $build_start->("ImportOrigin", $cur);
487 return if $nogenerate;
489 # Now we build it back up again
495 my $rm_tree_cached = sub {
497 runcmd @git, qw(rm --quiet -rf --cached), $subdir;
499 my $read_tree_debian = sub {
501 $rm_tree_cached->(qw(debian));
502 runcmd @git, qw(read-tree --prefix=debian/), "$treeish:debian";
504 my $read_tree_upstream = sub {
506 runcmd @git, qw(read-tree), $treeish;
507 $read_tree_debian->($build);
510 my $committer_authline = calculate_committer_authline();
513 mkdir $rd or $!==EEXIST or die $!;
515 foreach my $cl (qw(Debian), (reverse @brw_cl),
516 { SpecialMethod => 'RecordBreakwaterTip' },
517 qw(Upstream), (reverse @upp_cl)) {
519 $current_method = $cl;
522 my $method = $cl->{SpecialMethod} // $current_method;
523 my @parents = ($build);
524 my $cltree = $cl->{CommitId};
525 if ($method eq 'Debian') {
526 $read_tree_debian->($cltree);
527 } elsif ($method eq 'Upstream') {
528 $read_tree_upstream->($cltree);
529 } elsif ($method eq 'StartRewrite') {
532 } elsif ($method eq 'RecordBreakwaterTip') {
533 $breakwater = $build;
535 } elsif ($method eq 'DgitImportDebianUpdate') {
536 $read_tree_debian->($cltree);
537 $rm_tree_cached->(qw(debian/patches));
538 } elsif ($method eq 'DgitImportUpstreamUpdate') {
539 $read_tree_upstream->($cltree);
540 push @parents, map { $_->{CommitId} } @{ $cl->{OrigParents} };
544 $rewriting ||= $cl ne pop @processed;
545 my $newtree = cmdoutput @git, qw(write-tree);
547 $ch =~ s{^tree .*}{tree $newtree}m or confess "$ch ?";
548 $ch =~ s{^parent .*\n}{}m;
549 $ch =~ s{(?=^author)}{
550 map { "parent $_\n" } @parents
551 }me or confess "$ch ?";
553 $ch =~ s{^committer .*$}{$committer_authline}m
556 my $cf = "$rd/m$rewriting";
557 open CD, ">", $cf or die $!;
558 print CD $ch, "\n", $cl->{Msg} or die $!;
560 my @cmd = (@git, qw(hash-object));
561 push @cmd, qw(-w) if $rewriting;
562 push @cmd, qw(-t commit), $cf;
563 my $newcommit = cmdoutput @cmd;
564 confess "$ch ?" unless $rewriting or $newcommit eq $cl->{CommitId};
569 runcmd @git, qw(diff-tree --quiet), $input, $build;
571 return ($build, $breakwater);
574 sub get_head () { return git_rev_parse qw(HEAD); }
576 sub update_head ($$) {
577 my ($old, $new, $mrest) = @_;
578 runcmd @git, qw(update-ref -m), "git-debrebase $mrest", $new, $old;
582 badusage "no arguments to launder allowed";
583 my $old = get_head();
584 my ($tip,$breakwater) = walk $old;
585 update_head $old, $tip, 'launder';
586 # no tree changes except debian/patches
587 runcmd @git, qw(rm --quiet -rf debian/patches);
588 printf "# breakwater tip\n%s\n", $breakwater;
592 die if ($ARGV[0]//'') =~ m/^-/;
593 badusage "too many arguments to analyse" if @ARGV>1;
596 $old = git_rev_parse $old;
600 my ($dummy,$breakwater) = walk $old, 1,*STDOUT;
601 print "$breakwater BREAKWATER\n";
602 STDOUT->error and die $!;
605 my $toplevel = runcmd @git, qw(rev-parse --show-toplevel);
606 chdir $toplevel or die "chdir $toplevel: $!";
608 my $cmd = shift @ARGV;
611 $cmdfn = ${*::}{"cmd_$cmdfn"};
613 $cmdfn or badusage "unknown git-debrebase sub-operation $cmd";