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 Debian::Dgit qw(:DEFAULT $wa);
115 die "bad usage: $m\n";
121 my @cmd = qw(git config -z);
122 push @cmd, qw(--get-all) if wantarray;
124 my $out = cmdoutput @cmd;
125 return split /\0/, $out;
132 my ($type,$data) = git_cat_file $objid;
133 die unless $type eq 'commit';
134 $data =~ m/(?<=\n)\n/;
138 sub D_DEB () { return 0x1; } # debian/ (not including debian/patches/)
139 sub D_UPS () { return 0x2; } # upstream files
140 sub D_PAT_ADD () { return 0x4; } # debian/patches/ extra patches at end
141 sub D_PAT_OTH () { return 0x8; } # debian/patches other changes
143 our $rd = ".git/git-debrebase";
144 our $ud = "$rd/work";
147 sub commit_pr_info ($) {
149 return Data::Dumper->dump([$r], [qw(commit)]);
152 sub calculate_committer_authline () {
153 my $c = cmdoutput @git, qw(commit-tree --no-gpg-sign -m),
154 'DUMMY COMMIT (git-debrebase)', "HEAD:";
155 my ($h,$m) = get_commit $c;
156 $h =~ m/^committer .*$/m or confess "($h) ?";
160 # classify returns an info hash like this
162 # Hdr => # commit headers, including 1 final newline
163 # Msg => # commit message (so one newline is dropped)
165 # Type => (see below)
167 # Ix => $index # ie 0, 1, 2, ...
169 # Differs => return value from get_differs
171 # IsDggitImport => 'orig' 'tarball' 'unpatched' 'package' (as from dgit)
173 # NewMsg => # commit message, but with any [dgit import ...] edited
174 # # to say "[was: ...]"
184 # has additional entres in classification result
185 # Overwritten = [ subset of Parents ]
186 # Contributor = $the_remaining_Parent
188 # DgitImportUnpatched
189 # has additional entry in classification result
190 # OrigParents = [ subset of Parents ]
192 # BreakwaterUpstreamMerge
193 # has additional entry in classification result
194 # OrigParents = [ subset of Parents ]
199 my ($h,$m) = get_commit $objid;
201 my ($t) = $h =~ m/^tree (\w+)$/m or die $objid;
202 my (@ph) = $h =~ m/^parent (\w+)$/m;
213 foreach my $ph (@ph) {
217 Differs => (get_differs $t, $ph),
222 my ($type, @rest) = @_;
223 $r = { %$r, Type => $type, @rest };
228 $r = { %$r, Type => qw(Unknown) };
233 my $d = $r->{Parents}[0]{Differs};
234 if ($d == D_PAT_ADD) {
235 return $classify->(qw(AddPatches));
236 } elsif ($d & (D_PAT_ADD|D_PAT_OTH)) {
237 return $unknown->("edits debian/patches");
238 } elsif ($d == D_DEB) {
239 check if there were any deb before, is not this is
240 a packaging introduction (ie breakwater root)
241 return $classify->(qw(Packaging));
242 } elsif ($d == D_UPS) {
243 return $classify->(qw(Upstream));
244 } elsif ($d == (D_DEB|D_UPS)) {
245 return $classify->(qw(Mixed));
247 return $unknown->("no changes");
249 confess "internal error $objid ?";
253 return $unknown->("origin commit");
256 my @identical = grep { !$_->{Differs} } @p;
257 if (@p == 2 && @identical == 1) {
258 my @overwritten = grep { $_->{Differs} } @p;
259 confess "internal error $objid ?" unless @overwritten==1;
260 return $classify->(qw(Pseudomerge),
261 Overwritten => $overwritten[0],
262 Contributor => $identical[0]);
264 if (@p == 2 && @identical == 2) {
265 my @bytime = nsort_by {
266 my ($ph,$pm) = get_commit $_->{CommitId};
267 $ph =~ m/^committer .* (\d+) [-+]\d+$/m or die "$_->{CommitId} ?";
270 return $classify->(qw(Pseudomerge),
271 SubType => qw(Ambiguous),
272 Overwritten => $bytime[0],
273 Contributor => $bytime[1]);
276 my ($p_h, $p_m) = get_commit $p;
277 $p->{IsOrigin} = $p_h !~ m/^parent \w+$/m;
278 ($p->{IsDgitImport},) = $p_m =~ m/^\[dgit import ([0-9a-z]+) .*\]$/m;
280 my @orig_ps = grep { ($_->{IsDgitImport}//'X') eq 'orig' } @p;
282 if (!(grep { !$_->{IsOrigin} } @p) and
283 (@orig_ps >= @p - 1) and
284 $m2 =~ s{^\[(dgit import unpatched .*)\]$}{[was: $1]}m) {
286 return $classify->(qw(DgitImportUnpatched),
287 OrigParents => \@orig_ps);
290 my ($stype, $series) = git_cat_file "$t:debian/patches/series";
291 my $haspatches = $stype ne 'missing' && $series =~ m/^\s*[^#\n\t ]/m;
293 # How to decide about l/r ordering of breakwater merges ? git
294 # --topo-order prefers to expand 2nd parent first. There's
295 # already an easy rune to look for debian/ history anyway (git log
296 # debian/) so debian breakwater branch should be 1st parent; that
297 # way also there's also an easy rune to look for the upstream
298 # patches (--topo-order).
301 !$p[0]{IsOrigin} && # breakwater merge never starts with an origin
302 !($p[0]{Differs} & ~D_DEB) &&
303 !($p[1]{Differs} & ~D_UPS)) {
304 return $classify->(qw(BreakwaterUpstreamMerge),
305 OrigParents => [ $p[1] ]);
307 # xxx multi-.orig upstreams
309 return $unknown->("complex merge");
316 $wantdebonly,$depth) = @_;
317 # => ($tip, $breakwater_tip)
319 # go through commits backwards
320 # we generate two lists of commits to apply
321 my (@deb_cl, @ups_cl, @processed);
329 my ($appendinfo) = @_;
332 $ms .= "\n\n[git-debrebase $appendinfo]\n";
335 my $rewrite_from_here = sub {
336 push @processed, { SpecialMethod => 'StartRewrite' };
338 my $bomb = sub { # usage: return $bomb->();
339 print $report " Unprocessable" if $report;
342 return (undef,undef);
344 die "commit $cur: Cannot cope with this commit";
350 my $prprdelim = sub { print $report $prdelim if $report; $prdelim=""; };
353 return unless $report;
354 print $report $prdelim, @_;
360 push @deb_cl, { ExactlyParents => [] };
365 my $ty = $cl->{Type};
366 my $st = $cl->{SubType};
367 $prline->("$cl->{CommitId} $cl->{Type}");
368 $found{$ty. ( defined($st) ? "-$st" : '' )}++;
369 push @processed, $cl;
370 my $p0 = @[ $cl->{Parents} }==1 ? $cl->{Parents}[0]{CommitId} : undef;
371 if ($ty eq 'AddPatches') {
373 $rewrite_from_here->();
375 } elsif ($ty eq 'Packaging') {
379 } elsif ($ty eq 'Upstream') {
383 } elsif ($ty eq 'Mixed') {
386 my $cls = { $cl, $xmsg->("split mixed commit: $wh part") };
389 $queue->(\@deb_cl, "debian");
390 $queue->(\@ups_cl, "upstream");
391 $rewrite_from_here->();
393 } elsif ($ty eq 'Pseudomerge') {
394 print $report " Contributor=$ty->{Contributor}" if $report;
395 push @pseudomerges, $cl;
396 $rewrite_from_here->();
397 $cur = $ty->{Contributor};
399 } elsif ($ty eq 'BreakwaterUpstreamMerge') {
400 push @deb_cl, { ExactlyParents -> [$cur] };
401 $prline->("PreviousBreakwater");
403 } elsif ($ty eq 'DgitImportUnpatched') {
404 my $pm = $pseudomerges[-1];
406 # To an extent, this is heurstic. Imports don't have
407 # a useful history of the debian/ branch. We assume
408 # that the first pseudomerge after an import has a
409 # useful history or debian/, and ignore the histories
410 # from later pseudomerge. Often the first pseudomerge
411 # will be the dgit import of the upload to the actual
412 # suite intended by the non-dgit NMUer, and later
413 # pseudomerges may represent in-archive copies.
414 my $ovwrs = $pm->{Overwritten};
415 printf $report " PM=%s \@Overwr:%d", $pm, (scalar @$ovwrs)
420 my $ovwr = $ovwr->[0]{CommitId};
421 printf $report " Overwr=%s", $ovwr if $report;
422 # This import has a tree which is just like a
423 # breakwater tree, but it has the wrong history. It
424 # ought to have the previous breakwater (which the
425 # pseudomerge overwrote) as an ancestor. That will
426 # make the history of the debian/ files correct. As
427 # for the upstream version: either it's the same as
428 # was ovewritten (ie, same as the previous
429 # breakwater), in which case that history is precisely
430 # right; or, otherwise, it was a non-gitish upload of a
431 # new upstream version. We can tell these apart by
432 # looking at the tree of the supposed upstream.
435 SpecialMethod => 'DgitImportDebianUpdate',
436 $xmsg->("convert dgit import: debian changes")
438 my $differs = get_differs $ovwr, $cl->{Tree};
439 printf $report " Differs=%#x", $differs if $report;
440 if ($differs & D_UPS) {
441 printf $report " D_UPS" if $report;
442 # This will also trigger if a non-dgit git-based NMU
443 # deleted .gitignore (which is a thing that some of
444 # the existing git tools do if the user doesn't
445 # somehow tell them not to). Ah well.
448 SpecialMethod => 'DgitImportUpstreamUpdate',
449 $xmsg->("convert dgit import: upstream changes")
454 my ($basis,$dummy) = walk
456 $nogenerate, $report,
458 push @deb_cl, { ExactlyParents => [$basis] };
459 $rewrite_from_here->();
462 # Everything is from this import. This kind of import
463 # is already in valid breakwater format, with the
464 # patches as commits.
465 printf $report " NoPM" if $report;
466 push @deb_cl, { ExactlyParents => [$cur] };
467 # last thing we processed will have been the first patch,
468 # if there is one; which is fine, so no need to rewrite
469 # on account of this import
470 $prline->("ImportOrigin");
480 return (undef, $basis);
483 # Now we build it back up again
491 my $rm_tree_cached = sub {
493 runcmd @git, qw(rm --quiet -rf --cached), $subdir;
495 my $read_tree_debian = sub {
497 $rm_tree_cached->(qw(debian));
498 runcmd @git, qw(read-tree --prefix=debian/), "$treeish:debian";
500 my $read_tree_upstream = sub {
502 runcmd @git, qw(read-tree), $treeish;
503 $read_tree_debian->($build);
506 my $committer_authline = calculate_committer_authline();
509 mkdir $rd or $!==EEXIST or die $!;
511 foreach my $cl (qw(Debian), (reverse @deb_cl),
512 { SpecialMethod => 'RecordBreakwaterTip' },
513 qw(Upstream), (reverse @ups_cl)) {
515 $current_method = $cl;
518 my $method = $cl->{SpecialMethod} // $current_method;
519 my @parents = ($build);
520 my $cltree = $cl->{CommitId};
521 if ($method eq 'Debian') {
522 $read_tree_debian->($cltree);
523 } elsif ($method eq 'Upstream') {
524 $read_tree_upstream->($cltree);
525 } elsif ($method eq 'StartRewrite') {
528 } elsif ($method eq 'RecordBreakwaterTip') {
529 last if $wantdebonly;
530 $breakwater = $build;
532 } elsif ($method eq 'DgitImportDebianUpdate') {
533 $read_tree_debian->($cltree);
534 $rm_tree_cached->(qw(debian/patches));
535 } elsif ($method eq 'DgitImportUpstreamUpdate') {
536 $read_tree_upstream->($cltree);
537 push @parents, map { $_->{CommitId} } @{ $cl->{OrigParents} };
541 $rewriting ||= $cl ne pop @processed;
542 my $newtree = cmdoutput @git, qw(write-tree);
544 $ch =~ s{^tree .*}{tree $newtree}m or confess "$ch ?";
545 $ch =~ s{^parent .*\n}{}m;
546 $ch =~ s{(?=^author)}{
547 map { "parent $_\n" } @parents
548 }me or confess "$ch ?";
550 $ch =~ s{^committer .*$}{$committer_authline}m
553 my $cf = "$rd/m$rewriting";
554 open CD, ">", $cf or die $!;
555 print CD $ch, "\n", $cl->{Msg} or die $!;
557 my @cmd = (@git, qw(hash-object));
558 push @cmd, qw(-w) if $rewriting;
559 push @cmd, qw(-t commit), $cf;
560 my $newcommit = cmdoutput @cmd;
561 confess "$ch ?" unless $rewriting or $newcommit eq $cl->{CommitId};
566 runcmd @git, qw(diff-tree --quiet),
567 map { $wantdebonly ? "$_:debian" : $_ }
570 return ($build, $breakwater);
573 sub get_head () { return git_rev_parse qw(HEAD); }
575 sub update_head ($$) {
576 my ($old, $new, $mrest) = @_;
577 runcmd @git, qw(update-ref -m), "git-debrebase $mrest", $new, $old;
581 badusage "no arguments to launder allowed";
582 my $old = get_head();
583 my ($tip,$breakwater) = walk $old;
584 update_head $old, $tip, 'launder';
585 # no tree changes except debian/patches
586 runcmd @git, qw(rm --quiet -rf debian/patches);
587 printf "# breakwater tip\n%s\n", $breakwater;
591 die if ($ARGV[0]//'') =~ m/^-/;
592 badusage "too many arguments to analyse" if @ARGV>1;
595 $old = git_rev_parse $old;
599 my ($dummy,$breakwater) = walk $old, 1,*STDOUT;
600 print "$breakwater BREAKWATER\n";
601 STDOUT->error and die $!;
604 my $toplevel = runcmd @git, qw(rev-parse --show-toplevel);
605 chdir $toplevel or die "chdir $toplevel: $!";
607 my $cmd = shift @ARGV;
610 $cmdfn = ${*::}{"cmd_$cmdfn"};
612 $cmdfn or badusage "unknown git-debrebase sub-operation $cmd";