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 get_differs ($$) {
150 # This resembles quiltify_trees_differ, in dgit, a bit. But: we
151 # don't care about modes, or dpkg-source-unrepresentable changes,
152 # and we don't need the plethora of different modes.
158 my ($opts, $sfx, $fn) = @_;
159 my @cmd = (@git, qw(diff-tree -z --no-renames --name-only));
160 push @cmd "$_:$sfx" foreach $x, $y;
161 my $diffs = cmdoutput @bcmd, $x, $y;
162 foreach $f (split /\0/, $diffs) { $fn->(); }
165 $rundiff([], '', sub {
166 $differs |= $f eq 'debian' ? D_DEB : D_UPS;
169 if ($differs & D_DEB) {
171 $rundiff([], ':debian', sub {
172 $differs |= $f eq 'patches' ? D_PAT_OTH : D_DEB;
176 if ($differs & D_PAT_OTH) {
177 $rundiff([], ':debian/patches', sub {
186 sub commit_pr_info ($) {
188 return Data::Dumper->dump([$r], [qw(commit)]);
191 sub calculate_committer_authline () {
192 my $c = cmdoutput @git, qw(commit-tree --no-gpg-sign -m),
193 'DUMMY COMMIT (git-debrebase)', "HEAD:";
194 my ($h,$m) = get_commit $c;
195 $h =~ m/^committer .*$/m or confess "($h) ?";
199 # classify returns an info hash like this
201 # Hdr => # commit headers, including 1 final newline
202 # Msg => # commit message (so one newline is dropped)
204 # Type => (see below)
206 # Ix => $index # ie 0, 1, 2, ...
208 # Differs => return value from get_differs
210 # IsDggitImport => 'orig' 'tarball' 'unpatched' 'package' (as from dgit)
212 # NewMsg => # commit message, but with any [dgit import ...] edited
213 # # to say "[was: ...]"
223 # has additional entres in classification result
224 # Overwritten = [ subset of Parents ]
225 # Contributor = $the_remaining_Parent
227 # DgitImportUnpatched
228 # has additional entry in classification result
229 # OrigParents = [ subset of Parents ]
231 # BreakwaterUpstreamMerge
232 # has additional entry in classification result
233 # OrigParents = [ subset of Parents ]
238 my ($h,$m) = get_commit $objid;
240 my ($t) = $h =~ m/^tree (\w+)$/m or die $objid;
241 my (@ph) = $h =~ m/^parent (\w+)$/m;
252 foreach my $ph (@ph) {
256 Differs => (get_differs $t, $ph),
261 my ($type, @rest) = @_;
262 $r = { %$r, Type => $type, @rest };
267 $r = { %$r, Type => qw(Unknown) };
272 my $d = $r->{Parents}[0]{Differs};
273 if ($d == D_PAT_ADD) {
274 return $classify->(qw(AddPatches));
275 } elsif ($d & (D_PAT_ADD|D_PAT_OTH)) {
276 return $unknown->("edits debian/patches");
277 } elsif ($d == D_DEB) {
278 my ($ty,$dummy) = git_cat_file "$ph[0]:debian";
280 return $classify->(qw(Packaging));
281 } elsif ($ty eq 'missing') {
282 return $classify->(qw(BreakwaterStart));
284 return $unknown->("parent's debian is not a directory");
286 } elsif ($d == D_UPS) {
287 return $classify->(qw(Upstream));
288 } elsif ($d == (D_DEB|D_UPS)) {
289 return $classify->(qw(Mixed));
291 return $unknown->("no changes");
293 confess "internal error $objid ?";
297 return $unknown->("origin commit");
300 my @identical = grep { !$_->{Differs} } @p;
301 if (@p == 2 && @identical == 1) {
302 my @overwritten = grep { $_->{Differs} } @p;
303 confess "internal error $objid ?" unless @overwritten==1;
304 return $classify->(qw(Pseudomerge),
305 Overwritten => $overwritten[0],
306 Contributor => $identical[0]);
308 if (@p == 2 && @identical == 2) {
309 my @bytime = nsort_by {
310 my ($ph,$pm) = get_commit $_->{CommitId};
311 $ph =~ m/^committer .* (\d+) [-+]\d+$/m or die "$_->{CommitId} ?";
314 return $classify->(qw(Pseudomerge),
315 SubType => qw(Ambiguous),
316 Overwritten => $bytime[0],
317 Contributor => $bytime[1]);
320 my ($p_h, $p_m) = get_commit $p;
321 $p->{IsOrigin} = $p_h !~ m/^parent \w+$/m;
322 ($p->{IsDgitImport},) = $p_m =~ m/^\[dgit import ([0-9a-z]+) .*\]$/m;
324 my @orig_ps = grep { ($_->{IsDgitImport}//'X') eq 'orig' } @p;
326 if (!(grep { !$_->{IsOrigin} } @p) and
327 (@orig_ps >= @p - 1) and
328 $m2 =~ s{^\[(dgit import unpatched .*)\]$}{[was: $1]}m) {
330 return $classify->(qw(DgitImportUnpatched),
331 OrigParents => \@orig_ps);
334 my ($stype, $series) = git_cat_file "$t:debian/patches/series";
335 my $haspatches = $stype ne 'missing' && $series =~ m/^\s*[^#\n\t ]/m;
337 # How to decide about l/r ordering of breakwater merges ? git
338 # --topo-order prefers to expand 2nd parent first. There's
339 # already an easy rune to look for debian/ history anyway (git log
340 # debian/) so debian breakwater branch should be 1st parent; that
341 # way also there's also an easy rune to look for the upstream
342 # patches (--topo-order).
345 !$p[0]{IsOrigin} && # breakwater merge never starts with an origin
346 !($p[0]{Differs} & ~D_DEB) &&
347 !($p[1]{Differs} & ~D_UPS)) {
348 return $classify->(qw(BreakwaterUpstreamMerge),
349 OrigParents => [ $p[1] ]);
351 # xxx multi-.orig upstreams
353 return $unknown->("complex merge");
359 $nogenerate,$report) = @_;
360 # => ($tip, $breakwater_tip)
361 # (or nothing, if $nogenerate)
363 # go through commits backwards
364 # we generate two lists of commits to apply:
365 # breakwater branch and upstream patches
366 my (@brw_cl, @upp_cl, @processed);
373 my ($appendinfo) = @_;
376 $ms .= "\n\n[git-debrebase $appendinfo]\n";
379 my $rewrite_from_here = sub {
380 my $sp_cl = { SpecialMethod => 'StartRewrite' };
381 push @brw_cl, $sp_cl;
382 push @processed, $sp_cl;
387 my $prprdelim = sub { print $report $prdelim if $report; $prdelim=""; };
390 return unless $report;
391 print $report $prdelim, @_;
395 my $bomb = sub { # usage: return $bomb->();
396 print $report " Unprocessable" if $report;
399 return (undef,undef);
401 die "commit $cur: Cannot cope with this commit";
407 my $build_start = sub {
408 my ($msg, $parent) = @_;
411 no warnings qw(exiting); last;
416 my $ty = $cl->{Type};
417 my $st = $cl->{SubType};
418 $prline->("$cl->{CommitId} $cl->{Type}");
419 $found{$ty. ( defined($st) ? "-$st" : '' )}++;
420 push @processed, $cl;
421 my $p0 = @{ $cl->{Parents} }==1 ? $cl->{Parents}[0]{CommitId} : undef;
422 if ($ty eq 'AddPatches') {
424 $rewrite_from_here->();
426 } elsif ($ty eq 'Packaging') {
430 } elsif ($ty eq 'BreakwaterStart') {
431 $build_start->('FirstPackaging', $cur);
432 } elsif ($ty eq 'Upstream') {
436 } elsif ($ty eq 'Mixed') {
439 my $cls = { $cl, $xmsg->("split mixed commit: $wh part") };
442 $queue->(\@brw_cl, "debian");
443 $queue->(\@upp_cl, "upstream");
444 $rewrite_from_here->();
447 } elsif ($ty eq 'Pseudomerge') {
448 print $report " Contributor=$ty->{Contributor}" if $report;
449 push @pseudomerges, $cl;
450 $rewrite_from_here->();
451 $cur = $ty->{Contributor};
453 } elsif ($ty eq 'BreakwaterUpstreamMerge') {
454 $build_start->("PreviousBreakwater", $cur);
455 } elsif ($ty eq 'DgitImportUnpatched') {
456 my $pm = $pseudomerges[-1];
458 # To an extent, this is heuristic. Imports don't have
459 # a useful history of the debian/ branch. We assume
460 # that the first pseudomerge after an import has a
461 # useful history of debian/, and ignore the histories
462 # from later pseudomerges. Often the first pseudomerge
463 # will be the dgit import of the upload to the actual
464 # suite intended by the non-dgit NMUer, and later
465 # pseudomerges may represent in-archive copies.
466 my $ovwrs = $pm->{Overwritten};
467 printf $report " PM=%s \@Overwr:%d", $pm, (scalar @$ovwrs)
472 my $ovwr = $ovwrs->[0]{CommitId};
473 printf $report " Overwr=%s", $ovwr if $report;
474 # This import has a tree which is just like a
475 # breakwater tree, but it has the wrong history. It
476 # ought to have the previous breakwater (which the
477 # pseudomerge overwrote) as an ancestor. That will
478 # make the history of the debian/ files correct. As
479 # for the upstream version: either it's the same as
480 # was ovewritten (ie, same as the previous
481 # breakwater), in which case that history is precisely
482 # right; or, otherwise, it was a non-gitish upload of a
483 # new upstream version. We can tell these apart by
484 # looking at the tree of the supposed upstream.
487 SpecialMethod => 'DgitImportDebianUpdate',
488 $xmsg->("convert dgit import: debian changes")
490 my $differs = (get_differs $ovwr, $cl->{Tree});
491 printf $report " Differs=%#x", $differs if $report;
492 if ($differs & D_UPS) {
493 printf $report " D_UPS" if $report;
494 # This will also trigger if a non-dgit git-based NMU
495 # deleted .gitignore (which is a thing that some of
496 # the existing git tools do if the user doesn't
497 # somehow tell them not to). Ah well.
500 SpecialMethod => 'DgitImportUpstreamUpdate',
501 $xmsg->("convert dgit import: upstream changes")
504 $prline->(" Import");
505 $rewrite_from_here->();
506 $upp_limit //= $#upp_cl; # further, deeper, patches discarded
510 # Everything is from this import. This kind of import
511 # is already in valid breakwater format, with the
512 # patches as commits.
513 printf $report " NoPM" if $report;
514 # last thing we processed will have been the first patch,
515 # if there is one; which is fine, so no need to rewrite
516 # on account of this import
517 $build_start->("ImportOrigin", $cur);
525 return if $nogenerate;
527 # Now we build it back up again
533 my $rm_tree_cached = sub {
535 runcmd @git, qw(rm --quiet -rf --cached), $subdir;
537 my $read_tree_debian = sub {
539 $rm_tree_cached->(qw(debian));
540 runcmd @git, qw(read-tree --prefix=debian/), "$treeish:debian";
542 my $read_tree_upstream = sub {
544 runcmd @git, qw(read-tree), $treeish;
545 $read_tree_debian->($build);
548 my $committer_authline = calculate_committer_authline();
551 mkdir $rd or $!==EEXIST or die $!;
553 foreach my $cl (qw(Debian), (reverse @brw_cl),
554 { SpecialMethod => 'RecordBreakwaterTip' },
555 qw(Upstream), (reverse @upp_cl)) {
557 $current_method = $cl;
560 my $method = $cl->{SpecialMethod} // $current_method;
561 my @parents = ($build);
562 my $cltree = $cl->{CommitId};
563 if ($method eq 'Debian') {
564 $read_tree_debian->($cltree);
565 } elsif ($method eq 'Upstream') {
566 $read_tree_upstream->($cltree);
567 } elsif ($method eq 'StartRewrite') {
570 } elsif ($method eq 'RecordBreakwaterTip') {
571 $breakwater = $build;
573 } elsif ($method eq 'DgitImportDebianUpdate') {
574 $read_tree_debian->($cltree);
575 $rm_tree_cached->(qw(debian/patches));
576 } elsif ($method eq 'DgitImportUpstreamUpdate') {
577 $read_tree_upstream->($cltree);
578 push @parents, map { $_->{CommitId} } @{ $cl->{OrigParents} };
582 $rewriting ||= $cl ne pop @processed;
583 my $newtree = cmdoutput @git, qw(write-tree);
585 $ch =~ s{^tree .*}{tree $newtree}m or confess "$ch ?";
586 $ch =~ s{^parent .*\n}{}m;
587 $ch =~ s{(?=^author)}{
588 map { "parent $_\n" } @parents
589 }me or confess "$ch ?";
591 $ch =~ s{^committer .*$}{$committer_authline}m
594 my $cf = "$rd/m$rewriting";
595 open CD, ">", $cf or die $!;
596 print CD $ch, "\n", $cl->{Msg} or die $!;
598 my @cmd = (@git, qw(hash-object));
599 push @cmd, qw(-w) if $rewriting;
600 push @cmd, qw(-t commit), $cf;
601 my $newcommit = cmdoutput @cmd;
602 confess "$ch ?" unless $rewriting or $newcommit eq $cl->{CommitId};
607 runcmd @git, qw(diff-tree --quiet), $input, $build;
609 return ($build, $breakwater);
612 sub get_head () { return git_rev_parse qw(HEAD); }
614 sub update_head ($$$) {
615 my ($old, $new, $mrest) = @_;
616 runcmd @git, qw(update-ref -m), "git-debrebase $mrest", $new, $old;
620 badusage "no arguments to launder allowed";
621 my $old = get_head();
622 my ($tip,$breakwater) = walk $old;
623 update_head $old, $tip, 'launder';
624 # no tree changes except debian/patches
625 runcmd @git, qw(rm --quiet -rf debian/patches);
626 printf "# breakwater tip\n%s\n", $breakwater;
630 die if ($ARGV[0]//'') =~ m/^-/;
631 badusage "too many arguments to analyse" if @ARGV>1;
634 $old = git_rev_parse $old;
638 my ($dummy,$breakwater) = walk $old, 1,*STDOUT;
639 print "$breakwater BREAKWATER\n";
640 STDOUT->error and die $!;
643 my $toplevel = cmdoutput @git, qw(rev-parse --show-toplevel);
644 chdir $toplevel or die "chdir $toplevel: $!";
646 my $cmd = shift @ARGV;
649 $cmdfn = ${*::}{"cmd_$cmdfn"};
651 $cmdfn or badusage "unknown git-debrebase sub-operation $cmd";