git-debrebase: merge: Whitespace changes

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: Refactor interchange merge-base computation

We need to unpack this a bit so that we can insert some wrecknote
stuff, which will come in a moment.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: Fix a $nomerge message

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: Add a comment about rewriting

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: Remove a redundant check

MergedBreakwaters commits are now handled by a completely other branch
of the if, so this check is now redundant.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: mergedbreakwaters_anchor: Check consistency

Check that the result contains all the other anchors.  This is
ensured by walk, and required of VanillaMerges.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: Avoid needlessly walking MergedBreakwaters

We will not ever need to rewrite one of these, so we can just ask
keycommits for its anchor.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: test suite: Use t-ref-same-start

This is better than t-gdr-good laundered because (i) it is faster (ii)
it checks that the second laundry was a no-op (and did not rewrite
unnecessarily).

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: test suite: test avoiding re-merge of debian/

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: Avoid asking user to re-merge debian/

If it had conflicts, they will have been resolved earlier and we can
just take that resolution.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: test suite: Do a bit more testing

We can now call t-gdr-good laundered, because it works.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test suite: lib-gdr: t-gdr-good: Handle MergedBreakwaters

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: Add another xxx comment

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: walk: Reintroduce the local $read_tree_upstream sub

Now that we have the $tree_with_debian parameter to the global
read_tree_upstream, we can use it to avoid the git-write-tree call
and pass $build as the $tree_with_debian.

This thus partially reverts
  dfc092e0ec3d932e51a071000519477d0e33525c
  git-debrebase: Make read_tree_upstream global (nfc)

The combined effect is now that $read_tree_upstream does exactly the
git commands it did before.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: Get debian/ parts of new patch queue right

We want the newly generated delta queue commits to have the $newbase's
debian/.  Previously we let read_tree_upstream save and restore
debian/, but the contents of debian/ in the index are not right:
they're the result of the git merge, which is whatever is left over
from earlier.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: read_tree_upstream: New $tree_with_debian parameter

If not passed, this function does the same as it did before.

NFC.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: Generate some more debugging commits

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: keycommits: Handle merged-breakwater commits

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: keycommits: New $claimed_bw optional argument

No users yet, so NFC.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: Generate merged-breakwaters with correct metadata

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: keycommits: Break out $found_anchor

We are going to want to reuse this.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test suite: lib-gdr: t-gdr-good: Introduce t-gdr-analyse

NFC

This will allow us to recurse.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test suite: lib-gdr: t-gdr-good: Introduce $anal

NFC

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: Check that merge resolution is correct

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: Comment about laundry of merge of unstitched

I have concluded that this is not a problem avoidable by
git-debrebase, and that any trouble will be tolerable (at least, not
data loss).  The user should try to avoid doing this.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: Improve some xxx comments

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: test suite: Test both with and without `after'

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: Do not sometimes drop the first patch!

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Set $debugcmd_when_debuglevel = 2

This produces much better (less noisy) debug output.  git-debrebase
calls subcommands an awful lot and they are generally very
uninteresting.

If this now turns out to be insufficient, we will add additional calls
to printdebug.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Dgit.pm: printdebug: Selectively print $debugprefix, and add a doc

Remember whether the last printdebug ended in a \n.  If it didn't,
suppress printing the next newline.

I don't think anywhere calls printdebug without @_ right now.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Dgit.pm: printdebug: Strip empty strings from @_ and maybe quit early

Now the meat of the function always has nonempty $_[-1].

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Dgit.pm: printdebug: Break apart the print

NFC.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Dgit.pm: printdebug: Refactor slightly

This will make the next commits easier.  NFC.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Reverse argumennts to check diff error message

"Old, new" is more conventional.  This makes debug output and error
messages a bit easier to understand.

We must keep the backward order for get_differs because we need to
detect patch deletion, which shows up (reversed) as D_PAT_ADD.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Add a lot of printdebug

We also operated on sorted patch lists, and reorganise the prereq
computation slightly, to make this possible.

This `print DEBUG ... if' is not very nice.  This is needed because
printdebug unconditionally prints $debugprefix.  These new debug
prints are nearly the first which want to do multi-stage printing
outside dgit (and dgit normally operates without $debugprefix set).

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Dgit.pm: Introduce $debugcmd_when_debuglevel

This is also intended to be usefully localisable, and works by
localising $printdebug_when_debuglevel.

NFC.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Dgit.pm: Introduce $printdebug_when_debuglevel

This is intended to be usefully localisable.

NFC.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: Detect octopi later

If git-debrebase or dgit were to make an octopus merge, we should
recognise that as an unknown kind of gdr/dgit merge, not as a general
octopus merge.  This code motion leaves a spot where such a new kind
of merge could be classified.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: merge: Add an xxx

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: test suite: New gdr-merge test

This still has a lot of xxx's, but as it is right now it passes,
despite all the bugs.  This seems like a reasonable baseline for more
work.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: New merge handing, experimental and preliminary

Classification:
* Classify octopus merges as unknown, separately
* Classify 2-parent general merges as VanillaMerge, if feature enabled
* Introduce new merged-breakwater commit tag, classify as MergedBreakwaters
* Classify other gdr-generated 2-parent merges as unknown

Inspection (keycommits):
* VanillaMerge commits require laundry, much like dgit imports,
  and prevent discernment of the actual breakwater and anchor

Laundry (walk):
* Rewrite VanillaMerge using new merge_series function

Test suite:
* We get a slightly different error message in gdr-unprocessable, now.
  Later, when we make this non-experimental all the time, we are going
  to have to soup up gdr-unprocessable to make a worse test commit.

There are a lot of known bugs and infelicites, marked with "xxx".
There is a test (which will appear in a moment) but it's rather a poor
test, so there will be many uknown bugs.

So all of this is enabled only if GIT_DEBREBASE_EXPERIMENTAL_MERGE is
set in the environment.  (Strictly, only the classification is gated,
but that is sufficient.)  No significant functional change without
this setting.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: read_tree_upstream: New option $keep_patches

NFC with existing callers.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Improve ffq head recording message

Say `recorded _previous_ head' rather than `current head'.  Since by
the time the message appears, it's the previous head.  In particular,
`current' is very confusing if there were previous progress messages.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test suite: gitattributes: Change a comment from xxx to todo

This is not so critical.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

dgit: Fix a spelling error `$merginputs' to `$mergeinputs'

No functional change.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

dgit: Drop an obsolete xxx comment

This comment comes originally from a WIP commit
  e0ebbc5879c01d4c3e802101e146c5fb08bc9d49
I have in before-rebase.2016-11-07.  In that commit, @merginputs
did not exist.  Now it does, and that is how the ref comes out.
So this comment is obsolete.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Make read_tree_upstream global (nfc)

Save debian/ explicitly.  This is a bit annoying.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Make read_tree_debian global (nfc)

We can't easily do this to read_tree_upstream because that needs to
put debian/ back which is awkward.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test suite: lib-gdr: Break out t-gdr-prep-new-upstream

This does not include `git tag' for the upstream, because
gdr-newupstream wants to test the case where the tag is missing.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test suite: t-some-changes: Allow control of which files are edited

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Break out gbp_pq_export (nfc)

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: walk: Introduce $report_lprefix (nfc)

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

changelog: Start 6.7

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

changelog: finalise 6.6.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test suite: Fix gdr's calls to dgit when run formally in autopkgtest.

We should not pass `--dgit='.  That causes gdr to try to exec "".

Affects gdr-import-dgitview.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

changelog: Start 6.5

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

changelog: finalise 6.5

changelog: Categorise and reorder entries

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: scrap: Run git-rebase --abort if applicable

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: check for git-rebase in progress and abort most operations

scrap should abort, but doesn't yet.  status just prints a message.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test suite: gdr-subcommands: Drop obsolete line

We have done this another way and there will never be a push-hook like
this.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test suite: gdr-subcommands: Test scrap

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: scrap: Make it honour --noop-ok

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test suite: gdr-viagit: Test scrap

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: New feature `scrap'

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test suite: gdr-import-dgitview: Actually test convert-from-dgitview

This was always my intention with this test, of course.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Provide new convert-from-dgit-view operation.

The output is, unavoidably, not very pretty.

Closes:#905322.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Support --dgit option for finding dgit

Nothing uses this yet.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test suite: gdr-import-dgitview: New test for dgit dsc imports

So far this just tests the bugfixes I've just done.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Improve error message formatting from $bomb in walk

The : and ; were messed up.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Properly reject bare dgit dsc imports

A bare dgit dsc import is not a legal gdr branch, because the main
merge contains debian/patches/.  So it cannot be processed without
conversion.

Previously an attempt to launder such a branch would crash with a
stack trace, although keycommits would falsely claim it was OK

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Have resolve_upstream_commit set @$tried first

This allows a caller to know what tag name was successfully found,
by looking at $tried[-1];

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Break out resolve_upstream_version

NFC.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: resolve_upstream_version: Rename $upstream_version

This used to be called $new_upstream_version but this function is
called in other contexts too.

NFC.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase; Refactor resolve_upstream_version to take upstream version

NFC.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Use upstreamversion rather than open-coding

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Dgit.pm: Move upstreamversion from dgit

NFC.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Dgit.pm: Deconfuse argument orders of is_orig_file_of_p_v etc.

NFC.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

dgit: Improve error message for unknown suite, to suggest -d

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

Dgit.pm: Move is_orig_file_of_p_v and its dependencies from dgit

We want is_orig_file_of_p_v, so move it and all the things it depends
on.

NFC.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: factor out getoptions for subcommand

No functional change.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

dgit: Provide print-unapplied-treeish subcommand.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

dgit: Break out quilt_fakedsc2unapplied.

NFC.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

dgit: Rename --dgit-view-save to --save-dgit-view, leaving an alias.

This name makes more sense.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

dgit: Refactor %internal_object_save from $dgit_view_save

Preparatory to allow saving other internal objects.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test suite: editing a test script overrides DGIT_TESTS_PROGRESSIVE.

By making the script a dependency of the .ok stamp file.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test suite: Add t-tstunt-parsechangelog to many gdr tests.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

changelog: start 6.5

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

changelog: finalise 6.4

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase(1): Fix typo "unappled".

Closes:#905064.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

changelog: start 6.4

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

changelog: finalise 6.3

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Implement --help, providing a summary.

Provide a usage message.

I'm not convinced an option summary here is a good idea, mainly there
aren't any really important options.  (-f will be advertised when it
is relevant.)  So I have just summarised the most important operations.

Closes:#904990.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

dgit-maint-debrebase(7): Warn more against plain git rebase.

Specifically,
 - warn that pseudomerges must not be rebased over
 - warn and that git-rebase without a base should not be used
 - say to run gdr instead

Closes:#905004.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase(1): Warn against plain git rebase.

Re #905004.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase(5): Add ILLEGAL OPERATIONS section.

Re #905004.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

test suite; convert-from-gbp: test not identical in upstream files

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Improve error message for convert-from-gbp upstream discrepancy

Closes:#905078.

I don't intend to print `small' diffs because I think that kind of
dwim is probably more annoying than helpful.

I don't intend to factor this kind of pattern out throughout because:
The messages for these commands are generally not the actual diff rune
dgit or gdr ran, for various reasons (notably that dgit/gdr often use
object hashes).  The result is that there is little duplication.

So I think the situation now is OK.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: Improve error messages for bad options.

* GetOptions calls warn().  So we need a wrapper which disables
  or $SIG{__WARN__} (which prints a stack trace).
* Put the call to badusage in the wrapper.
* Change the messages to be clearer about what is meant.
* Add the program name to the badusage message.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>

git-debrebase: convert-from-gbp: Use same algorithm for finding upstream

The version is in the changelog, and thus upstream commitish can then
be found from as new-upstream.

This also means we now give the same error message, which is much
better than the one we had before.

Closes:#905062.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>