dgit: Option parsing: combine parseopts_late_defaults and finalise_opts_opts

No functional change.

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

dgit: Option parsing: Move parseopts_late_defaults textually

No functional change.

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

dgit: Option parsing: Fix distro/suite reference bugs

By moving parseopts_late_defaults call.

This is conceptually the same as finalise_late_opts, so do it there.
We have to add a missing call to build_prep_early, as otherwise we
call massage_dbp_args too soon.

Also move a nopushing() call until after $isuite is set

All of this means that we now honour $isuite (and sometimes $idistro)
much better.

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

dgit: Option parsing: Make function parseopts_late_defaults

No functional change.

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

dgit: Option parsing: Provide call to build_prep_early in build_source

This means $isuite is set a bit earlier, etc.

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

dgit: Option parsing: quilt-fixup: Use build_prep_early

This means $isuite is set early enough.

As a side-effect, we correctly reject -p.

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

test suite: Allow overriding of -dtest-dummy dgit argument

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

test suite: lib-reprepro: Allow configuring for another distro

We can now make a second `distro' this way, with some additional
optional arguments.

No functional change for existing call sites.

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

test suite: lib-reprepro: Add a couple of comments

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

test suite: Refactor lib-reprepro

* Break out t-reprepro-cfg.
* Introduce the variables `mir' and `etcapt' to many functions.

No functional change.

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

test suite: Add t-stunt-parsechangelog to a few tests which were missing it.

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

test suite: Parse Dgit: field properly (look for initial \w+ only)

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

dgit: import-dsc: Parse `Dgit:' field properly, using only \w+

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

test suite: gbp-orig: Add a missing -m

Without this, it would run an editor if stdout was a tty (!)

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

dgit-badcommit-fixup: Do not investigate symrefs

Closes:#850547.

Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>
dgit-badcommit-fixup:
* Do not investigate symrefs.  Closes:#850547.

test suite: lib-orig-include-exclude Strip block count out of find -ls

This is unstable!  I have just seen it vary spontaneously in the
middle of a test run, resulting in this:

 ++ diff -u ../before ../after
 --- ../before   2017-01-08 21:39:48.176033763 +0000
 +++ ../after    2017-01-08 21:39:48.208033833 +0000
 @@ -1,2 +1,2 @@
 -   208380      4 -rw-rw-r--   9 ian      ian           236 Nov  6 20:49 /home/ian/things/Dgit/2dgit/tests/tmp/orig-include-exclude-chkquery/mirror/pool/main/example_1.0.orig-docs.tar.gz
 -   208381      4 -rw-rw-r--   8 ian      ian           373 Nov  6 20:49 /home/ian/things/Dgit/2dgit/tests/tmp/orig-include-exclude-chkquery/mirror/pool/main/example_1.0.orig.tar.gz
 +   208380      4 -rw-rw-r--  10 ian      ian           236 Nov  6 20:49 /home/ian/things/Dgit/2dgit/tests/tmp/orig-include-exclude-chkquery/mirror/pool/main/example_1.0.orig-docs.tar.gz
 +   208381      4 -rw-rw-r--   9 ian      ian           373 Nov  6 20:49 /home/ian/things/Dgit/2dgit/tests/tmp/orig-include-exclude-chkquery/mirror/pool/main/example_1.0.orig.tar.gz
 + test 1 = 0
 + t-report-failure

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

test suite: lib-orig-include-exclude: Introduce origs_findls

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

Arrange to pass --debug-quick-random to gpg-agent.

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

badcommit-rewrite: Fix operation using installed version of fixup.

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

test suite: Generalise DGIT_TEST_REAL_<FOO> handling

No overall functional change.

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

test suite: tests/run-all: Use a bit more cpu

Make allowance for tests that do a bit of downloading or whatever.

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

changelog: start 2.17~ (again)

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

changelog: finalise 2.16.2

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

dgit-badcommit-fixup; In --check mode, exit status 2 if things are not fine

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

dgit-badcommit-fixup: Fix crash when running for 2nd time in bare repo.

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

changelog: start 2.17~ (again)

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

changelog: finalise 2.16.1

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

dgit-badcommit-fixup: New mode --check which is readonly.

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

changelog: start 2.17~

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

changelog: finalise 2.16

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

dgit-badcommit-fixup: Put our version in our rewrite map commit message

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

dgit-badcommit-fixup: Skip our own output refs

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

dgit-badcommit-fixup: Instructions comment

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

Shell bug: avoid local=$(....) as it defeats set -e

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

test suite: badcommit-rewrite: Check HEAD manipulation is right

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

dgit-badcommit-fixup: Do not detach HEAD - this is not needed

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

test suite: New test for history-rewriting

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

dgit-badcommit-fixup: Now seems to work

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

dgit-badcommit-fixup: Debugging (commented out)

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

test suite: lib-core: Provide DGIT_BADCOMMIT_FIXUP

So that tests can test this program.

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

test suite: t-make-badcommit: Generate a commit we can fix up

This is very like a commit generated by #849041.  The wrongness is the
same.  The difference is that it's not in a dgit-generated merge, but
the fixup script doesn't care about that.

The existing test case which uses this (drs-push-rejects) still sees
badness and is therefore still happy.

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

dgit-badcommit-fixup: Bugfixes

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

Change name of rewrite map ref (needs an extra / for git)

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

test suite: Break out t-make-badcommit

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

dgit: Honour dgit-rewrite-map

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

Dgit.pm: Provide git_cat_file

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

dgit-badcommit-fixup: Implement dgit-rewrite-map

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

dgit-badcommit-fixup: Break out filter_updates

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

dgit-badcommit-fixup: Remove obsolete variable

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

dgit-badcommit-fixup: Do update sameness filtering a bit later

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

dgit-badcommit-fixup: Remove failing approach

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

dgit-badcommit-fixup: Ship

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

dgit-badcommit-fixup: Rename

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

badcommit-fixup: Merge from a filtered view of my personal playground

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

changelog: start 2.16~

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

badcommit-fixup: remove some debug

badcommit-fixup: dummy add seems workingish

badcommit-fixup: break out $bare

badcommit-fixup: wip dummy commits

badcommit-fixup: wip dummy commits

changelog: finalise 2.15 (will not be uploaded)

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

badcommit-fixup: break out $bare

badcommit-fixup: reorganise arg parsing

test suite: Do not tolerate any stderr output from git-fsck

We must tolerate
  notice: HEAD points to an unborn branch (master)
  notice: No default references
which are generated by some of our existing test cases and are pretty
much harmless.

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

test suite: Introduce t-expect-fsck-fail

No functional change.

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

test suite: Make expect_fsck_fail contain objids

This will make it possible to be stricter.

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

Merge tag dgit/2.14 into `defence in dgit-repos-server' branch

Fix up semantic conflict: the new git-fsck test (in 2.14) fails on the
new broken commits (in the `defence in dgit-repos-server' branch).

We need to disable this.  Ideally we would make a tighter test, but
that's too much to do in a merge.

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

test suite: Test bad commit defence in drs-push-rejects

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

dgit-repos-server: Defend against bad commits (from eg #849041).

Right now, this causes the test `debpolicy-quilt-gbp' to fail,
because in this branch #849041 is unfixed.

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

badcommit-fixup: only detach bare

badcommit-fixup: fix a git-

badcommit-fixup: seems to work

badcommit-fixup: seems to do main thing

badcommit-fixup: wip

badcommit-fixup: wip g-f-b

badcommit-fixup: wip g-f-b

badcommit-fixup: before no g-f-b

badcommit-fixup: more wip

changelog: finalise 2.14

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

badcommit-fixup wip

changelog: Document changes

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

dgit: Do not generate bogus commits

There were two places in the code where ill-formed commits were
generated: both psedumerges.

One is in dgit import-dsc, a fairly minor problem.

The other is in --overwrite and is very bad because that's an
important option.

Nothing in git seems to notice, unless you run git-fsck.  Even a git
server does not, by default.  However, some other popular git servers
do reject these broken commits.

I have tested this changes with git-fsck (by using the test suite
patches which follow in this series) and now everything seems fine.

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

test suite: Clean out erroneous reflog in a worktree

git-reflog expire --expire-unreachable=now --all

Without this, this worktree causes git-fsck to complain!

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

test suite: Run git-fsck a lot

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

test suite: Add some missing calls to t-ok.

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

test suite: Introduce t-ok.

No functional change.

 perl -i~ -pe 's/^echo ok\.$/t-ok/' tests/tests/*[^~]
 perl -i~ -pe 's/^echo ok\.$/t-ok/' tests/setup/*[^~]
 perl -i~ -pe 's/^echo done\.$/t-ok/' tests/tests/*[^~]
 perl -i~ -pe 's/^echo done\.$/t-ok/' tests/setup/*[^~]

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

changelog: start 2.14

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

changelog: start 2.14

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

changelog: finalise 2.13

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

changelog: document dgit-sponsorship fix

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

dgit-sponsorship(7): Update in light of fixed #844129

Signed-off-by: Sean Whitton <spwhitton@spwhitton.name>

changelog: document dgit-maint-gbp fix

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

dgit-maint-gbp(7): Remove reference to closed bug.

Thanks Felipe Sateler for the bug report.

Signed-off-by: Sean Whitton <spwhitton@spwhitton.name>

Import fix: Switch back to unpa branch on patch import iterations.

In particular, do not fail utterly if dpkg-source and gbp
disagree. Closes:#848843.

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

Improve "cannot represent change" message

Print the git old and new modes too.

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

quilt fixup: Permit creation of patches which delete files

By psssing --include-removal to dpkg-source, and tolerating it when we
do our quilt fixup analysis.

dpkg-source has supported this since at least stretch.

Closes:#848901.

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

changelog: start 2.13

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

changelog: finalise 2.12

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

absurd/git: Do not fail to import a .dsc containing patches to .git/ (!)

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

absurd/git Do not fail to import certain weird .dscs

A .dsc containing patches which patch files multiple times would be
rejected by dpkg-source due to #848611.

Work around this by arranging for the patch of interest to never be
the first patch, by introducing (and then evaporating) a dummy patch.

Yuk.

Closes:#848391.

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