+dgit (1.5~~) unstable; urgency=medium
+
+ Incompatible change:
+ * dgit sbuild does not pass -A to sbuild. Consequently the default
+ build is now simply sbuild's default. With older sbuilds it was
+ possible to override dgit's -A by passing another option. But this
+ has been changed recently and now this default setting is very awkward
+ to change for the dgit user.
+
+ Major new feature:
+ * --quilt=gbp, --quilt=dpm, --quilt=unpacked: Introduce facility for
+ split view (dgit/mainiainer view), to improve compatibility with some
+ other workflow tools.
+
+ New checks and improved behaviours:
+ * When running dpkg-buildpackage, cope if user specified -g or -G.
+ * dgit sbuild: check that the set of .changes files found is as we
+ expect, before calling mergechanges. Re:#800060.
+ * dgit sbuild: Rename the used-up .changes files to `.inmulti' to
+ avoid accidental use of the wrong one (by software, or by users).
+ * dgit sbuild: Check that the binary .changes file doesn't contain a
+ .dsc.
+ * Introduce --rm-old-changes to delete previous builds' changes files.
+ * Remove any pre-existing _source.changes file before building source,
+ as a safety check.
+ * No longer tolerate a multitude of .changes files when doing push.
+ Instead, insist on a single one. Closes:#800110.
+ * dgit sbuild no longer deletes extranious .changes files; instead
+ we rely on --rm-old-changes, or failing that, fail early.
+ * When doing quilt linearisation, treat upstream .gitignores not
+ in the toplevel the same way we treat ones in the toplevel.
+ * When automatically generating quilt patch, honour GIT_COMMITTER_DATE
+ for filename creation (makes filename deterministic in test suite).
+ * New --overwrite option, replaces need to for user to use
+ git merge -s ours. Closes:#838718.
+ * When generating quilt patches from git commits, make patches that
+ look quite like git-format-patch output (rather than strange things
+ based on an obselete interpretation of DEP-3).
+ * When generating quilt patches from git commits, honour (and strip)
+ any Gbp-Pq headers (that we understand).
+ * Several dgit-generated commits now have slightly better annotations
+ from dgit about what it was doing.
+ * Make --quilt=gbp the default for dgit gbp-build.
+ * New tag format (for dgit view) archive/debian/VERSION.
+ * Before committing to push, check that .dsc and .changes correspond.
+ Closes:#800060.
+ * Better error message if non-split-brain patch stack no longer
+ applies (due to new upstream version, or user messing with it).
+ Closes:#833025.
+ * Better error message if HEAD contains changes unrepresentable
+ by `3.0 (quilt)'. Closes:#834618.
+
+ Infrastructure:
+ * Better error handling in dgit-repos-policy-debian.
+ * dgit-repos-policy-debian.: fix git-cat-file-handling with multiple
+ taints in db (!).
+
+ Documentation:
+ * Document the dgit-distro.DISTRO.quilt-mode config setting.
+ * Clarify the --clean= options' documentation. Closes:#800054.
+ * Discourage use of the --PROGRAM:OPTION escape hatch. (Apropos
+ of various bug reports including #800060 and #833025.)
+ * Document the expected form of HEAD for each --quilt= mode.
+
+ Bugfixes:
+ * When cleaning up after failed clone, stat the to-be-cleaned-up
+ directory before running rmtree on it. Closes:#796773.
+ * Do not call "warn" on failure of cleanup handler in END block
+ (since warn has been made fatal and aborts the cleanup chain).
+ * Print better error message (with `fail' rather than `die') if
+ `dgit clone' cannot create the destination directory.
+ * Properly substitute $changesfile in one of the `You can retry'
+ messages. Closes:#800078.
+ * Pass --ch:* and -v options to dpkg-buildpackage when building
+ source. Fixes bad Perl poetry syntax. Closes:#829121.
+ * When synthesing a commit from a .dsc from the archive, stop
+ internal git reset from printing a confusing message about HEAD.
+ * Turn off git gc in the private working areas.
+ * Do not fail to do some important quilt processing in some
+ --quilt modes.
+ * Fix two calls to chdir without proper error checking.
+ * Fix a couple of bugs in error reporting.
+ * Fix up .orig detection to be less trustful of (ambiguous) filenames.
+ * Fix up .orig detection to correctly allow exactly the right
+ .orig-comp components accorging to dpkg-source(1).
+ * Fix up .orig detection to handle .asc's (and other signatures)
+ of .orig components the same way as their tarballs.
+ * Tidy up refs/dgit-fetch/ after dgit fetch (if successful).
+ * Fix handling of in-archive copies.
+ * Don't break if user has push.followTags=true. Closes:#827878.
+ * Arrange for the special dgit remote to be skipped by git fetch --all
+ etc. Closes:#827892.
+ * Allow local git config options to override user-global ones,
+ as is proper. Closes:#835858.
+ * Get sense of failed clone failed cleanup error reporting check
+ correct. Closes:#796773.
+
+ Test suite:
+ * When sbuild fails, do not crash due to sed not finding the log
+ file. Instead, simply tolerate the absence of the log file.
+ * Put --no-arch-all in build-modes-sbuild act, not only its real_act.
+ Cosmetic change only.
+ * Set GIT_COMMITTER_DATE and GIT_AUTHOR_DATE and increment them
+ explicitly in drs-push-rejects test. This avoids date dependencies
+ which can cause that test to fail on fast computers.
+ * Remove some spurios .debs from the example_1.0.tar.
+ * Increase sqlite_busy_timeout in debpolicy-dbretry, because
+ zealot is very slow and we need to give the other processes time
+ to rollback and release the lock.
+ * Test quilt single-debian-patch.
+ * Provide `tartree-edit gitfetchinfo' etc. to help with comparing
+ different test case git working tree tarballs.
+ * Test dgit-repos-policy-debian with multiple (identical, as it happens)
+ existing taints.
+ * Provide better log output for certain failures.
+
+ --
+
dgit (1.4) unstable; urgency=high
Bugfixes: