chiark / gitweb /
stgit
16 years agoRename branch.*.stgitformatversion to branch.*.stgit.stackformatversion.
Yann Dirson [Tue, 12 Jun 2007 21:57:07 +0000 (22:57 +0100)]
Rename branch.*.stgitformatversion to branch.*.stgit.stackformatversion.

Nearly as discussed on the ml, with the name slightly modified to
better allow other types of PatchSet.

16 years agoMove stack format-version checks into Stack class.
Yann Dirson [Tue, 12 Jun 2007 21:57:07 +0000 (22:57 +0100)]
Move stack format-version checks into Stack class.

This defines an interface which is called from PatchSet, and can
(must) be implemented in classes inheriting PatchSet.

16 years agoAbstract a PatchSet object out of Series.
Yann Dirson [Tue, 12 Jun 2007 21:57:07 +0000 (22:57 +0100)]
Abstract a PatchSet object out of Series.

The goal is to move into PatchSet all functionnality that is not
dependant of Series being a real stack, paving the way for handling
parallel (hydra) stacks, or any other way to organize patches.

16 years agoUse get/set_name for a stack's name.
Yann Dirson [Tue, 12 Jun 2007 21:57:07 +0000 (22:57 +0100)]
Use get/set_name for a stack's name.

It is much more clear to use get_name() than get_branch(), and this is
a step towards factoring stuff into a PatchSet class.

16 years agoHave mail take the default for --prefix from stgit.mail.prefix.
Yann Dirson [Tue, 12 Jun 2007 21:57:07 +0000 (22:57 +0100)]
Have mail take the default for --prefix from stgit.mail.prefix.

It is more convenient to set the prefix once and for all in the config
file than each time on command line.

16 years agoHave series show empty hidden patches as hidden rather than empty.
Yann Dirson [Tue, 12 Jun 2007 21:57:07 +0000 (22:57 +0100)]
Have series show empty hidden patches as hidden rather than empty.

It is probably more important to show the hidden status of an empty
patch, since "clean" will not delete it anyway, and the display would
be ambiguous for a patch tagged "0" just before a patch tagged "!".

16 years agoTest "stg rebase" after "stg commit"
Yann Dirson [Tue, 12 Jun 2007 21:57:06 +0000 (22:57 +0100)]
Test "stg rebase" after "stg commit"

Two new tests for "stg rebase":

  1. Try to rebase to a commit that is ahead of HEAD. This should
     work, and does.

  2. Try to commit a patch, and then rebase. This doesn't work,
     because "stg rebase" aborts if orig-base != base, and "stg
     commit" doesn't update orig-base. (It does work if "stg rebase"
     is given the --force flag.)

Signed-off-by: Karl Hasselström <kha@treskal.com>
Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoChanged rebasing safety check to look for reachability of stack base (gna bug #9181).
Yann Dirson [Tue, 12 Jun 2007 21:57:06 +0000 (22:57 +0100)]
Changed rebasing safety check to look for reachability of stack base (gna bug #9181).

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoAdd my release scripts to the contrib directory
Catalin Marinas [Fri, 8 Jun 2007 22:46:11 +0000 (23:46 +0100)]
Add my release scripts to the contrib directory

Signed-off-by: Catalin Marinas <catalin.marinas@gmail.com>
16 years agoFactor out common functionality into the utils.py file
Catalin Marinas [Fri, 8 Jun 2007 22:18:04 +0000 (23:18 +0100)]
Factor out common functionality into the utils.py file

Sequences for writing or reading lists to/from files are moved to the
stgit.utils file.

Signed-off-by: Catalin Marinas <catalin.marinas@gmail.com>
16 years agoFix the hidden patches functionality (bug #9077)
Catalin Marinas [Fri, 8 Jun 2007 22:13:34 +0000 (23:13 +0100)]
Fix the hidden patches functionality (bug #9077)

This patch fixes the misleading functionality of hiding/unhiding
patches. Only unapplied patches can be hidden and they are now on a
separate list. Commands like 'stg push --all' will not push hidden
patches.

Signed-off-by: Catalin Marinas <catalin.marinas@gmail.com>
16 years agoFix bug #9284 - error messages include abbreviated command
Catalin Marinas [Thu, 7 Jun 2007 22:13:57 +0000 (23:13 +0100)]
Fix bug #9284 - error messages include abbreviated command

Signed-off-by: Catalin Marinas <catalin.marinas@gmail.com>
16 years agoFixed typo in contrib/stg-whatchanged.
Yann Dirson [Wed, 6 Jun 2007 22:01:52 +0000 (00:01 +0200)]
Fixed typo in contrib/stg-whatchanged.

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoCatch early trying rebasing to unknown ref, and add testcase.
Yann Dirson [Wed, 6 Jun 2007 22:01:47 +0000 (00:01 +0200)]
Catch early trying rebasing to unknown ref, and add testcase.

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoRobustify rebase test: check patches are reapplied.
Yann Dirson [Wed, 6 Jun 2007 22:01:42 +0000 (00:01 +0200)]
Robustify rebase test: check patches are reapplied.

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoFix removal of series to nuke the formatversion config item.
Yann Dirson [Wed, 6 Jun 2007 21:05:22 +0000 (23:05 +0200)]
Fix removal of series to nuke the formatversion config item.

Since this parameter is now used to decide if the branch has already
been initialised, not removing it forbids to create a stack with the
same name as one that was deleted.

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoFix removal of series with non-existant trash dir.
Yann Dirson [Wed, 6 Jun 2007 21:05:17 +0000 (23:05 +0200)]
Fix removal of series with non-existant trash dir.

Otherwise it is not possible to delete a stack that never had a patch
deleted (eg. a newborn stack).

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoFix deletion and move of a hidden patch (gna bug #9244).
Yann Dirson [Mon, 4 Jun 2007 21:14:43 +0000 (23:14 +0200)]
Fix deletion and move of a hidden patch (gna bug #9244).

We must unhide a patch before we delete it, or we get an exception
while trying to unhide a non-existent patch.

Similarly, when renaming a hidden patch, we must take care of hiding
an unhiding at the right moment.

Both problems introduced in commit
841c7b2a6b015e3ab3331713cf10caf7a4fd0b49 when first indroducing patch
hiding.

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoAdd new --diff-opts/-O flag to diff- and status-related commands.
Yann Dirson [Thu, 31 May 2007 22:34:38 +0000 (00:34 +0200)]
Add new --diff-opts/-O flag to diff- and status-related commands.

This new flag allows to pass arbitrary flags to the git-diff calls
underlying several StGIT commands.  It can be used to pass flags
affecting both diff- and status-generating commands (eg. -M or -C), or
flags only affecting diff-generating ones (eg. --color, --binary).

It supercedes --binary for commands where it was allowed,
'-O --binary' is now available for the same functionality.

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoMake diff flags handling more modular.
Yann Dirson [Thu, 31 May 2007 22:34:33 +0000 (00:34 +0200)]
Make diff flags handling more modular.

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoCall external commands without a shell where possible.
Yann Dirson [Thu, 31 May 2007 22:34:28 +0000 (00:34 +0200)]
Call external commands without a shell where possible.

On my dev box it consistently improves performance when timing the
full testsuite:

before:

user    2m22.509s
sys     0m50.695s

user    2m23.565s
sys     0m49.399s

user    2m23.497s
sys     0m49.675s

after:

user    2m20.261s
sys     0m45.687s

user    2m21.485s
sys     0m46.427s

user    2m20.581s
sys     0m45.367s

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoAdd 2 new contrib scripts.
Yann Dirson [Thu, 31 May 2007 22:34:23 +0000 (00:34 +0200)]
Add 2 new contrib scripts.

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoFix the automatic patch name generation for the pick command
Catalin Marinas [Tue, 22 May 2007 20:45:53 +0000 (21:45 +0100)]
Fix the automatic patch name generation for the pick command

It was broken by a recent commit to unify the patch name generation.

Signed-off-by: Catalin Marinas <catalin.marinas@gmail.com>
16 years agoStGIT: Support Cc: fields in stg mail --auto
Petr Baudis [Tue, 22 May 2007 20:45:53 +0000 (21:45 +0100)]
StGIT: Support Cc: fields in stg mail --auto

Aside of Signed-off-by and Acked-by fields, Cc fields may occur in the
message footer; include them in the auto-composed mail Cc list.

Signed-off-by: Petr Baudis <pasky@suse.cz>
16 years agoAdd a "make tags" target
Catalin Marinas [Mon, 21 May 2007 20:58:39 +0000 (21:58 +0100)]
Add a "make tags" target

Signed-off-by: Catalin Marinas <catalin.marinas@gmail.com>
16 years agoDon't use / as separatar since it is common i branch names
Robin Rosenberg [Mon, 21 May 2007 20:58:39 +0000 (21:58 +0100)]
Don't use / as separatar since it is common i branch names

Don't use / as separatar since it is common i branch names

Signed-off-by: Robin Rosenberg <robin.rosenberg@dewire.com>
16 years agoUpdate the bash prompt from 'applied' instead of the obsolete 'current'
Robin Rosenberg [Mon, 21 May 2007 20:58:39 +0000 (21:58 +0100)]
Update the bash prompt from 'applied' instead of the obsolete 'current'

Signed-off-by: Robin Rosenberg <robin.rosenberg@dewire.com>
söndag 20 maj 2007 skrev Yann Dirson:
> Note that "tail -1" gives a warning with newer versions, "tail -n 1"
> should be the proper call.
My man page doesn't mention -N being deprecated, but ok since -n 1 seems
to work here too.

> Also I'm not sure it is a good way to look at "applied" file, since
> Karl IIRC has plans to change this.  Better call "stg top" and not
> touch that again :)

Calling stg is too slow to be be used here. I that command in my first draft
for this function and people complained (see the thread named "Bash snippet
to show branch and patch in bash prompt"). It takes ~ 0.15s on here which is
very noticable, barely below my pain threshold.

We'll update the prompt when and if Karl breaks this.

It'd probably drain my battery too :/

-- robin

 contrib/stgbashprompt.sh |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

16 years agoRefactor message printing
Karl Hasselström [Mon, 21 May 2007 20:58:39 +0000 (21:58 +0100)]
Refactor message printing

Feed all those little progress and status messages through a central
place where such things as newlines, indentation, and suppressing
output when not on a TTY can be taken care of once and for all.

This patch takes care of almost, but not quite, all output. The most
notable leftover is the printing of help messages.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoAdd --binary flag to commands that generate diffs
Karl Hasselström [Mon, 21 May 2007 20:58:38 +0000 (21:58 +0100)]
Add --binary flag to commands that generate diffs

This just passes the --binary option to git-diff-*, which causes the
generated diffs to contain an applyable diff even when binary files
differ. It's necessary to do this if you want to mail patches to
binary files.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoTest the format version upgrade code
Karl Hasselström [Mon, 21 May 2007 20:58:38 +0000 (21:58 +0100)]
Test the format version upgrade code

This test contains tarballs of repositories created with older
versions of StGIT. It also contains the script used to generate them,
but at some point we will lose the ability to easily run old versions
-- for example, if git changes incompatibly -- so tarballs will be the
only practical option for sufficiently old versions.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoUpgrade old StGIT branches to new-format metadata
Karl Hasselström [Mon, 21 May 2007 20:58:38 +0000 (21:58 +0100)]
Upgrade old StGIT branches to new-format metadata

There used to be a "stg branch --convert" command that switched
between "old" and "new" format metadata. But my recent metadata
cleanup patches introduced a "new new" format, and more are hopefully
on the way, so it was time to start versioning the metadata format
explicitly in order to avoid future headaches.

This patch removes the "stg branch --convert" command, and makes StGIT
automatically upgrade older formats to the latest format. It stores
the format (as an integer) in the config file. The current metadata
format version number is 2 (the "old" format is 0, and the "new"
format is 1).

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoHave only a single command in each test_expect_failure
Karl Hasselström [Mon, 21 May 2007 20:58:38 +0000 (21:58 +0100)]
Have only a single command in each test_expect_failure

Otherwise, we can't know which one failed.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoFix config caching so that get, set, get works
Karl Hasselström [Mon, 21 May 2007 20:58:38 +0000 (21:58 +0100)]
Fix config caching so that get, set, get works

The config caching was never invalidated or updated, which caused the
two gets to always return the same value regardless of the value
passed to set.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoRemove obsolete files when deleting a branch
Karl Hasselström [Tue, 15 May 2007 19:38:40 +0000 (21:38 +0200)]
Remove obsolete files when deleting a branch

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoStore branch description in the config file
Karl Hasselström [Tue, 15 May 2007 17:10:02 +0000 (18:10 +0100)]
Store branch description in the config file

Instead of storing the branch description in an StGIT-specific file,
store it in the git config file, where tools other than StGIT can read
and write it.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoUse list comprehension in uncommit.py for Python 2.3
Catalin Marinas [Tue, 15 May 2007 17:10:02 +0000 (18:10 +0100)]
Use list comprehension in uncommit.py for Python 2.3

The (None for ...) is not allowed in Python 2.3. Changed to [...]

Signed-off-by: Catalin Marinas <catalin.marinas@gmail.com>
16 years agoUncommit to a named commit
Karl Hasselström [Tue, 15 May 2007 17:10:02 +0000 (18:10 +0100)]
Uncommit to a named commit

Add a new flag to "stg uncommit": --to. This flag takes a committish
parameter, and uncommits everything up until that commit.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoIf any uncommit would fail, don't uncommit anything
Karl Hasselström [Tue, 15 May 2007 17:10:01 +0000 (18:10 +0100)]
If any uncommit would fail, don't uncommit anything

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoRemove an unnecessary parameter to make_patch_name
Karl Hasselström [Tue, 15 May 2007 17:10:00 +0000 (18:10 +0100)]
Remove an unnecessary parameter to make_patch_name

The "alternative" parameter was a boolean that indicated whether we
were interested in testing if the first generated name was acceptable,
or if we would just always accept it. But that can be accomplished by
giving an "unacceptable" function that always returns False, so
there's no need for an additional parameter.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoGenerate patch names of more uniform length
Karl Hasselström [Tue, 15 May 2007 17:10:00 +0000 (18:10 +0100)]
Generate patch names of more uniform length

Cut the generated patch name at 30 characters after getting rid of
unwanted characters, not before. This gives patch names of more
uniform length, since consecutive unwanted characters were replaced by
a single dash.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoMake the "name" argument to "stg new" optional
Karl Hasselström [Tue, 15 May 2007 17:09:59 +0000 (18:09 +0100)]
Make the "name" argument to "stg new" optional

If no name is given, one is generated from the commit message. This
can be very handy.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoRename "bury" back to "sink".
Yann Dirson [Tue, 15 May 2007 17:09:58 +0000 (18:09 +0100)]
Rename "bury" back to "sink".

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoStop recording branch.*.remote to '.' for local parents.
Yann Dirson [Tue, 15 May 2007 17:09:57 +0000 (18:09 +0100)]
Stop recording branch.*.remote to '.' for local parents.

We should be using the remote name of the remote branch in
branch.*.merge, instead of the local one as we currently do (I'll
ensure this in a later patch).

As a consequence, setting "remote" to '.' and "merge" to the name of a
*local* parent would become inconsistent with the remote case.  Anyway
the local case should make use of pull-policy=rebase instead, and this
change revealed that we were fetching useless information in the local
case (no need for a remote).

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoFix typo in the patchlogs copying code
Catalin Marinas [Tue, 15 May 2007 17:09:55 +0000 (18:09 +0100)]
Fix typo in the patchlogs copying code

Signed-off-by: Catalin Marinas <catalin.marinas@gmail.com>
16 years agoStore branch description in the config file
Karl Hasselström [Mon, 14 May 2007 19:28:10 +0000 (21:28 +0200)]
Store branch description in the config file

Instead of storing the branch description in an StGIT-specific file,
store it in the git config file, where tools other than StGIT can read
and write it.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoUse "stg applied" instead of reading the applied file directly
Karl Hasselström [Sun, 13 May 2007 18:27:30 +0000 (20:27 +0200)]
Use "stg applied" instead of reading the applied file directly

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoTest patch order, not just number of patches
Karl Hasselström [Sun, 13 May 2007 18:27:25 +0000 (20:27 +0200)]
Test patch order, not just number of patches

In addition to testing the number of applied patches, also test their
order. And test the order of unapplied patches too while we're at it.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoMake patch deletion test more specific
Karl Hasselström [Sun, 13 May 2007 18:27:20 +0000 (20:27 +0200)]
Make patch deletion test more specific

Don't just test the number of applied and unapplied patches; also test
their name and order.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoDocument patch syntax.
Yann Dirson [Sun, 13 May 2007 14:11:54 +0000 (16:11 +0200)]
Document patch syntax.

Acked-by: Karl Hasselström <kha@treskal.com>
Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoMake the "name" argument to "stg new" optional
Karl Hasselström [Fri, 11 May 2007 01:40:05 +0000 (03:40 +0200)]
Make the "name" argument to "stg new" optional

If no name is given, one is generated from the commit message. This
can be very handy.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoFix popping on non-active branches
Karl Hasselström [Tue, 8 May 2007 23:32:46 +0000 (01:32 +0200)]
Fix popping on non-active branches

Resetting HEAD on the current branch when popping a patch on another
branch isn't a great idea.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoTeach bash to complete branch names in some places.
Yann Dirson [Sun, 6 May 2007 23:13:57 +0000 (01:13 +0200)]
Teach bash to complete branch names in some places.

This patch causes bash completion to look for branch names instead of
patch names for "branch" and "rebase".

Unfortunately, the current completion framework does not seem flexible
enough to get branch completion after "-b" flags, but I feel things
are already better this way.

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoDon't use patches/<branch>/current
Karl Hasselström [Sun, 6 May 2007 15:13:44 +0000 (17:13 +0200)]
Don't use patches/<branch>/current

The name of the current patch, if any, is always the last line of
patches/<branch>/applied (and there is no current patch if and only if
the "applied" file is empty). So use that instead, and stop having to
worry about keeping the redundant "current" file up-to-date.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoCopy patchlogs when cloning a stack or picking a patch.
Yann Dirson [Sat, 5 May 2007 14:00:28 +0000 (16:00 +0200)]
Copy patchlogs when cloning a stack or picking a patch.

This will allow to keep patch history accross those operations, and
see when they diverged when looking at the respective patchlogs.  With
more work, that will allow to locate the common ancestor when sync'ing
a patch across branches.

The work on "pick" can still be improved: currently the patchlog is
preserved on "pick patch[@branch][//top]", but we could make it work
for //top.old as well (or any rev in the patchlog when we'll be able
to name those).

Patchlog preservation in pick can only work when referencing a patch
by name.  When picking an arbitrary commit we could also improve
things by recording a "pick" operation instead of a "new".

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoCleanup variable names in pick.
Yann Dirson [Sat, 5 May 2007 14:00:23 +0000 (16:00 +0200)]
Cleanup variable names in pick.

I'm going to need patch objects there, so avoid using variable "patch"
to store only a patchname.

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoAvoid contrib/stg-swallow deleting unrelated empty patches.
Yann Dirson [Sat, 5 May 2007 13:58:07 +0000 (15:58 +0200)]
Avoid contrib/stg-swallow deleting unrelated empty patches.

Let's just check ourselves if the patch is empty after pushing instead
of using clean: no more surprise with patches getting abusively
cleaned, and we should now be faster when there are many unapplied
patches.

Signed-off-by: Yann Dirson <ydirson@altern.org>
16 years agoFix deleting series trash directory
Roland Dreier [Thu, 3 May 2007 18:29:41 +0000 (11:29 -0700)]
Fix deleting series trash directory

When deleting a series, the loop that iterates through the files in
__trash_dir needs to add __trash_dir to the filename that it passes to
os.remove().

Signed-off-by: Roland Dreier <rolandd@cisco.com>
16 years agoDon't use refs/bases/<branchname>
Karl Hasselström [Sun, 29 Apr 2007 22:15:52 +0000 (00:15 +0200)]
Don't use refs/bases/<branchname>

It's silly to save the stack base in a ref when it can trivially be
computed from the bottommost applied patch, if any. (If there are no
applied patches, it's simply equal to HEAD.)

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agodash doesn't do curly brace expansion, so avoid that
Karl Hasselström [Sun, 29 Apr 2007 21:09:55 +0000 (23:09 +0200)]
dash doesn't do curly brace expansion, so avoid that

I have dash as /bin/sh, and it doesn't seem to expand stuff in curly
braces, which breaks t1800. Work around this by not using curly brace
expansion to generate test data.

Signed-off-by: Karl Hasselström <kha@treskal.com>
16 years agoString comparison is done with "=", not "=="
Karl Hasselström [Sun, 29 Apr 2007 21:09:50 +0000 (23:09 +0200)]
String comparison is done with "=", not "=="

The "test" builtin in bash accepts both "==" and "=", but it didn't
work with dash as /bin/sh.

Signed-off-by: Karl Hasselström <kha@treskal.com>
17 years agoFailed recursive merging can silently lose date in StGIT
Catalin Marinas [Tue, 17 Apr 2007 16:29:32 +0000 (17:29 +0100)]
Failed recursive merging can silently lose date in StGIT

The logic in git.merge(recursive = True) is that if
git-merge-recursive fails, deal with the unmerged entries via diff3
and interactive merge (if 'autoimerge' is set to 'yes' in
config). There is a situation, however, when git-merge-recursive fails
because of an untracked file that would be overwritten by the
merge. In this case, git-merge-recursive aborts the operation but
without any unmerged entries in the index. StGIT considers that the
merge lead to an empty patch without reporting any error.

This commit makes it a bit safer by detecting whether there are
unmerged index entries and, if not, further raises an error informing
the user of the problem. The patch is still pushed as empty but the
user is informed of the failure and the possibility of running 'push
--undo'.

A better solution, but which require a bit more work, is to
distinguish between the merge failures and, in this specific case,
abort the push completely and revert the patch to its original state
(and popped from the stack).

Signed-off-by: Catalin Marinas <catalin.marinas@gmail.com>
17 years agoDocument the new 'stg branch --description' features.
Yann Dirson [Thu, 12 Apr 2007 20:15:36 +0000 (22:15 +0200)]
Document the new 'stg branch --description' features.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoRevert commit "Move man page stg(1) to stg(7)"
Catalin Marinas [Wed, 11 Apr 2007 23:01:10 +0000 (00:01 +0100)]
Revert commit "Move man page stg(1) to stg(7)"

This commit is reverting 5906924135f886527c1e947d100073067783eaa2 and
also moves stg.txt to section 1 in Documentation/Makefile.

Signed-off-by: Catalin Marinas <catalin.marinas@gmail.com>
17 years agoFix the 'bury' command completion to include all patches
Catalin Marinas [Wed, 11 Apr 2007 23:01:10 +0000 (00:01 +0100)]
Fix the 'bury' command completion to include all patches

Signed-off-by: Catalin Marinas <catalin.marinas@gmail.com>
17 years agoMove man page stg(1) to stg(7)
Catalin Marinas [Tue, 10 Apr 2007 22:09:17 +0000 (23:09 +0100)]
Move man page stg(1) to stg(7)

Signed-off-by: Catalin Marinas <catalin.marinas@gmail.com>
17 years agoAdd "stg bury" command, with the functionality of contrib/stg-sink.
Yann Dirson [Tue, 10 Apr 2007 18:27:02 +0000 (20:27 +0200)]
Add "stg bury" command, with the functionality of contrib/stg-sink.

This is the rewrite in python of by stg-sink written in perl.

I changed the name to "bury" since it seems more descriptive of what
it does, despite being less of an opposite to "float" than "sink" was.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoAdd --description option to branch
Catalin Marinas [Tue, 10 Apr 2007 21:47:10 +0000 (22:47 +0100)]
Add --description option to branch

This options allows one to set the description text of a branch.

Signed-off-by: Catalin Marinas <catalin.marinas@gmail.com>
17 years agoStop advertising 'pull' as the only operation blocked on protected stacks.
Yann Dirson [Tue, 10 Apr 2007 21:47:10 +0000 (22:47 +0100)]
Stop advertising 'pull' as the only operation blocked on protected stacks.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoMake the documentation of options more consistent.
Yann Dirson [Tue, 10 Apr 2007 21:47:10 +0000 (22:47 +0100)]
Make the documentation of options more consistent.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoFix doc cross-refs.
Yann Dirson [Tue, 10 Apr 2007 21:47:10 +0000 (22:47 +0100)]
Fix doc cross-refs.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoAdd doc for 'branch'.
Yann Dirson [Tue, 10 Apr 2007 21:47:10 +0000 (22:47 +0100)]
Add doc for 'branch'.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoAdd doc for 'clone' and 'init'.
Yann Dirson [Tue, 10 Apr 2007 21:47:10 +0000 (22:47 +0100)]
Add doc for 'clone' and 'init'.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoDon't use section 7 for main manpage.
Yann Dirson [Tue, 10 Apr 2007 21:47:09 +0000 (22:47 +0100)]
Don't use section 7 for main manpage.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoFix typo in help string.
Yann Dirson [Tue, 10 Apr 2007 21:47:09 +0000 (22:47 +0100)]
Fix typo in help string.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoCleanup parent information on stgit branch deletion.
Yann Dirson [Fri, 16 Mar 2007 22:45:45 +0000 (22:45 +0000)]
Cleanup parent information on stgit branch deletion.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoAdd support for unsetting config vars.
Yann Dirson [Fri, 16 Mar 2007 22:45:45 +0000 (22:45 +0000)]
Add support for unsetting config vars.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoAdd a script to quickly run an interactive test session.
Yann Dirson [Fri, 16 Mar 2007 22:45:45 +0000 (22:45 +0000)]
Add a script to quickly run an interactive test session.

This uses the capabilities of the test library to provide a throw-away
playground to test how the current stgit tree behaves.

To use it, run the following command:
(cd t && ./interactive.sh)

You'll get a shell in a brand new git tree.  Exiting that shell will
get rid of this temporary directory.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoRemove debugging output from contrib/stg-sink
Yann Dirson [Fri, 16 Mar 2007 22:45:45 +0000 (22:45 +0000)]
Remove debugging output from contrib/stg-sink

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoCopy parentbranch setting on 'stg branch --clone'.
Yann Dirson [Fri, 16 Mar 2007 22:45:44 +0000 (22:45 +0000)]
Copy parentbranch setting on 'stg branch --clone'.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoDrop utils.remove_dirs() in favor of os.removedirs().
Yann Dirson [Fri, 16 Mar 2007 22:45:44 +0000 (22:45 +0000)]
Drop utils.remove_dirs() in favor of os.removedirs().

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoMake use exception raised by removedirs.
Yann Dirson [Fri, 16 Mar 2007 22:45:44 +0000 (22:45 +0000)]
Make use exception raised by removedirs.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoFix diagnostic messages on patch deletion and simplify others.
Yann Dirson [Fri, 16 Mar 2007 22:45:44 +0000 (22:45 +0000)]
Fix diagnostic messages on patch deletion and simplify others.

Also make use standard os.removedirs(), allowing for the same
simplification.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoFix stack deletion when orig-base is present.
Yann Dirson [Fri, 16 Mar 2007 22:45:43 +0000 (22:45 +0000)]
Fix stack deletion when orig-base is present.

The introduction of the orig-base file makes the stack deletion fail
to remove the stack's directory, leaving the stack in a zombie state,
where attempting to recreate it was refused because of the directory
already existing, but attempting to delete it failed since the applied
and unapplied files were not found.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoAdd an empty line before signed-off-by
Robin Rosenberg [Fri, 16 Mar 2007 22:45:43 +0000 (22:45 +0000)]
Add an empty line before signed-off-by

Default to the standard convention of having an empty line
before the Signed-off line when addign a signature using the
--sign option.

Signed-off-by: Robin Rosenberg <robin.rosenberg@dewire.com>
Signed-off-by: Catalin Marinas <catalin.marinas@gmail.com>
17 years agoFixed t2102-pull-policy-rebase to really test 'rebase' policy.
Yann Dirson [Mon, 12 Mar 2007 22:44:22 +0000 (22:44 +0000)]
Fixed t2102-pull-policy-rebase to really test 'rebase' policy.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoMake 'stg cp' 2nd form safe.
Yann Dirson [Mon, 12 Mar 2007 22:44:21 +0000 (22:44 +0000)]
Make 'stg cp' 2nd form safe.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoSome clarifications to the main doc.
Yann Dirson [Mon, 12 Mar 2007 22:44:21 +0000 (22:44 +0000)]
Some clarifications to the main doc.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoDo not link to docs that will never be written.
Yann Dirson [Mon, 12 Mar 2007 22:44:21 +0000 (22:44 +0000)]
Do not link to docs that will never be written.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoAdd 'stg cp' command.
Yann Dirson [Mon, 12 Mar 2007 22:44:21 +0000 (22:44 +0000)]
Add 'stg cp' command.

Supports copying several files and dirs into existing dir, as well as
copying single file or dir with a new name.

In the case of directory copying, only copies files known to git (ie,
no generated files).  Should maybe add --all as alternative.

Does not check whether we're going to lose data (overwriting any
modified file, or any file not tracked by git).  This should probably
be done first, before doing any real change (preparing the changes
in-memory, then do the work if OK).  Will give us --dry-run at no
additional cost.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoOnly check for upstream merges if not fast-forwarding
Catalin Marinas [Mon, 12 Mar 2007 22:44:20 +0000 (22:44 +0000)]
Only check for upstream merges if not fast-forwarding

Signed-off-by: Catalin Marinas <catalin.marinas@gmail.com>
17 years ago[Revert] Make 'push --undo' safer
Catalin Marinas [Thu, 7 Dec 2006 22:02:42 +0000 (22:02 +0000)]
[Revert] Make 'push --undo' safer

(reverting of commit 539fb2921209bd3fd6165baa057f76f474aa5cb4)

Make the push undoing a bit safer by asking the user to first clean
the local changes or conflicts rather than silently resetting them.

Signed-off-by: Catalin Marinas <catalin.marinas@gmail.com>
17 years agoFix bash completion to not garble the screen with an error message.
Yann Dirson [Tue, 6 Mar 2007 20:18:44 +0000 (21:18 +0100)]
Fix bash completion to not garble the screen with an error message.

Maybe we could refuse to complete further when the command was not
recognized, but at least this fix avoids the annoying behaviour.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoAdd contrib/stg-sink: sink a patch to given location (mirrors float).
Yann Dirson [Fri, 2 Mar 2007 21:34:29 +0000 (21:34 +0000)]
Add contrib/stg-sink: sink a patch to given location (mirrors float).

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoAdd contrib/stg-mdiff: display diffs of diffs.
Yann Dirson [Fri, 2 Mar 2007 21:34:29 +0000 (21:34 +0000)]
Add contrib/stg-mdiff: display diffs of diffs.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoAdd contrib/stg-gitk: helper script to run gitk
Yann Dirson [Fri, 2 Mar 2007 21:34:29 +0000 (21:34 +0000)]
Add contrib/stg-gitk: helper script to run gitk

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoAdd contrib/stg-cvs: helper script to manage a mixed cvs/stgit working copy
Yann Dirson [Fri, 2 Mar 2007 21:34:28 +0000 (21:34 +0000)]
Add contrib/stg-cvs: helper script to manage a mixed cvs/stgit working copy

This is an early prototype only, use with care, and be sure to read the
LIMITATIONS section in the script comments.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoAdd contrib/stg-swallow: completely merge an unapplied patch into current one
Yann Dirson [Fri, 2 Mar 2007 21:34:28 +0000 (21:34 +0000)]
Add contrib/stg-swallow: completely merge an unapplied patch into current one

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoAdd contrib/stg-fold-files-from: pick selected changes from a patch above current one
Yann Dirson [Fri, 2 Mar 2007 21:34:28 +0000 (21:34 +0000)]
Add contrib/stg-fold-files-from: pick selected changes from a patch above current one

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoAdd contrib/stg-show-old: show old version of a patch.
Yann Dirson [Fri, 2 Mar 2007 21:34:28 +0000 (21:34 +0000)]
Add contrib/stg-show-old: show old version of a patch.

Signed-off-by: Yann Dirson <ydirson@altern.org>
17 years agoAdd contrib/stg-whatchanged: look at what would be changed by refreshing.
Yann Dirson [Fri, 2 Mar 2007 21:34:28 +0000 (21:34 +0000)]
Add contrib/stg-whatchanged: look at what would be changed by refreshing.

This script outputs a "metadiff" (diff of diffs, ie. diff between 2
versions of a single patch).  This is somewhat a proof of concept, and some
work is indeed needed in the way to present the results.  Consider
filtering the output at least through colordiff.

I have 2 uses for this script:
- when simply editing a patch, provides a 3rd way to check what I've done
  before refreshing (in addition to "stg diff" and "stg diff -r //bottom")
- most usefully, when resolving conflicts caused by a push, to ease
  checking of the merge result.

Signed-off-by: Yann Dirson <ydirson@altern.org>