chiark / gitweb /
Richard Kettlewell [Sun, 12 Jul 2009 17:23:12 +0000]
Correct tests/ clean target
Richard Kettlewell [Mon, 6 Jul 2009 19:10:16 +0000]
skip tests/hg-clone with ancient mercurial
Richard Kettlewell [Mon, 6 Jul 2009 19:04:30 +0000]
Fix wc invocation in tests
Richard Kettlewell [Sun, 5 Jul 2009 15:55:39 +0000]
Rewrote p4_commit() to handle all four cases of -m and the file list.
In one of the cases this implies that vcs invokes the user's editor
directly.
Richard Kettlewell [Sun, 5 Jul 2009 15:17:37 +0000]
For p4, vcs log with no further args now applies to current directory
and children, not the whole depot. This is more consistent with the
other vcs commands when used with p4.
Richard Kettlewell [Sat, 23 May 2009 10:29:22 +0000]
mention static-http in CHANGES
Richard Kettlewell [Sat, 23 May 2009 10:28:03 +0000]
Support hg's static-http scheme, and use that in tests.
(Why is everything but bzr so bad at getting raw repositories via http?)
Richard Kettlewell [Sat, 23 May 2009 10:27:54 +0000]
Look in right place for vcs executable in tests.
Richard Kettlewell [Sun, 17 May 2009 16:07:41 +0000]
Correct killfds logic in assemble(), which got broken in the command
execution rewrite.
Richard Kettlewell [Sun, 17 May 2009 15:56:12 +0000]
Move clone tests to their own files so they can be skipped when chiark
is down. Also added one for hg.
Richard Kettlewell [Sun, 17 May 2009 15:36:48 +0000]
Test 'vcs -n add'.
Richard Kettlewell [Sat, 16 May 2009 12:58:52 +0000]
distcheck-clean
Richard Kettlewell [Sat, 16 May 2009 12:54:07 +0000]
Rearrange. Source code goes into src/ and much is moved into a separate
libvcs.a. Tests go into tests/ (those that weren't there already) and
use the library too (where they aren't shell scripts). Tests renamed to
t- so that tests/ being on the path doesn't break things.
Richard Kettlewell [Sat, 16 May 2009 11:11:22 +0000]
Fix 'vcs -n rm -f' for darcs (and test 'vcs -n rm -f').
Richard Kettlewell [Sat, 16 May 2009 11:03:15 +0000]
Test 'vcs -n revert'.
Richard Kettlewell [Sat, 16 May 2009 10:55:21 +0000]
Test 'vcs -n clone' for git/bzr.
Richard Kettlewell [Sat, 16 May 2009 10:52:35 +0000]
Test 'vcs -n up'
Richard Kettlewell [Sat, 16 May 2009 08:51:59 +0000]
Test 'vcs -n edit'.
Richard Kettlewell [Sat, 16 May 2009 08:45:28 +0000]
Test 'vcs -n commit' and 'vcs diff'.
Richard Kettlewell [Sat, 16 May 2009 08:30:07 +0000]
tests log PWD to the right place
Richard Kettlewell [Sat, 16 May 2009 08:15:21 +0000]
Kill a few stray tabs.
Richard Kettlewell [Fri, 15 May 2009 21:07:26 +0000]
Page test depends on having a tty as stdout.
Richard Kettlewell [Fri, 15 May 2009 19:31:52 +0000]
Merge test(-related) improvements
Richard Kettlewell [Fri, 15 May 2009 19:30:11 +0000]
Restore dry-run for 'vcs commit' with p4. This stuff needs testing
better...
Richard Kettlewell [Wed, 13 May 2009 19:42:39 +0000]
Replace digit() with equivalent and built-in isdigit().
I don't remember why it was like that to start with.
Richard Kettlewell [Wed, 13 May 2009 19:34:39 +0000]
Lose P4Where::P4Where(), because not used.
Richard Kettlewell [Tue, 12 May 2009 22:02:32 +0000]
Test case for pager redirection.
Richard Kettlewell [Tue, 12 May 2009 21:52:31 +0000]
Remove LDADD assignment, redundant since revno 111.
Richard Kettlewell [Mon, 11 May 2009 21:36:55 +0000]
Issue "bzr up" for bound branches and "bzr pull" otherwise (for "vcs
up").
Richard Kettlewell [Sun, 10 May 2009 15:48:54 +0000]
Rearrange xml.cc in a slightly more idiomatic way.
Richard Kettlewell [Sun, 10 May 2009 15:06:54 +0000]
correct changelog name for debian-native package
Richard Kettlewell [Sun, 10 May 2009 15:06:40 +0000]
build-depends on libexpat-dev
Richard Kettlewell [Sun, 10 May 2009 15:05:22 +0000]
Link to Google code.
Richard Kettlewell [Sun, 10 May 2009 15:02:25 +0000]
Update compatibility list in README
Richard Kettlewell [Sun, 10 May 2009 14:35:03 +0000]
Placate compiler.
Richard Kettlewell [Sun, 10 May 2009 14:33:03 +0000]
"vcs clone" test for git now uses a branch created by git 1.1.3, which
allows tests to pass on older versions.
Richard Kettlewell [Sun, 10 May 2009 14:15:51 +0000]
Both pager and main vcs operation run in subprocesses when a pager is in
use, allowing finer control over when to terminate, what exit status to
return to the ultimate caller, etc.
Richard Kettlewell [Sat, 9 May 2009 14:54:31 +0000]
Link to Expat and mention it in README.
Richard Kettlewell [Sat, 9 May 2009 12:17:20 +0000]
Update CHANGES.
Richard Kettlewell [Sat, 9 May 2009 12:01:34 +0000]
Glibc's iconv() doens't know "char", so call nl_langinfo() instead.
Also remember to call setlocale(LC_CTYPE) so we actually have the right
encoding!
Richard Kettlewell [Sat, 9 May 2009 11:56:39 +0000]
Placate compiler.
Richard Kettlewell [Sat, 9 May 2009 11:55:10 +0000]
Merge subversion XML support.
Richard Kettlewell [Sat, 9 May 2009 11:52:51 +0000]
vcs revert for svn now invokes 'svn status --xml' and uses the output
of that to determine what to revert. Ordinary 'svn status' output
varies too much between versions to be usabel.
Richard Kettlewell [Sat, 9 May 2009 11:16:14 +0000]
(Optionally) suppress redundant character data in XML parser.
Richard Kettlewell [Sat, 9 May 2009 11:09:02 +0000]
Expat-based XML parser.
Intended to be used for parsing 'svn status --xml' output; the non-XML
output changes in format between versions making it rather hard to
parse reliably.
Richard Kettlewell [Sat, 9 May 2009 11:08:45 +0000]
More verbose test output.
Richard Kettlewell [Sat, 9 May 2009 09:31:32 +0000]
Restore '-v' behavior too(!)
Richard Kettlewell [Sat, 9 May 2009 09:08:07 +0000]
Restore dry-run mode(!) and remove a bogus 'OK'.
Richard Kettlewell [Sat, 9 May 2009 09:05:50 +0000]
Check for Curl library and header directly rather than using Curl's
Autoconf macro. This doesn't make much difference on Linux but on the
Mac means we can use the native Curl rather than (e.g.) Fink's, so it
saves a substantial build dependency.
Richard Kettlewell [Sat, 9 May 2009 09:00:07 +0000]
Only hide test output in a logfile if TESTLOG=true. When invoked
interactively the output now goes to stdout as you might expect.
Richard Kettlewell [Fri, 8 May 2009 21:53:27 +0000]
All tests should now cope without srcdir being set.
Richard Kettlewell [Fri, 8 May 2009 21:50:39 +0000]
Mac build fixes.
Richard Kettlewell [Fri, 8 May 2009 21:41:47 +0000]
Improve utils.cc test coverage a little.
Delete some more unused code.
Richard Kettlewell [Fri, 8 May 2009 21:15:40 +0000]
Remove obsolete P4Opened.
Richard Kettlewell [Mon, 4 May 2009 21:08:41 +0000]
Merge latest p4 fix
Richard Kettlewell [Mon, 4 May 2009 21:08:30 +0000]
Update CHANGES
Richard Kettlewell [Mon, 4 May 2009 21:07:15 +0000]
Add some missing p4_decode() calls.
Richard Kettlewell [Mon, 4 May 2009 20:59:12 +0000]
Specify --debug twice for verbose debug output
Normal debug output is relatively terse: commands executed but not their
input or output.
Verbose debug output includes full details of command input and output,
the conclusions of listfiles(), etc.
Richard Kettlewell [Mon, 4 May 2009 20:52:39 +0000]
Merge latest P4 fixes.
Richard Kettlewell [Mon, 4 May 2009 20:48:18 +0000]
Forgot to cope with newly added (etc) files in last change.
Now we 'p4 where' to find the local of any 'p4 opened' output that 'p4
have' doesn't help us out with.
Richard Kettlewell [Mon, 4 May 2009 20:36:38 +0000]
Substantially rewrite p4 'vcs status'.
The new model is:
- get p4 to tell us what is open using 'p4 opened'
- get p4 to tell us what it know about, and what its local paths are,
using 'p4 have'
- translate p4's absolute local paths to relative local paths
- enumerate all files as before and add any non-ignored ones as
status=?
We keep the results in a map<> with filesystem ordering, so we don't
need to merge the two sources of information in a complicated way to get
a sensible output order.
This probably obsoletes some of the code in p4utils.cc, so a revisit
might be worthwhile if it doesn't turn out to be useful for anything
else.
Richard Kettlewell [Mon, 4 May 2009 12:46:53 +0000]
Further attempts to get p4 vcs status right. Not there yet...
Richard Kettlewell [Sun, 3 May 2009 11:03:56 +0000]
Suppress unwanted error output from 'p4 where'.
There remains a crash due to output misinterpretation...
Richard Kettlewell [Sun, 3 May 2009 10:17:01 +0000]
Split out p4utils.cc.
Richard Kettlewell [Sun, 3 May 2009 10:07:55 +0000]
Merge 'p4 opened' fix
Richard Kettlewell [Sun, 3 May 2009 10:04:34 +0000]
Suppress output from empty 'p4 opened'.
Richard Kettlewell [Sun, 3 May 2009 09:56:07 +0000]
Merge up to 'vcs -n' fix
Richard Kettlewell [Sun, 3 May 2009 09:54:59 +0000]
Restore 'vcs -n'
Richard Kettlewell [Sun, 3 May 2009 09:52:17 +0000]
Debugging output
Richard Kettlewell [Sun, 3 May 2009 09:37:43 +0000]
Don't ship t-execute, t-version!
Richard Kettlewell [Sat, 2 May 2009 18:56:35 +0000]
Further use of 3-way execute
Richard Kettlewell [Sat, 2 May 2009 18:53:48 +0000]
Continue logic of standardizing on vector<string> for command lines.
Richard Kettlewell [Sat, 2 May 2009 17:54:56 +0000]
Standardize on vector<string> for command lines.
Richard Kettlewell [Sat, 2 May 2009 17:50:17 +0000]
Three-way version of execute()
Richard Kettlewell [Sat, 2 May 2009 17:34:59 +0000]
Split up construction and initialization of reader/writer monitor
classes, so they can be more usefuly used conditionally.
Richard Kettlewell [Sat, 2 May 2009 13:42:25 +0000]
Merge command execution rewrite
Richard Kettlewell [Sat, 2 May 2009 13:38:18 +0000]
Execution rewrite reaches inject()
Richard Kettlewell [Sat, 2 May 2009 13:29:55 +0000]
More execution rewrite.
Richard Kettlewell [Sat, 2 May 2009 13:25:20 +0000]
More command execution rewrite
Richard Kettlewell [Sat, 2 May 2009 13:10:50 +0000]
Basic testing for execute.cc
Richard Kettlewell [Sat, 2 May 2009 12:49:40 +0000]
Rewrite execute() to use a more general program-execution framework.
The other program execution functions aren't done yet.
Richard Kettlewell [Wed, 29 Apr 2009 18:30:49 +0000]
vcs status now shows deleted files properly for perforce.
Richard Kettlewell [Wed, 29 Apr 2009 18:30:43 +0000]
better command debug output
Richard Kettlewell [Wed, 29 Apr 2009 17:54:38 +0000]
Merge more tests
Richard Kettlewell [Tue, 28 Apr 2009 18:12:50 +0000]
Check that help/version options exist
Richard Kettlewell [Tue, 28 Apr 2009 18:08:07 +0000]
More p4 status tests
Richard Kettlewell [Tue, 28 Apr 2009 18:01:46 +0000]
Check that perforce ignore support ignores non-regular files
Richard Kettlewell [Tue, 28 Apr 2009 17:54:10 +0000]
Merge from vcs.dev
Richard Kettlewell [Mon, 27 Apr 2009 22:51:35 +0000]
make clean removes test log files
Richard Kettlewell [Mon, 27 Apr 2009 22:46:29 +0000]
Use scripts/htmlman from DisOrder
Richard Kettlewell [Mon, 27 Apr 2009 22:44:06 +0000]
Update docs + CHANGES
Richard Kettlewell [Mon, 27 Apr 2009 22:38:26 +0000]
Merge from coverage branch
Functional changes:
- revert should work on cvs in presence of conflicts
- don't add directories with p4
Many tests added.
Richard Kettlewell [Mon, 27 Apr 2009 22:36:36 +0000]
.bzrignore
Richard Kettlewell [Mon, 27 Apr 2009 22:28:24 +0000]
Link with -lgcov when doing coverage testing.
This ought to help with tracing inside forks (but it doesn't seem to...)
Richard Kettlewell [Mon, 27 Apr 2009 22:28:15 +0000]
A couple more error case tests.
Richard Kettlewell [Mon, 27 Apr 2009 22:18:01 +0000]
Typo fix
Richard Kettlewell [Mon, 27 Apr 2009 22:14:22 +0000]
Don't try to p4 add directories.
Richard Kettlewell [Mon, 27 Apr 2009 22:00:38 +0000]
Handle 'cvs -n up' exit status in the face of conflicts.
Tests exercise reversion for CVS a bit more.
Richard Kettlewell [Mon, 27 Apr 2009 21:44:05 +0000]
Use 'vcs clone' instead of native clone in tests.
Richard Kettlewell [Mon, 27 Apr 2009 21:40:50 +0000]
Add some expected failure tests.