| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
This makes it easier to tweak log verbosity and redirect logs to other
handlers. For example, the WSGI servers are unstable with stderr
closed, and crash with an IOError if they try to print a warning to
stderr.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
The bulk of the work is in regard to XML, with new BugDir.xml and
.from_xml methods to support the new <bugdir> entity. I also split
the guts import_xml's ._run method into sub-methods to make the import
logic more obvious.
|
| |
|
| |
|
| |
|
|
|
|
| |
Sometimes saxutils returns unicode.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
| |
This happens most often when a previous crash leaves an empty
directory `abc` in .be/.../bugs/abc/ or
.be/.../bugs/.../comments/abc/. The new exception ensures the error
message is "Bug/Comment X missing value file" which tells you where to
look for the repository corruption not "summary is None" which told
you nothing.
|
| |
|
| |
|
| |
|
|\ |
|
| | |
|
|/ |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
$ python test.py libbe.command.init
Doctest: libbe.command.init.Init ... FAIL
...
-----------------------
File ".../libbe/command/init.py", line 47, in libbe.command.init.Init
Failed example:
ui.run(cmd)
Exception raised:
Traceback (most recent call last):
...
File "/tmp/be.wtk/libbe/command/init.py", line 97, in _run
bd = libbe.bugdir.BugDir(storage, from_storage=False)
File "/tmp/be.wtk/libbe/bugdir.py", line 185, in __init__
self.save()
File "/tmp/be.wtk/libbe/bugdir.py", line 228, in save
self.save_settings()
File "/tmp/be.wtk/libbe/bugdir.py", line 204, in save_settings
mf = mapfile.generate(self._get_saved_settings())
File "/tmp/be.wtk/libbe/storage/util/settings_object.py", line 230, in _get_saved_settings
self, self._setting_name_to_attr_name(k))
File "/tmp/be.wtk/libbe/storage/util/properties.py", line 194, in _fget
value = fget(self)
File "/tmp/be.wtk/libbe/storage/util/properties.py", line 329, in _fget
primer(self)
File "/tmp/be.wtk/libbe/storage/util/settings_object.py", line 69, in prop_load_settings
self.load_settings()
File "/tmp/be.wtk/libbe/bugdir.py", line 194, in load_settings
self.settings = mapfile.parse(settings_mapfile)
File "/tmp/be.wtk/libbe/storage/util/mapfile.py", line 123, in parse
c = yaml.load(contents)
...
File "/usr/lib/python2.6/site-packages/yaml/reader.py", line 213, in update_raw
data = self.stream.read(size)
AttributeError: 'NoneType' object has no attribute 'read'
...
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Note that it only upgrades on-disk versions, so you can't use a
non-VCS storage backend whose version isn't your command's current
storage version. See #bea/110/bd1# for reasoning. To see the on-disk
storage version, look at
.be/version
To see your command's supported storage version, look at
be --full-version
I added test_upgrade.sh to exercise the upgrade mechanism on BE's own
repository.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
duplicate_bugdir() works, but for the vcs backends, it could require
shelling out for _every_ file read. This could, and probably will, be
horribly slow. Still it works ;).
I'm not sure what a better implementation would be. The old
implementation checked out the entire earlier state into a temporary
directory
pros: single shell out, simple upgrade implementation
cons: wouldn't work well for HTTP backens
I think a good solution would run along the lines of the currently
commented out code in duplicate_bugdir(), where a
VersionedStorage.changed_since(revision)
call would give you a list of changed files. diff could work off of
that directly, without the need to generate a whole duplicate bugdir.
I'm stuck on how to handle upgrades though...
Also removed trailing whitespace from all python files.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Targeting normal bugs
With "be depend". I think we should remove the "target" field from
bugs, and move target dependencies over into the "be depend"
framework.
* be target list
Would become "be list --severity target". A target "severity" would
keep target bugs distinct from other bug/issue types.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This flag allows us to skip unittest and testsuite declaration if we
woln't need them. It speeds up simple be calls a suprising amount.
With Testing=True (the old behavior):
wking@thor:be.wtk$ time ./be > /dev/null
real 0m0.393s
user 0m0.340s
sys 0m0.048s
With TESTING=False (the new behavior):
be.wtk$ time ./be > /dev/null
real 0m0.216s
user 0m0.152s
sys 0m0.064s
This adjustment was inspired by Jakub Wilk's Debian bug:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=559295
|
|
|
|
|
|
|
|
| |
The old allow_changes and allow_new_comments didn't have separate
handling for extra_strings, like import_xml will need. It also didn't
have a way to specify what to do if an illegal change occurs.
Sometimes you'll want to raise an exception, but sometimes you'll want
to ?silently? ignore the change.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This makes Bug.add_comment simpler. Also makes Bug.from_xml() more
robust, since it no longer depends on the order in which the XML file
lists the comments. The previous Bug.from_xml() would have choked on
<be-xml>
<bug>
<comment>
<uuid>B</uuid>
<in-reply-to>A</in-reply-to>
</comment>
<comment>
<uuid>A</uuid>
</comment>
</bug>
</be-xml>
because when B was being added, the referenced A hadn't yet been
noticed.
|
|
|
|
|
|
|
|
|
|
|
| |
This is a pretty critical feature, dunno how I missed it before.
I also added a little check to both Bug and Comment.from_xml() so that
xml_string can take an ElementTree Element as well as the usual raw
string/unicode. This avoids repeated string <-> Element conversions.
Added Bug.add_comment() which handles the addition of a Comment
instance, matching .in_reply_to, checking .uuid uniqueness, etc.
|