| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For self.parsed_version[0] >= 2, was returning the entire output of
._u_invoke_client() (status, output, and error). Now it just returns
the output, which is what we want. This fixes Chris' failed test:
======================================================================
FAIL: Should get file contents as committed to specified revision.
----------------------------------------------------------------------
Traceback (most recent call last):
File "libbe/vcs.py", line 852, in test_revision_file_contents_as_committed
self.test_contents['rev_1'], committed_contents)
AssertionError: 'Lorem ipsum' != (0, 'Lorem ipsum', '')
----------------------------------------------------------------------
|
| |
|
|
|
|
|
| |
The version string is useful information to have around, while the
help string is probably not. For example, we use it in
darcs.Darcs._vcs_get_file_contents() to construct an incantation
appropriate to the version we're dealing with.
|
| |
|
|
|
|
|
|
|
| |
*.__del__() is run some unspecified time after the refcount for an
object is reduced to zero. Sometimes that means that the rest of the
world has already been deallocated, which makes life difficult,
especially when Python won't attempt to construct stack traces inside
*.__del__(). We were always (hopefully ;) calling del(*) anyway,
so we just replace those calls with *.cleanup()
|
| |
|
|
|
| |
If an exception is raised during command exection, self.ret may be
left at None.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Allowed values for both are "yes" or "no", case insensitive.
Setting Confirm:no disables the confirmation email letting you know
your new bug was successfully submitted.
Setting Subscribe:yes automatically subscribes you to the new bug
you're submitting immediately, without reqiring an extra control
email.
In the case where both Confirm:yes and Subscribe:yes, the submitter
will only recieve the subscription email (which shows the "be show"
view of the new bug) and not the submission email (which shows the
output of all the executed submission commands).
Both of these pseudoheaders were in response to a
Would you like a direct response concerning your comments?
checkbox on the web frontend that generates emails for a BE repo.
When the checkbox is set, we set Confirm:yes Subscribe:yes, and the
submitter gets a nice pretty confirmation email and will be
automatically notified of any future action on their bug. When the
checkbox is not set, they get no response or submission emails.
Also fixed up some bugs in in the subscriber_emails() parsing. We now
use ordered_subscriptions to ensure that we parse the "DIR"
subscription first, since we don't want to include the same
information twice, and the user might be subscribed to DIR:all and a
particular bug (which would be wierd, but is easily avoidable). This
also ensures that the more central DIR level changes appear first in
the email.
We also check that there is an entry for a particular bug in bug_index
before attempting to grab it, which could raise KeyErrors otherwise.
Finally, we check to make sure we don't double-include bugs for which
the user is subscribed to both that bug's new and DIR:new.
A final minor correction is the capitalization of the first word of
our log entries. The logs are pretty cluttered, and the
capitalization helps the lines I care about stand out. A better
solution would be to come up with a cleaner idea of what to write to
the logs...
|
| |
|
|
|
| |
Because subscribe.get_bugdir_subscribers() only looks for subscribers
in bugs that are already in memory.
|
| |
|
|
|
|
| |
It had been eating the last line of the message (which wasn't a
problem when I had a blank line before the footer, but was when I had
neither footer nor trailing blank line).
|
| |
|
|
|
| |
To conform with Python convention for "hiding" class attributes only
intended for internal use.
|
| |
|
|
|
| |
Check that the string has the right tag before attempting to parse it
:p.
|
| |
|
|
|
| |
It's useful to have some way to see the current subscription status
from the command line ;).
|
| |
|
|
|
|
|
|
|
|
| |
The
if new_fields[u'date'] != None:
bit avoids attemting to normalize missing dates (which fails). You
get messages with missing dates when comment_message_to_xml() is
called recursively for multipart messages.
Also fixed some unicode keys (['X'] -> [u'X']) for consistency.
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Switched from cStringIO to StringIO in be-handle-mail because:
(from http://docs.python.org/library/stringio.html)
"Unlike the memory files implemented by the StringIO module, those
provided by this [cStringIO] module are not able to accept Unicode
strings that cannot be encoded as plain ASCII strings."
I'm not sure what all the fuss with sys.__stdin__ had been about in
Command.run(), but I took it out and everything still seems to work
;).
Also fix Makefail (again!) to install under $HOME by default.
|
| |
|
|
|
|
|
|
|
| |
"comments" should always be a directory if it exists, but if it is
something else, be should fail rather than return an empty comment
list.
This fix was inspired by some browsing around in response to Thomas
Habets' version of a fix for missing "bugs" directory.
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Two major benefits:
1) Programatic access to the BE version. Previously the
version-string logic was internal to the commandline interface ./be,
which left the other interfaces hanging.
2) Simple to override auto-generated revision id version with any
string you like (by setting libbe.version._VERSION, if that's the way
you want to run things or the way your packaging system requires.
Note that other modules should never read libbe.version._VERSION,
since it may not be defined. Instead they should use the output of
libbe.version.version().
|
| | |
|
| | |
|
| |
|
|
|
| |
Because otherwise we run into problems when running be from inside
cron jobs.
|
| |
|
|
|
|
| |
I changed the class name in commit
wking@drexel.edu-20090718201613-n242qfl4s3j3kfjf
but missed two references.
|
| |
|
|
| |
Also a few minor tweaks to the module imports.
|
| | |
|
| | |
|
| |\ |
|
| | | |
|
| | |
| |
| |
| |
| | |
Also moved pre-YAML mapfile handling in mapfile.parse() into
upgrade.Upgrade_1_0_to_2._upgrade_mapfile().
|
| | | |
|
| | | |
|
| |/ |
|
| | |
|
| | |
|
| |
|
|
| |
Also fleshed out the be-handle-mail docstring.
|
| | |
|
| |
|
|
|
|
|
|
|
|
| |
Previously:
$ be init
$ be list
...
File ".../libbe/bugdir.py", line 537, in list_uuids
for uuid in os.listdir(self.get_path("bugs")):
OSError: [Errno 2] No such file or directory: '.../.be/bugs'
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
| |
Also adjusted update_copyright.sh to ignore Gianluca's temporary use
of
gianluca <gian@galactica>
since he's now using
Gianluca Montecchi <gian@grys.it>
on becommands/html.py.
|
| |\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Changes to merge with my branch:
* Added some reference to "be html" or "becommands/html.py" in his
new bugs' summaries. Bug titles should make clear to which aspect
of our growing repository they refer.
* Fixed unittests in becommands/html.py.
* execute() kwarg "test"->!"manipulate_encodings
* bugdir.simple_bug_dir() -> bugdir.SimpleBugDir()
* bd.cleanup() at end of unittests
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
