Rick Hillegas wrote:
On 2/16/11 12:59 PM, Mike Matrigali wrote:
Rick Hillegas wrote:
On 2/16/11 11:56 AM, Knut Anders Hatlen wrote:
Mike Matrigali<[email protected]> writes:
Just a reminder to not put colon after your derby number in svn
commits.
This defeats the automatic mechanism that JIRA uses to make your svn
commits show up in the JIRA history for an issue. It really makes
backporting fixes hard when you can't see the commits in the JIRA.
For instance the following svn commit will not show up:
DERBY-4711: Hung thread after another thread is interrupted
If a thread fails while waiting for a lock, remove that thread from
the queue.
While the following will:
DERBY-4711 Hung thread after another thread is interrupted
If a thread fails while waiting for a lock, remove that thread from
the queue.
That's odd... I've always been using a colon, and looking at
https://issues.apache.org/jira/browse/DERBY-5032 the following commit
message indeed does show up when I click on "Subversion Commits":
,----
| DERBY-5032: derived InternalXact attempts (and fails) to override
a method from base
|
| Patch contributed by Dave Brosius<[email protected]>.
`----
This is the first I have heard about this problem. Like Knut, I have
been putting colons in my commit messages for years. When did people
start noticing this problem? That may help us track it down.
Thanks,
-Rick
Looking at it closer it does seem like sometimes it works and sometimes
it does not. I just assumed it was consistent. It does not seem to be
consistent with a user also. For instance DERBY-5015 by knut does not
show a svn commit for me, while 5032 does.
Thanks, Mike. Why do you suspect the problem is the colon? Maybe there
is some other detail about the 5032 message which triggers this problem.
Thanks,
-Rick
I just remember early on (meaning years ago) having problems with some
sort of punctuation being right next to the derby number. I assumed
that the way it works is that there is a scanner that runs over the
comments looking for jira
numbers of format XXXX-NNNN , and that at least at some point is was not
smart enough to deal with anything other than whitespace around the
token. Given the number of entries we have with colon's it must be
something else. Every time I see a missing one the comment has the
colon and I have never seen a missing entry where there was no colon but
these are just mostly observations while I have been working on
backports and could just be explained by most of the entries having
colons.
