On 20/03/2012 15:22, Colm O hEigeartaigh wrote:
Hi,
I think this would be a useful convention to follow. A more important
thing (imo) is to make sure that the JIRA number is in the subject of
the svn commit, so that you can search for a particular issue in JIRA,
and see what commit fixed it.
...but we should already be following this, isn't it?
2012/3/19 Francesco Chicchiriccò<[email protected]>:
Hi all,
in the past, when reporting an issue that was judged significant, we used
(as best-practice) to create a dedicated test case so that we could keep
sure that the fix for that significant issue was still working as
development was proceeding.
For this reason you can find many test cases named "issueXXX" where XXX is
the issue number on GoogleCode; execute the following command from the
directory where you've checked out /trunk if you want to have an idea:
egrep -ri issue[0-9]+ core/src/test/java/ | grep public
After switching at ASF, we barely continued such habit, only changing a bit
the name pattern for test cases to "issueSYNCOPEYYY" where SYNCOPE-YYY is
the issue number on ASF JIRA; execute the following command from the
directory where you've checked out /trunk if you want to have an idea:
grep -ri issueSYNCOPE core/src/test/java/ | grep public
Now I am wondering:
1. Do you think this could be useful? I do, mainly for the reason reported
at the beginning: once you provide a fix for a significant issue, I believe
it's important to be sure that a later commit won't brake that fix.
2. Is there any best practice about this at ASF?
TIA
Regards.
--
Francesco Chicchiriccò
Apache Cocoon PMC and Apache Syncope PPMC Member
http://people.apache.org/~ilgrosso/
