ant-ivy-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Xavier Hanin" <>
Subject Re: code review please: issue 475
Date Sat, 21 Apr 2007 12:12:09 GMT
Hello Jeff,

On 4/21/07, jeff <> wrote:
> could someone review this fix?


essentially, i added a check in IvyInstall to see if there are any
> unresolved dependencies, and if so, throw a BuildException. this is the
> default behavior, but it can be over ridden by setting haltunresolved=false
> on the install task.

This looks pretty fine. One minor thing: I think haltonunresolved instead of
haltunresolved would be more consistent with the naming used in other cases
(like haltonfailure from the resolve task).

i also added a new test class, IvyInstallTest, that tests the missing
> dependency case for haltunresolved true and false.


also, i found another test case that was failing. see
> line 117 for the failed test, and
> (diffs) for the fix.

Yes, I have fixed the problem for the release, but haven't backported the
fix to the trunk yet. I will do that when I port some changes done for the
release to the trunk. BTW, the fix is in the test and not in the
IBiblioResolver class, since we have agreed to use instead
of ibiblio as default root (ibiblio is now considered only as a mirror of

diffs attached, thanks.
> p.s., i do have apache SVN write access, which i think applies to all
> projects? so if it's okay here, should be able to commit myself, if the
> change is acceptable.

I don't know apache rules very well, but I would be surprised if you have
write access to our project. Indeed AFAIK  commiting directly is considered
as being a commiter, which requires a vote in our PPMC.


> Ahhh...imagining that irresistible "new car" smell?
> Check out new cars at Yahoo! Autos.<*;_ylc=X3oDMTE1YW1jcXJ2BF9TAzk3MTA3MDc2BHNlYwNtYWlsdGFncwRzbGsDbmV3LWNhcnM->

Learn Ivy at ApacheCon:
Manage your dependencies with Ivy!

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message