ant-ivy-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From James Poli <>
Subject Cache problem?
Date Tue, 26 Oct 2010 11:46:51 GMT
I'm evaluating moving to Ivy 2.2.0 from 2.0.0 for our builds.  I've noticed one of our integration
tests on Windows started failing at 2.2.0.  It's finding a module in the cache now when it
didn't before.  I've included the pseudo code below.  Note we do some special processing of
the module we're actually building so we want to know if it exists in the repository before
we actually build it:).

pseudo code for building project A:
   clean repository
   findrevision module A
      if found
         resolve module A
      build module A
      publish module A
      resolve module A

The pseudo code above runs twice in an AntUnit test.  The first run passes but the second
run fails; the findrevision gets a hit in the cache (found) so it tries to resolve the module
but that fails since repository is has been cleaned.
My first inclination is that the test needs to clean the cache between runs.  But then again
I started thinking how does Ivy validate the cache?  Shouldn't findrevision know the module
isn't valid anymore in the cache similar to the resolve?  It appears to work this way in 2.0.0.

Perhaps a bug got fixed in the cache management?

What's the best way to know if your cache has expired if the repository has been cleaned?

Anyway thanks for any help in resolving this (no pun intended :)).

James L. Poli
SAS Institute Inc.
Sas Campus Drive (R4251)
Cary, NC 27513
Phone: 919-531-6066
SAS...  The Power to Know

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