[ https://issues.apache.org/jira/browse/NPANDAY-599?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14105206#comment-14105206 ]
Lars Corneliussen edited comment on NPANDAY-599 at 8/21/14 8:48 AM: -------------------------------------------------------------------- We talked about making i a per-project somehow, didn't we? {panel} [07.08.2014 12:04:04] Lars Corneliussen: iirc somehow the deps were cached between reactor projects [07.08.2014 12:04:34] Lars Corneliussen: but I can't remember what happend then... if it failed, or if it just included pdbs you wouldn't want... [07.08.2014 12:06:07] Brett Porter (NPanday team): can see why it could be a problem, maybe better just to have the cache be mapped per project instead [07.08.2014 12:09:19] Lars Corneliussen: if per-project is available, it seems much more approprieate than per-lookup1 [07.08.2014 12:22:07] Brett Porter (NPanday team): no per project lookup, but we can do that in the maps {panel} was (Author: lcorneliussen): We talked about making i a per-project somehow, didn't we? [code] [07.08.2014 12:04:04] Lars Corneliussen: iirc somehow the deps were cached between reactor projects [07.08.2014 12:04:34] Lars Corneliussen: but I can't remember what happend then... if it failed, or if it just included pdbs you wouldn't want... [07.08.2014 12:06:07] Brett Porter (NPanday team): can see why it could be a problem, maybe better just to have the cache be mapped per project instead [07.08.2014 12:09:19] Lars Corneliussen: if per-project is available, it seems much more approprieate than per-lookup1 [07.08.2014 12:22:07] Brett Porter (NPanday team): no per project lookup, but we can do that in the maps [code] > Cache resolve-result reactor-wide to avoid pdb/gac/++ beeing resolved > multiple times > ------------------------------------------------------------------------------------ > > Key: NPANDAY-599 > URL: https://issues.apache.org/jira/browse/NPANDAY-599 > Project: NPanday > Issue Type: Sub-task > Components: Maven Plugins > Reporter: Lars Corneliussen > Fix For: 1.5.0-incubating > > > Currently it seems like our custom contributors/resolvers execute for every > possible path (artifact.getDependencyTrail() changes...). > We can cache what we resolved reactor-wide in a separate component. > Cachekey would be artifact.getId(), right? > Should we let each contributor implement the cache > (ArtifactResolvingContributor) - or should it be generically handled in the > artifact resolver? -- This message was sent by Atlassian JIRA (v6.2#6252)