Some Eclipse Foundation services are deprecated, or will be soon. Please ensure you've read this important communication.

Bug 337216

Summary:	Previously resolved maven components ignore maven provider settings
Product:	z_Archived	Reporter:	Matt Biggs <zebbedi>
Component:	Buckminster	Assignee:	buckminster.core-inbox <buckminster.core-inbox>
Status:	NEW ---	QA Contact:
Severity:	normal
Priority:	P3	CC:	thomas
Version:	unspecified
Target Milestone:	---
Hardware:	PC
OS:	Linux
Whiteboard:

Description Matt Biggs

2011-02-15 09:34:02 EST

Build Identifier: helios-maintenance

It appears that on the helios-maintenance branch, some changes have been made to the MavenCSpecBuilder. The changes look correct, but has a knock on effect. 

It means that when materialising against a previously materialised workspace, the maven components are iterated with a 'local' reader instead of a maven reader. This in turn means that the changes i submitted in bug 299491 along with any other settings specified in your maven reader are ignored.

Line 86 of the MavenCSpecBuilder calls addDependencies, but if the component is already on disk, the reader it passes in is a local reader and therefore has no knowledge of things like transitive or scopes or map entries. 

Somehow i guess it needs to pass in the preferences specified for the maven provider, but i have no idea how to achieve that. 

Reproducible: Always

Steps to Reproduce:
1. Perform a buckminster 'import' on a workspace that has some maven dependencies ensuring that you specify a scope or transitive=false.
2. Re-perform a buckminster 'import' on the same workspace
3. Buckminster will ignore your maven provider preferences (eg transitive) and try to resolve everything including all its sub-dependencies etc.

Comment 1 Matt Biggs

2011-06-15 08:21:59 EDT

Hope nobody minds me pinging this. It's a bit of a show stopper for being able to use the latest buckminster. 

I can no longer materialize more than once.

Comment 2 Thomas Hallgren

2011-06-15 08:41:51 EDT

Matt, we don't have much resources to put on this right now. A way forward could be if you provided a patch. If you do, I promise to give it priority.

Comment 3 Matt Biggs

2011-06-15 09:03:46 EDT

Cheers, I'll take a look. May just see if we can suppression checking previously materialized components via a preference or something.