Community
Participate
Working Groups
It is problematic, that the project-names are referenced in .classpath files. I'd like random IDs better, since that way project-dependencies could be shared with other people in a team. The way it is now, all people in a team have to have the same project-names to be abled to do that.
What would this random ID be then and how could it be shared amongst developpers ?
The Project-ID could be an addition property of a project. It than must be stored in the .project file and referenced by the .classpath files instead of the project names. The Project-ID would be a 32-bit (or longer) integer value, and it would be chosen by random if a project is created. Even among diffenrent computers, there shouldn't be much collisions.
BTW: i assume that the .project file is shared among the developers. I also filed Bug 69267 since the project-name isn't really honoured when a project is checked out. So the project-names are pretty much the badest thing to use for a file like .classpath which is shared among many developers in a team.
Then, once you checked out a project. It references subsequent projects, and only tells you about unbound IDs. How can you figure the missing ones, do you check them all out ? Having the name show up is likely a very good hint for your teammates, and you simply agree that people should change the name on checkout (unless they want to fixup all dependent project classpaths).
OK, but than Bug 69267 should be "fixed". Either make Eclipse keeping the name of the project and the project-name in the .project-file consistent, or don't store the project-name in the .project file, since it isn't necessarily used by Eclipse.
Comment#4 should have been reading: "[...] you simply agree that people should NOT change the name on checkout [...]". Bug 69267 looks like a real inconsistency, which is misleading I believe. Closing this one, since no action is planned on the build path.