Bug 140262

Summary:	Fup of bug 135180, '_' should not be used in plugin qualifier
Product:	[Eclipse Project] JDT	Reporter:	Olivier Thomann <Olivier_Thomann>
Component:	Core	Assignee:	Olivier Thomann <Olivier_Thomann>
Status:	RESOLVED INVALID	QA Contact:
Severity:	normal
Priority:	P3	CC:	dj.houghton, jeffmcaffer, markus.kell.r, philippe_mulet
Version:	3.2
Target Milestone:	3.2 RC3
Hardware:	PC
OS:	Windows XP
Whiteboard:

Description Olivier Thomann

2006-05-04 17:13:36 EDT

Fup of bug 135180, we should remove the '_' from the plugin qualifier.

Comment 1 Jeff McAffer

2006-05-04 17:30:48 EDT

+1

Comment 2 Olivier Thomann

2006-05-04 17:35:59 EDT

Markus,

Please cast your vote.

Comment 3 DJ Houghton

2006-05-04 17:49:25 EDT

+1

Comment 4 Olivier Thomann

2006-05-04 17:54:26 EDT

I propose x631 instead of v_631 since we have to garantee that v_631 < x631.
Or for further versions:
v_631 < x632.

If I simply drop the '_', this invariant is broken.
v_631 > v632.

So using a 'x' instead of 'v' should be good enough.

Comment 5 Markus Keller

2006-05-04 18:02:44 EDT

+1 for 3.2 RC3.

Comment 6 Olivier Thomann

2006-05-04 18:30:33 EDT

Finally I plan to use 'ver'. 'v_' was meant for version. Using 'x' doesn't mean anything. Using 'ver' keeps the tag small enough.
If not good, we can still change it for 'x' afterwards, whereas changing it for 'x' could not be changed back to 'ver'.

Comment 7 Olivier Thomann

2006-05-04 22:05:01 EDT

Verified for 3.2RC3 using I20060504-2000.

Comment 8 Philipe Mulet

2006-05-05 05:55:07 EDT

I prefer 'ver' over 'x' for the reasons indicated as well. Still I don't like this workaround as it may break our version ordering, and is not addressing the proble in general (for all the versions we already emitted, thinking of other branches). We carefully chose version names which would merge well together (incrementing from v_5xx to v_6xx from 3.1 to 3.2 to leave some room for further maintenance drops). This is all broken now... 

Rather than asking clients to stop exposing the bug in plugin ID/version decoding, we should try harder to fix the parsing algorithm... note that I may be missing why it cannot be addressed...

Comment 9 Olivier Thomann

2006-05-05 12:37:22 EDT

Reopen to close as INVALID.

Comment 10 Olivier Thomann

2006-05-05 12:37:41 EDT

Closing as INVALID.
No changes were required from JDT/Core.