Bugzilla – Full Text Bug Listing
| Summary: | [JSDT Bridge] [API] Enable the connector for other JavaScript support | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | [WebTools] JSDT | Reporter: | Roy Ganor <ganoro> | ||||||
| Component: | General | Assignee: | Nitin Dahyabhai <thatnitind> | ||||||
| Status: | RESOLVED WONTFIX | QA Contact: | Nitin Dahyabhai <thatnitind> | ||||||
| Severity: | enhancement | ||||||||
| Priority: | P3 | CC: | cmjaun, david_williams, evgeny.c, greg, jin.phd | ||||||
| Version: | 3.1 | Keywords: | helpwanted | ||||||
| Target Milestone: | --- | ||||||||
| Hardware: | All | ||||||||
| OS: | All | ||||||||
| Whiteboard: | |||||||||
| Bug Depends on: | 266534 | ||||||||
| Bug Blocks: | |||||||||
| Attachments: |
|
||||||||
|
Description
Roy Ganor
Created attachment 125501 [details]
patch for sse
So far I've resisted the idea of letting extensions choose their priority, as you eventually arrive at the same conflicts later on when multiple extensions have decided that they're the most important one. In fact, your updated schema says as much. Are you certain you wouldn't want to keep contributing to the current implementation? Bug 243886 has the only outstanding patch that I can find, so far, and the DLTK implementation lacks a number of refinements already in JSDT. Created attachment 125941 [details]
updated version of a patch
Updated version of a patch, schema was fixed a bit to support priority for an additional configuration option.
(In reply to comment #3) > Are you certain you wouldn't want to keep contributing to the current > implementation? Bug 243886 has the only outstanding patch that I can find, so > far, and the DLTK implementation lacks a number of refinements already in JSDT. > yes we are certain ;) The patch is processing the extension priorities as floats instead of integers. Is that intentional or something that didn't get updated between the two revisions? This is intentional, this is to allow injections of functionality in between two other implementations. Like, when there is JSDT with default priority 0 and DLTK with declared priority 1, someone can provide outline implementation that overrides JSDT but is ruled out by DLTK with priority 0.5, and float gives us a confidence that the required intermediate level will always be available. This is just another touch of configurability and changing this to integers won't break anything as of now. I think it's inappropriate to have a 'priority' in an participating environment like Eclipse plugins. The standard pattern to allow alternatives, if that is allowed at all, is to use user preferences, or product preferences, or both. Or, am I misunderstanding what this is doing? (I'm skim reading :) No, that is what it is for, and the possible implications just with different Galileo packages and installations is worrisome. I absolutely want to avoid a *user* preference if I can, though. I understand the problem, and I can see why this solution is attractive, but I just can't solve it in this way. The sourceViewerConfiguration already has a predictable list of target IDs in a predetermined order and computed at runtime to solve this kind of problem (see http://www.eclipse.org/webtools/wst/components/sse/designs/EditorConfiguration.html and the use of org.eclipse.wst.sse.ui.internal.provisional.extensions.ConfigurationPointCalculator) based on what the user is doing rather than on simply what is installed. The intended effect of a priority attribute would increase the support burden of any upstream adopter product that didn't also take steps to revert back to the current behavior when the alternate is installed. I'm not going to mark this as wontfix since I understand the problem you're faced with, but I'm rejecting this patch and approach. Categorizing JSDT bugzillas for planning purposes. Resolving given the lack of further push back and contributions in this area. I expect it will be reopened if there's still interest in pursuing this. |