Support special IDREF validation via ExtensionsErrorReporter.isPresent#1849
Support special IDREF validation via ExtensionsErrorReporter.isPresent#1849merks merged 1 commit intoeclipse-pde:masterfrom
Conversation
|
For without this fix we have these warnings: With the fix we have only these warnings: These are legitimate because those are not in the target platform in the self-hosted debug launch. |
HannesWell
left a comment
There was a problem hiding this comment.
From what you have described, it sounds like there is no good solution for the problem, besides reworking the specification and validation of extension-points in general.
Of course the latter would be better then to have special handling of certain extension-point values in PDE, but I would be ok with this.
OTOH, a relatively simple generic solution could be introducing the possibility to declare a custom validator (for attributes?) when defining a new extension-point? Then platform.ui could inject the custom validation for the affected points.
- The quicklinks.exsd defines two attributes that are specified as IDREFs to command IDs but they actually permit a command spec and even wildcards. - This leads to warnings when those aspects are used and those warnings are impossible for the user to fix. - If we disable the basedOn type for those, the extension attributes no longer support Browse... to find a command, which makes authoring very inconvenient. - Ideally PDE is made aware of these special attribute so support specialized validation.
07229c9 to
a492478
Compare
|
FYI @jonahgraham Woo hoo, with this change, the EPP environment can be provisioned without any warnings: |



Support special IDREF validation via ExtensionsErrorReporter.isPresent
In EPP there are a bunch of quicklinks definitions, e.g.,
https://github.com/eclipse-packaging/packages/blob/d0d154a80c075e857dcf531396546686ec844dba/packages/org.eclipse.epp.package.committers/plugin.xml#L390-L439
So we end up with lots of warnings that we can't eliminate: