bug#61371: 30.0.50; Adding support for jdt:// file scheme in eglot

[Theodor Thornhill]

Danny Freeman <danny@dfreeman.email> writes:

> Danny Freeman <danny@dfreeman.email> writes:
>
>> Theodor Thornhill <theo@thornhill.no> writes:
>>
>>> Michael Albinus <michael.albinus@gmx.de> writes:
>>>
>>>> Theodor Thornhill via "Bug reports for GNU Emacs, the Swiss army knife
>>>> of text editors" <bug-gnu-emacs@gnu.org> writes:
>>>>
>>>>> Hi!
>>>>
>>>> Hi Theo,
>>>>
>>>>> When using Eglot along with Java there are some hacks that needs to be
>>>>> made to make things functional.  One of the more important things is to
>>>>> add support for the jdt:// file scheme that is sent when you try to
>>>>> go-to-definition on a system or third party lib.
>>>>>
>>>>> As you can see, we need to query the server with
>>>>> :java/classFileContents, and also register with the server to send it
>>>>> with the :extendedClientCapabilities on initialization.
>>>>>
>>>>> Is there a place this code could live? Maybe in the new java-ts-mode?
>>>>
>>>> There is the jarchive package on GNU ELPA. Could this be used, perhaps 
>>>> extended?
>>>>
>>>>> Theo
>>>>
>>>> Best regards, Michael.
>>>
>>> Maybe! Don't know what it does, but I'll check it out - thanks!
>>>
>>> Theo
>>
>> JDT urls are outside the scope of jarchive. I would think of them less
>> as URLs and more as tokens to be decoded by the LSP server that provides
>> them. JDT urls are intended to be sent back to the LSP server using a
>> special extension method.
>>
>> There is an issue open in the eglot-java repo to implement this but I
>> have been too busy to get around to it
>> https://github.com/yveszoundi/eglot-java/issues/6
>>
>> Someone has responded there with some code similar to jarchive that will
>> parse the JDT urls and try to open them, but I wouldn't consider that a
>> permanent solution. JDT urls are not standardized and subject to change
>> by the JDT LSP maintainers.
>
> I think this bug can be closed now, it was fixed in the eglot-java
> package. See: https://github.com/yveszoundi/eglot-java/issues/6
>
> Thank you,
> -- 
> Danny Freeman

I agree :-)