On 06/09/2022 19:07, Ihor Radchenko wrote:
> Max Nikulin writes:
> 
>> Debugging `org-file-apps' and `org-open-file' regexp subgroups I noticed
>> an issue: the code distorts case of the link components making them
>> invalid. The patch to fix the bug is attached, the commit message
>> contains an example of the problem.
>>
>> Maybe I break some use case, but I am unaware when namely downcased link
>> must be used.
> 
> I tracked this dlink business down to
> 75563bf71e6df356a5ae77a93152fcf913378107.
> The relevant ML discussion is in
> https://orgmode.org/list/4B51D104.9090502@jboecker.de

It is the thread I mentioned in the commit message.

> So, it is probably safe to drop it, especially if tests are passing.

There is no tests for `org-open-file'. Checking of file existence, 
running an asynchronous process make it harder to create such tests.

>>   		    ;; First, try matching against apps-dlink if we
>>   		    ;; get a match here, store the match data for
>>   		    ;; later.
>> -		    (let ((match (assoc-default dlink apps-dlink
>> -						'string-match)))
>> +		    (let* ((case-fold-search t)
>> +                           (match (assoc-default link apps-dlink
>> +                                                 'string-match)))
> 
> With this patch, `apps-dlink' name becomes completely confusing.
> Is there any way to get rid of it as well? (or maybe rename to something
> more reasonable).

Evey attempt to read this function gives a new surprise. I agree that 
dlink has no sense any more, see the new patches.