Always match friendly sessions for special buffers #3432
Conversation
cider-repl.el
Outdated
| @@ -1746,72 +1746,76 @@ constructs." | |||
| (mapconcat #'identity (cider-repl--available-shortcuts) ", ")))) | |||
| (error "No command selected"))))) | |||
|
|
|||
| (defvar cider--project-agnostic-buffer-names | |||
There was a problem hiding this comment.
Just check that some of those names are not configurable.
There was a problem hiding this comment.
Btw, it seems to me this is more or less cider-ancillary-buffers, so I guess you should use it instead.
vemv
force-pushed
the
friendly-special-buffers
branch
from
7d52b87
to
f963e35
Compare
|
I pushed some unrelated changes, please ignore for now |
vemv
force-pushed
the
friendly-special-buffers
branch
from
f963e35
to
7d52b87
Compare
|
Reviewable again! |
|
Some tests are failing (other than the Windows tests). |
vemv
force-pushed
the
friendly-special-buffers
branch
from
7d52b87
to
f4477b4
Compare
|
The tests pass again if I remove the second commit that observed Perhaps, it had some nuance? ...For completeness, here's the removed commit: 7d52b87 |
|
Perhaps the list is not full at the time the test is being run? (you'll see that buffers add themselves there after their definitions) Maybe you can just set it to some explicit value for the purpose of the test? |
vemv
force-pushed
the
friendly-special-buffers
branch
from
f4477b4
to
0ae3e03
Compare
| @@ -60,6 +60,12 @@ | |||
| (before-each | |||
| (setq sesman-sessions-hashmap (make-hash-table :test #'equal) | |||
| sesman-links-alist nil | |||
| cider-ancillary-buffers (seq-filter (lambda (s) | |||
There was a problem hiding this comment.
Adding this fixed the tests.
|
Looks good to me now. |
|
Just passing by to say "thank you!" for this. It happened only sporadically to me, and I've always thought it was a problem with my configuration, but I've never really sit down to debug it better |
|
❤️🍻! We now also have this troubleshooting page for session linking: https://docs.cider.mx/cider/troubleshooting.html#interactions-fail-with-no-linked-cider-sessions |
*cider-error*,*cider-result*, and possibly others, always deserve to get an associated repl, if possible.This normally already is the case, because their
default-directorymatches that of the current repl.However, for some reason, occasionally their default-directory will be something like
~/.emacs.dwhich matches no repl's dir.So this PR fixes that heisenbug.
But it also is an overall improvement, because conceptually,
*cider-error*/*cider-result*belong to no project in particular, so if a repl is available, it should be considered 'friendly' to it.Cheers - V