search

Wilfred / ag.el

An Emacs frontend to The Silver Searcher
http://agel.readthedocs.org/en/latest/
525 stars 61 forks source link

Slightly more intuitive ag-reuse-buffers behavior #111

Closed gordonmcshane closed 8 years ago

gordonmcshane commented 8 years ago

This is very helpful behavior when you have multiple ag results buffers open, and need to re-run ag on them from time to time, but don't want to clobber *ag search*.

Wilfred commented 8 years ago

This seems confusing to me. Could you clarify?

Suppose I search for functions in my git repo, so I have a buffer called *ag search text:defun dir:~/projects/ag.el/*. With this change, if I run M-x ag-project RET defvar RET, I have a buffer called *ag search text:defun dir:~/projects/ag.el/* that contains a text search for "defvar". I don't think that's what users expect.

Are you manually renaming buffers, or using some other workflow?

gordonmcshane commented 8 years ago

Ahh, I hadn't thought of that.

RE: renaming buffers, yes, what I do is: ag RET some-symbol-name RET /some/path RET which I then rename. I also do this with (compile), but compilation-mode works as I expect. Say I visit *renamed compilation buffer* and press g. I get recompilation in that same buffer using the original parameters passed to (compile), however, do this in an ag-mode buffer with ag-reuse-buffers set to non-nil and whatever was in the default *ag search* is clobbered and replaced with an ag search using the parameters from *renamed ag search buffer* instead of just re-using that same buffer. To give you a use case, I might have a few ag search buffers open that serve as a "TODO list" of deprecated function usage. As I update the code, I often will return to the ag buffer and press g to get an updated count of matches.

I guess what I was attempting to accomplish was having ag-mode behave more like compilation-mode in regards to (recompile).

Wilfred commented 8 years ago

Thanks, that's really helpful.

So, I think the goal here is to allow users to run recompile in renamed buffers. I think that's definitely worth supporting, although this current PR doesn't quite achieve that. This is probably something that should work regardless of the value of ag-reuse-buffers.

gordonmcshane commented 8 years ago

I absolutely agree that my solution falls short. If I have time, I'll revisit and see if I can't come up with something to handle the situation in a generic fashion.

Wilfred commented 8 years ago

I'm going to close this. If you do come up with a generic solution, I'd be very happy to merge.