Fix the tab completion of paths containing \t\n
#10357
Conversation
| c,r = test_complete(s) | ||
| @test r == endof(s)-4:endof(s) | ||
| @test "space\\ .file\"" in c | ||
| for whitspace in @windows? [" "] : [" ","\t","\n"] |
There was a problem hiding this comment.
The @windows? macro is on the copping block (eg @StefanKarpinski hates it because of the un intuitive parsing). Not sure how to write this as pretty without it though.
There was a problem hiding this comment.
it's been on the chopping block for a while now. @StefanKarpinski when were you going to rewrite this as @os windows ? a : b?
|
I will look at the Travis failures tommorrow. |
| @test r == endof(s)-4:endof(s) | ||
| @test "space\\$(whitspace).file" in c | ||
|
|
||
| s = @windows? "cd(\"β $dir_space\\\\space" : "cd(\"β $dir_space/space" |
There was a problem hiding this comment.
rather than splitting this on OS here, I think you could just interpolate Base.path_separator into the string
|
The current state of the pull are able to complete paths with With this in mind I think the correction discussed in #10336, which I reverted in 2b81792 should be Should I reintroduce the correction in #10336 or a new pull and then close this in current state, for future reference if someone really would like to introduce this small feature which nobody seems to handle elegantly. cc @JeffBezanson, @vtjnash. |
|
I'm going to close this due to the above. |
|
I do think reverting 2b81792 would be a more accurate and correct representation of what we're supporting. Hopefully it's a rare corner case that not many people will hit though. |
This is my fix to the issue with path completion for paths containing
\t\ndiscussed in #10336.