fixes Wrong colour for keyword "length" in keyword list #206 #207
Conversation
| syn keyword elixirKeyword contained abs bit_size byte_size div elem hd length | ||
| syn keyword elixirKeyword contained map_size node rem round tl trunc tuple_size | ||
| " Kernel functions | ||
| syn match elixirKernelFunction contained containedin=elixirGuard '\<is_atom\|is_binary\|is_bitstring\|is_boolean\|is_float\|is_function\|is_integer\|is_list\|is_map\|is_nil\|is_number\|is_pid\|is_port\>\([ (]\)\@=' |
There was a problem hiding this comment.
I think we should use a group on these regex: \<\(is_atom\|..., right?
There was a problem hiding this comment.
I thought so too, but from testing it doesn't seem to make a difference from what I can tell (AFAIK it doesn't matter as long as we don't make use of the captures in the regex). Maybe there's an edge case I'm missing?
There was a problem hiding this comment.
In theory without the group the \< and \> are not applied to each word in the list. So, this: \<\(A\|B\)\> is equal to \(\<A\>\|\<B\>\), whilst \<A\|B\> isn't.
There was a problem hiding this comment.
Better example:
AB
A
B
- with
\<A\|B\>will match with all 3 lines; - with
\<\(A\|B\)\>will match only with the last two lines.
There was a problem hiding this comment.
@jbodah are you going to make this change or should I take this one over?
jbodah
force-pushed
the
kernel-functions
branch
from
80e7b39 to
5fe9dec
Compare
|
Here's a simplified version of the problem:
And here's this branch:
|
jbodah
force-pushed
the
kernel-functions
branch
from
5fe9dec to
7adebea
Compare
|
@kassio should be all set now |
|
Just fix the conflicts with master and I'll hit the green button. |
…s#206 Note that this also removes highlighting in some places that I found it to be useful (e.g. using `is_pid` in a function body) but that can be discussed in a different PR
jbodah
force-pushed
the
kernel-functions
branch
from
7adebea to
d52089a
Compare
|
@kassio rebased |
fixes Wrong colour for keyword "length" in keyword list #206
Note that this also removes highlighting in some places that I found it
to be useful (e.g. using
is_pidin a function body) but that can bediscussed in a different PR