[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
bug#28905: 11.90.2.2017-07-25; TeX-documentation-texdoc blocks
From: |
Ikumi Keita |
Subject: |
bug#28905: 11.90.2.2017-07-25; TeX-documentation-texdoc blocks |
Date: |
Mon, 05 Nov 2018 03:33:38 +0900 |
Hi Mosè,
>>>>> Mosè Giordano <address@hidden> writes:
> I didn't try the patch, just read it. The idea should be to launch an
> asynchronous process and show its output, correct?
Yes.
> Only one comment: is it necessary to use
> `start-process-shell-command'? Wouldn't `start-process' be
> sufficient?
Yes, mostly. The two would make little difference. I followed the
previous code which used `shell-command-to-string'. (Actually, other
parts of this function pay attention whether to use underlying shell or
not, as the comments between the code tell:
;; Note: `shell-command-to-string' uses shell, only
;; `call-process' looks at `exec-path', thus only here makes
;; sense to use `executable-find' to test whether texdoc is
;; available.
and
;; Called without prefix argument: just run "texdoc --view <pkg>" and
;; show the output, so that the user is warned in case it doesn't find
;; the documentation or "texdoc" is not available.
)
As far as I can see, the difference between `start-process' and
`start-process-shell-command' would be noticeable only when:
(1) texdoc is not available on the user's machine. In that case,
`start-process' would end in lisp error while
`start-process-shell-command' would continue to run and show the error
message "texdoc: No such file or directory" which the shell outputs.
(2) the values of `exec-path' and PATH of the invoked shell differ. If
one of them contains texdoc and other doesn't, calling texdoc might
success or not, accordingly.
Best,
Ikumi Keita