[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
bug#16045: 24.3.50; rgrep can't work
From: |
Eli Zaretskii |
Subject: |
bug#16045: 24.3.50; rgrep can't work |
Date: |
Wed, 04 Dec 2013 19:24:07 +0200 |
> Date: Wed, 4 Dec 2013 13:57:45 +0800
> From: zijianyue <zijianyue@163.com>
>
> Hello,rgrep worked well before I use this version 24.3.50.
> Now,rgrep always found nothing in emacs,lgrep works well.I don't know why.
Michael, this is due to this commit of yours (almost a year ago!):
revno: 111276
committer: Michael Albinus <michael.albinus@gmx.de>
branch nick: trunk
timestamp: Thu 2012-12-20 11:15:38 +0000
message:
* progmodes/grep.el (rgrep): Escape command line. Sometimes, it
is too long for Tramp. See discussion in
<http://thread.gmane.org/gmane.emacs.tramp/8233/focus=8244>.
* progmodes/compile.el (compilation-start): Remove line escape template.
Command lines with newlines are non-portable: Windows shells don't
support them, and don't treat backslashes specially anyway (that's why
we have shell-quote-argument). So what that change did on Windows is
add literal \r\n strings to the command, making it wrong at best and
un-parsable at worst.
Since the original problem was with Tramp, i.e. with remote files, I
suggest the patch below; it worked for me on Windows. Please see if
there's anything wrong with it; if not, I will commit.
Btw, I cannot say I like the fact that the command displayed in the
*grep* buffer is different from what is actually passed to the shell.
It makes debugging much harder when some error occurs. In my case,
'find' displayed several warnings like this:
find: warning: Filenames usually don't contain slashes (though pathnames do).
That means that '-name .#*\
' will probably evaluate to false all the time on this system. You might
find the '-wholename' test more useful, or perhaps '-samefile'. Alternatively,
if you are using GNU grep, you could use 'find ... -print0 | grep -FzZ .#*\
'.
and I stared at this for several minutes trying to figure out what the
heck was it talking about, since the command displayed in the buffer
had no backslashes at all. I needed to look at the command line in
GDB to see what is actually being sent. So I think this is not a good
idea at all.
Here's the patch I suggest:
=== modified file 'lisp/progmodes/grep.el'
--- lisp/progmodes/grep.el 2013-05-24 20:54:38 +0000
+++ lisp/progmodes/grep.el 2013-12-04 17:10:42 +0000
@@ -1005,7 +1005,9 @@ to specify a command to run."
(mapconcat
#'shell-quote-argument
(split-string files)
- (concat "\\\n" " -o " find-name-arg " "))
+ (concat
+ (if (file-remote-p dir) "\\\n")
+ " -o " find-name-arg " "))
" "
(shell-quote-argument ")"))
dir
@@ -1026,7 +1028,9 @@ to specify a command to run."
(concat "*/"
(cdr ignore)))))))
grep-find-ignored-directories
- "\\\n -o -path ")
+ (if (file-remote-p dir)
+ "\\\n -o -path "
+ " -o -path "))
" "
(shell-quote-argument ")")
" -prune -o "))
@@ -1044,7 +1048,9 @@ to specify a command to run."
(shell-quote-argument
(cdr ignore))))))
grep-find-ignored-files
- "\\\n -o -name ")
+ (if (file-remote-p dir)
+ "\\\n -o -name "
+ " -o -name "))
" "
(shell-quote-argument ")")
" -prune -o "))))))
- bug#16045: 24.3.50; rgrep can't work, zijianyue, 2013/12/04
- bug#16045: 24.3.50; rgrep can't work,
Eli Zaretskii <=
- bug#16045: 24.3.50; rgrep can't work, Stefan Monnier, 2013/12/04
- bug#16045: 24.3.50; rgrep can't work, Michael Albinus, 2013/12/05
- bug#16045: 24.3.50; rgrep can't work, Eli Zaretskii, 2013/12/05
- bug#16045: 24.3.50; rgrep can't work, Michael Albinus, 2013/12/05
- bug#16045: 24.3.50; rgrep can't work, Eli Zaretskii, 2013/12/05
- bug#16045: 24.3.50; rgrep can't work, Michael Albinus, 2013/12/05
- bug#16045: 24.3.50; rgrep can't work, Stefan Monnier, 2013/12/05
- bug#16045: 24.3.50; rgrep can't work, Michael Albinus, 2013/12/06