Re: [1.7] Backslash incorrectly triggers DOS style path warning

2009-04-22 Thread Corinna Vinschen
On Apr 22 14:50, Christopher Faylor wrote: > Hmm. This would mean that you can't debug glob/noglob using strace > though... I guess I'll check in a fix for the common case where you > don't want to be surprised by having your quoted wildcards expanded. > The other case is probably only useful for

Re: [1.7] Backslash incorrectly triggers DOS style path warning

2009-04-22 Thread Christopher Faylor
On Wed, Apr 22, 2009 at 02:31:02PM -0400, Christopher Faylor wrote: >On Wed, Apr 22, 2009 at 07:08:26PM +0200, Corinna Vinschen wrote: >>On Apr 22 12:59, Christopher Faylor wrote: >>> On Wed, Apr 22, 2009 at 06:49:44AM -0600, Eric Blake wrote: >>> >> In other words, strace is mistakenly performing

Re: [1.7] Backslash incorrectly triggers DOS style path warning

2009-04-22 Thread Christopher Faylor
On Wed, Apr 22, 2009 at 07:08:26PM +0200, Corinna Vinschen wrote: >On Apr 22 12:59, Christopher Faylor wrote: >> On Wed, Apr 22, 2009 at 06:49:44AM -0600, Eric Blake wrote: >> >> In other words, strace is mistakenly performing glob expansion on the >> >> subsidiary arguments to the program being tr

Re: [1.7] Backslash incorrectly triggers DOS style path warning

2009-04-22 Thread Corinna Vinschen
On Apr 22 12:59, Christopher Faylor wrote: > On Wed, Apr 22, 2009 at 06:49:44AM -0600, Eric Blake wrote: > >> In other words, strace is mistakenly performing glob expansion on the > >> subsidiary arguments to the program being traced, when we really wanted to > >> trace ls with a literal argument o

Re: [1.7] Backslash incorrectly triggers DOS style path warning

2009-04-22 Thread Christopher Faylor
On Wed, Apr 22, 2009 at 06:49:44AM -0600, Eric Blake wrote: >> In other words, strace is mistakenly performing glob expansion on the >> subsidiary arguments to the program being traced, when we really wanted to >> trace ls with a literal argument of "*". > >And this strace limitation (bug?) was jus

Re: [1.7] Backslash incorrectly triggers DOS style path warning

2009-04-22 Thread Eric Blake
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 According to Eric Blake on 4/22/2009 6:24 AM: > COMPREPLY=($( awk 'BEGIN {FS=","} > /^\s*[^|\#]/ {for (i=1; i<=2; ++i) { \ > gsub(" .*$", "", $i); \ >

Re: [1.7] Backslash incorrectly triggers DOS style path warning

2009-04-22 Thread Eric Blake
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 According to Ken Brown on 4/22/2009 5:58 AM: >> It is maybe getting globbed on the command-line because not >> protected by >> quoting and it contains pattern match chars? > > The text in question is in the definition of _known_hosts() in > /etc/bas

Re: [1.7] Backslash incorrectly triggers DOS style path warning

2009-04-22 Thread Ken Brown
On 4/21/2009 10:06 PM, Dave Korn wrote: Mark J. Reed wrote: On Tue, Apr 21, 2009 at 7:50 PM, Christopher Faylor wrote: Cygwin isn't scanning command lines looking for backslashes to scold you about. Glad to hear it The line in question was somehow used as an argument to open() or stat() or

Re: [1.7] Backslash incorrectly triggers DOS style path warning

2009-04-21 Thread Dave Korn
Mark J. Reed wrote: > On Tue, Apr 21, 2009 at 7:50 PM, Christopher Faylor wrote: >> Cygwin isn't scanning command lines looking for backslashes to scold you >> about. > > Glad to hear it > >> The line in question was somehow used as an argument to open() >> or stat() or access() or some other fu

Re: [1.7] Backslash incorrectly triggers DOS style path warning

2009-04-21 Thread Mark J. Reed
On Tue, Apr 21, 2009 at 7:50 PM, Christopher Faylor wrote: > Cygwin isn't scanning command lines looking for backslashes to scold you > about. Glad to hear it >  The line in question was somehow used as an argument to open() > or stat() or access() or some other function which takes a filename >

Re: [1.7] Backslash incorrectly triggers DOS style path warning

2009-04-21 Thread Christopher Faylor
On Tue, Apr 21, 2009 at 07:41:06PM -0400, Mark J. Reed wrote: >On Monday, April 20, 2009, Christopher Faylor wrote: >> It isn't really erroneous. ??Something (awk) is trying to open a filename >> with that includes backslashes, so cygwin1.dll thinks that it is trying >> to open a DOS path. > >Excep

Re: [1.7] Backslash incorrectly triggers DOS style path warning

2009-04-21 Thread Mark J. Reed
On Monday, April 20, 2009, Christopher Faylor wrote: > It isn't really erroneous.  Something (awk) is trying to open a filename > with that includes backslashes, so cygwin1.dll thinks that it is trying > to open a DOS path. Except it looks like, from what's pasted above, that it's actually a non-f

Re: [1.7] Backslash incorrectly triggers DOS style path warning

2009-04-20 Thread Christopher Faylor
On Mon, Apr 20, 2009 at 09:37:01AM -0400, Ken Brown wrote: > On 4/20/2009 9:29 AM, Ken Brown wrote: >> I just installed the bash-completion package. To see how it worked, I >> typed >> $ ssh >> This yielded: >> $ ssh cygwin warning: >> MS-DOS style path detected: BEGIN {FS=","} >>

Re: [1.7] Backslash incorrectly triggers DOS style path warning

2009-04-20 Thread Ken Brown
On 4/20/2009 9:29 AM, Ken Brown wrote: I just installed the bash-completion package. To see how it worked, I typed $ ssh This yielded: $ ssh cygwin warning: MS-DOS style path detected: BEGIN {FS=","} /^\s*[^|\#]/ {for (i=1; i<=2; ++i) { \

[1.7] Backslash incorrectly triggers DOS style path warning

2009-04-20 Thread Ken Brown
I just installed the bash-completion package. To see how it worked, I typed $ ssh This yielded: $ ssh cygwin warning: MS-DOS style path detected: BEGIN {FS=","} /^\s*[^|\#]/ {for (i=1; i<=2; ++i) { \ gsub(" .*$", "", $i)