On Jul 9, 2000, "Mark E." <[EMAIL PROTECTED]> wrote:
> The below is all true. I can only add the Cygwin uses a leading '//'
> specially (as Posix allows) to encode drivenames for their Unix emulation.
Not any more (i.e., this is now deprecated). The new recommended
Unix-line pathname for c:/
On Sun, Jul 09, 2000 at 08:27:00PM -0400, Mark E. wrote:
>[To: line drastically trimmed]
>
>The below is all true. I can only add the Cygwin uses a leading '//'
>specially (as Posix allows) to encode drivenames for their Unix emulation.
>For example, 'c:/windows' can and will be encoded as '//c/
[To: line drastically trimmed]
The below is all true. I can only add the Cygwin uses a leading '//'
specially (as Posix allows) to encode drivenames for their Unix emulation.
For example, 'c:/windows' can and will be encoded as '//c/windows'. So
preserving the first double-slash would a good t
On Jul 6, 2000, Paul Eggert <[EMAIL PROTECTED]> wrote:
> dir al dn eg
> ..
>1 ..
^
Hey! I had forgotten the `;s,^$,.,' at the end!
-
On Jul 7, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
>> "Paul" == Paul Eggert <[EMAIL PROTECTED]> writes:
Paul> I still think the expr-based solution that I sent to
Paul> autoconf-patches yesterday is the way to go.
> I agree, I like it. I think we should develop the expr-essionism of
From: Akim Demaille <[EMAIL PROTECTED]>
Date: 07 Jul 2000 19:53:35 +0200
There is one question though: how sure are you of the portability of
this \| construct?
I'm pretty sure it was in Unix Version 7. I've been using it for
years. The Hylafax administration scripts use it, and th
> "Paul" == Paul Eggert <[EMAIL PROTECTED]> writes:
Paul> I still think the expr-based solution that I sent to
Paul> autoconf-patches yesterday is the way to go.
I agree, I like it. I think we should develop the expr-essionism of
Autoconf, I'm tired to the classical echo | sed paintings :)
From: Bernard Dautrevaux <[EMAIL PROTECTED]>
Date: Thu, 6 Jul 2000 17:26:48 +0200
under UN*X, '//1' should give '/',
No, under some older flavors of Unix, leading // is a special path
name: it refers to a "super-root" and is used to access other
machines' files. Leading ///, , etc
> Subject: Re: [patch] acgeneral.m4 (dirname usage)
>
>
> >>>>> "Bernard" == Bernard Dautrevaux
> <[EMAIL PROTECTED]> writes:
>
> Bernard> If a path starts by a double '/' (that is for any UNC path,
> Bernard> quite often found
> From: Alexandre Oliva <[EMAIL PROTECTED]>
> Date: 06 Jul 2000 12:30:15 -0300
>
> What does `dirname' do, on those platforms, when given such special
> pathnames?
I'm guessing that you are now talking about `dirname' from Bash or
from Sh-utils, yes?
If so, then it does pretty much the same thi
> From: Akim Demaille <[EMAIL PROTECTED]>
> Date: 06 Jul 2000 17:03:52 +0200
>
> Foreignward compatibility is indeed desired. But are you sure we're
> losing here? Mark, or Eli, what do you think we should do? We're
> trying to implement dirname. `al' is the current CVS solution.
Sorry, I don
From: Akim Demaille <[EMAIL PROTECTED]>
Date: 06 Jul 2000 17:03:52 +0200
echo $dir | sed '/^\/*$/!s,//*$,,;s,[^/]*$,,;s,//*$,/,;/^\/$/!s,/$,,'
It's not portable to use "echo". On some hosts, "echo" mishandles
backslashes and/or names beginning with "-".
It's more portable to use "expr",
On Jul 6, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
> But are you sure we're losing here?
Make that `moving loseward' :-)
> Mark, or Eli, what do you think we should do? We're trying to
> implement dirname. `al' is the current CVS solution.
What does `dirname' do, on those platforms, w
On Jul 6, 2000, Bernard Dautrevaux <[EMAIL PROTECTED]> wrote:
> If a path starts by a double '/' (that is for any UNC path, quite often
> found on DOS boxes), we will strip one of these leading '/'
Nope. We'll only do it if we're given `//server00', for example; this
will become `/'. For `//s
> "Bernard" == Bernard Dautrevaux <[EMAIL PROTECTED]> writes:
Bernard> If a path starts by a double '/' (that is for any UNC path,
Bernard> quite often found on DOS boxes), we will strip one of these
Bernard> leading '/'; thus the directory part of
Bernard> "//server00/home/me/myfile" (which
> -Original Message-
> From: Akim Demaille [mailto:[EMAIL PROTECTED]]
> Sent: Thursday, July 06, 2000 2:05 PM
> To: Lars J. Aas
> Cc: Alexandre Oliva; [EMAIL PROTECTED]; [EMAIL PROTECTED];
> [EMAIL PROTECTED]
> Subject: Re: [patch] acgeneral.m4 (dirname u
Akim> undocumentedward-underscoredward compatible.
Alexandre> These words are getting Awkimward.
Lars> And you are getting awkimward compatible...
:) :) Hi'm rooling ze warld!
On Thu, Jul 06, 2000 at 08:13:03AM -0300, Alexandre Oliva wrote:
: On Jul 6, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
: > undocumentedward-underscoredward compatible.
:
: These words are getting Awkimward.
And you are getting awkimward compatible...
Lars J
On Jul 6, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
> undocumentedward-underscoredward compatible.
These words are getting Awkimward.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD st
> "Lars" == Lars J Aas <[EMAIL PROTECTED]> writes:
Lars> On Thu, Jul 06, 2000 at 10:06:10AM +0200, Akim Demaille wrote: :
Lars> So until we know we have it right for sure, it should not be
Lars> public.
Lars> So what does that mean? Just undocumented in the .texi file, or
Lars> underscoring
On Jul 6, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
>> "Alexandre" == Alexandre Oliva <[EMAIL PROTECTED]> writes:
Alexandre> It works with Solaris' sed, and is significant more
Alexandre> complicated because of some lossage in it.
> I don't think we need this
Indeed, we don't, Solaris
> "Alexandre" == Alexandre Oliva <[EMAIL PROTECTED]> writes:
Alexandre> It works with Solaris' sed, and is significant more
Alexandre> complicated because of some lossage in it.
I don't think we need this, I never heard of anything like this. And
the Solaris system I tried works fine:
Last
On Thu, Jul 06, 2000 at 10:06:10AM +0200, Akim Demaille wrote:
: So until we know we have it right for sure, it should not be public.
So what does that mean? Just undocumented in the .texi file, or
underscoring the macro name?
Lars J
On Jul 6, 2000, Martin Wilck <[EMAIL PROTECTED]> wrote:
> I think "!" after a sed address expression is not portable -
> actually, it seems to work only with GNU sed.
Thanks. How about this, then:
al=`echo $dir | sed
's%^\(..*/[^/][^/]*\)//*$%\1%;s%^\(.[^/]*\)//*$%\1%;s%[^/]*$%%;s%^\(..*/[^/
> Hm, I've used this many times, and have never been trapped. Which sed
> are you thinking about?
Sorry, I was wrong.
Problem was: I had a space between the "!" and the sed command. :(
*That* is only supported by GNU sed.
> "Martin" == Martin Wilck <[EMAIL PROTECTED]> writes:
>> > al=`echo $dir | sed
>> '/^\/*$/!s%//*$%%;s%[^/]*$%%;s%//*$%/%;/^\/$/!s%/$%%'` ^^
>> ^^
Martin>
Martin> ^ I think "!" after a sed address expression is not portable -
Martin> actually, it seems to work only with GNU sed.
Hm,
> > al=`echo $dir | sed '/^\/*$/!s%//*$%%;s%[^/]*$%%;s%//*$%/%;/^\/$/!s%/$%%'`
> ^^ ^^
^
I think "!" after a sed address expression is not portable -
actually, it
On Jul 6, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
> How about using commas instead of %?
That's my preference too. I was just keeping the style of the
original sed script.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer
How about using commas instead of %? Karl Berry pushed me several
times to `,' because it results in something more readable:
sed '/^\/*$/!s,//*$,,;s,[^/]*$,,;s,//*$,/,;/^\/$/!s,/$,,'
sed '/^\/*$/!s%//*$%%;s%[^/]*$%%;s%//*$%/%;/^\/$/!s%/$%%'
and I tend to agree with him.
> "Alexandre" == Alexandre Oliva <[EMAIL PROTECTED]> writes:
Akim> Almost right :)
Alexandre> Almost? Seems right to me.
Oh, yeah! I momentarily thought you were named texi2dvi, sorry :)
I'm approving and applying your patch.
On Jul 6, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
>> "Alexandre" == Alexandre Oliva <[EMAIL PROTECTED]> writes:
Alexandre> Try this one:
>>> al=`echo $dir | sed
>>> '/^\/*$/!s%//*$%%;s%[^/]*$%%;s%//*$%/%;/^\/$/!s%/$%%'`
> Almost right :)
Almost? Seems right to me.
--
Alexandre O
> "Alexandre" == Alexandre Oliva <[EMAIL PROTECTED]> writes:
Alexandre> Try this one:
>> al=`echo $dir | sed
>> '/^\/*$/!s%//*$%%;s%[^/]*$%%;s%//*$%/%;/^\/$/!s%/$%%'`
Almost right :)
~/src/ace % cat /tmp/dirname.sh nostromo 10:59
#! /bin/sh
format="%12s %1
On Jul 6, 2000, Akim Demaille <[EMAIL PROTECTED]> wrote:
> Some more work will be needed.
Try this one:
> al=`echo $dir | sed '/^\/*$/!s%//*$%%;s%[^/]*$%%;s%//*$%/%;/^\/$/!s%/$%%'`
^^ ^^
--
Alexandre Oliva Enjoy Guarana', see
> "Alexandre" == Alexandre Oliva <[EMAIL PROTECTED]> writes:
Alexandre> The patch is fine
Applied.
Alexandre> I'm not 100% sure about its portability, though.
There is nothing risky, IMHO, in there.
Some more work will be needed.
src/ace/patches % cat /tmp/dirname.sh
> "Ralf" == Ralf Corsepius <[EMAIL PROTECTED]> writes:
Ralf> With your patch however, I would expect people to start using
Ralf> AC_SHELL_DIRNAME instead of dirname and therefore I would expect
Ralf> it to hit some time.
Personally I'd prefer that you don't make it public, it is too young.
M
"Lars J. Aas" wrote:
>
> On Wed, Jul 05, 2000 at 05:12:19PM +0200, Ralf Corsepius wrote:
> : "Lars J. Aas" wrote:
> : >
> : > +# AC_SHELL_DIRNAME(PATHNAME)
> : > +# --
> : > +# Remove last slash and trailing text.
> : > +# Not all systems have dirname, so we emulate it wit
On Jul 5, 2000, "Lars J. Aas" <[EMAIL PROTECTED]> wrote:
> +[echo $1 | sed 's%/[[^/][^/]]*$%%'])
This is not strictly correct. It doesn't handle relative pathnames
nor `/'. The patch is fine, since it just preserves current behavior,
and it makes it easier for someone to fix the existing lim
On Wed, Jul 05, 2000 at 05:12:19PM +0200, Ralf Corsepius wrote:
: "Lars J. Aas" wrote:
: >
: > +# AC_SHELL_DIRNAME(PATHNAME)
: > +# --
: > +# Remove last slash and trailing text.
: > +# Not all systems have dirname, so we emulate it with sed.
: > +# This macro must be usab
"Lars J. Aas" wrote:
>
> +# AC_SHELL_DIRNAME(PATHNAME)
> +# --
> +# Remove last slash and trailing text.
> +# Not all systems have dirname, so we emulate it with sed.
> +# This macro must be usable from inside ` `.
> +define([AC_SHELL_DIRNAME],
> +[echo $1 | sed 's%/[[^/][
On Wed, Jul 05, 2000 at 04:10:54PM +0200, Akim Demaille wrote:
: Lars> There's obviously one set of ""s too much here. Should the
: Lars> redundant pair be removed in the macro definition or the macro
: Lars> invocations?
:
: I'd say from the macro.
:
: BTW, you should state in the comment of A
> "Bernard" == Bernard Dautrevaux <[EMAIL PROTECTED]> writes:
Bernard> I'd rather suppress them from the macro invokations; when
Bernard> reusing AC_SHELL_DIRNAME I thus has less to type :-)
Right, but I personally want to have full control over this kind of
things. And just imagine it was
> -Original Message-
> From: Lars J. Aas [mailto:[EMAIL PROTECTED]]
> Sent: Wednesday, July 05, 2000 3:17 PM
> To: [EMAIL PROTECTED]; [EMAIL PROTECTED]
> Subject: Re: [patch] acgeneral.m4 (dirname usage)
>
>
> On Wed, Jul 05, 2000 at 03:05:13PM +0200, Lars J
> "Lars" == Lars J Aas <[EMAIL PROTECTED]> writes:
Lars> There's obviously one set of ""s too much here. Should the
Lars> redundant pair be removed in the macro definition or the macro
Lars> invocations?
I'd say from the macro.
BTW, you should state in the comment of AC_SHELL_DIRNAME that
On Wed, Jul 05, 2000 at 03:05:13PM +0200, Lars J. Aas wrote:
: +define([AC_SHELL_DIRNAME],
: +[echo "$1" | sed 's%/[[^/][^/]]*$%%'])
[...]^ ^
: + ac_dir=`AC_SHELL_DIRNAME("$ac_file")`
[...] ^^
: +ac_dir=`AC_SHELL_DIRNAME("$ac_file")`
[...]
44 matches
Mail list logo
