Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-20 Thread Peter Eisentraut
Akim Demaille writes: > Do you people think we should continue explaining how to do things > without Automake? I think yes, because otherwise you decrease the overall usability of your product. -- Peter Eisentraut [EMAIL PROTECTED] http://yi.org/peter-e/

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-20 Thread Jim Meyering
Akim Demaille <[EMAIL PROTECTED]> wrote: | > "Peter" == Peter Eisentraut <[EMAIL PROTECTED]> writes: | | Peter> Alexandre Oliva writes: | >> Anyway, automake uses `echo timestamp > file', instead of touch, | >> and I've always thought the reason was to avoid having an empty | >> file. | | Pet

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-20 Thread Akim Demaille
| Akim Demaille writes: | > Peter> The reason (which is documented somewhere in the auto* manuals) | > | > I don't know where. | | autoconf.texi | | @node Automatic Remaking, , Build Directories, Makefile Substitutions | @subsection Automatic Remaking | | [snip] | | The @file{stamp-} files a

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-20 Thread Peter Eisentraut
Akim Demaille writes: > Peter> The reason (which is documented somewhere in the auto* manuals) > > I don't know where. autoconf.texi @node Automatic Remaking, , Build Directories, Makefile Substitutions @subsection Automatic Remaking [snip] The @file{stamp-} files are necessary because the t

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-20 Thread Akim Demaille
> "Peter" == Peter Eisentraut <[EMAIL PROTECTED]> writes: Peter> Alexandre Oliva writes: >> Anyway, automake uses `echo timestamp > file', instead of touch, >> and I've always thought the reason was to avoid having an empty >> file. Peter> The reason (which is documented somewhere in the aut

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-20 Thread Peter Eisentraut
Alexandre Oliva writes: > Anyway, automake uses `echo timestamp > file', instead of touch, and > I've always thought the reason was to avoid having an empty file. The reason (which is documented somewhere in the auto* manuals) is that on some (BSD?) systems a 'touch' of an empty file doesn't upd

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-20 Thread Akim Demaille
> "Alexandre" == Alexandre Oliva <[EMAIL PROTECTED]> writes: Alexandre> I wouldn't mind having the testsuite failing in case an Alexandre> empty file can't be created, but I'd be very concerned if Alexandre> autoconf or configure scripts generated by it depended on Alexandre> this ability, fo

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-20 Thread Alexandre Oliva
On Mar 20, 2001, Akim Demaille <[EMAIL PROTECTED]> wrote: >> "Alexandre" == Alexandre Oliva <[EMAIL PROTECTED]> writes: Alexandre> IIRC, there are filesystems that don't support zero-sized Alexandre> files. > ISTR that for a time we believed AIX was one such FS, but it has never > been demo

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-20 Thread Akim Demaille
> "Alexandre" == Alexandre Oliva <[EMAIL PROTECTED]> writes: >> Anyway, we really need to know how to create portably empty files, >> we use it at other places IIRC. Alexandre> Do we? A quick grep didn't show any such place. Alexandre> IIRC, there are filesystems that don't support zero-si

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-19 Thread Alexandre Oliva
On Mar 19, 2001, Akim Demaille <[EMAIL PROTECTED]> wrote: > Anyway, we really need to know how to create portably empty files, we > use it at other places IIRC. Do we? IIRC, there are filesystems that don't support zero-sized files. So we shouldn't depend on it, and we shouldn't depend on bein

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-19 Thread Bob Proulx
I missed something along the way and could not find it in the mail archive. Could someone recap why the following is not a good way to create empty files? It should be a defined behavior for all POSIX shells. I understand that the subject line implies that this does not work under Ultrix. I am

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-19 Thread Harlan Stenn
I got much farther with /bin/sh5: ... configure: creating ./config.status sed: Unexpected End-of-file /bin/sh5: ^Jsrcdir=..^JINSTALL=/usr/bin/install: not found configure: WARNING: you should use --build, --host, --target configure: WARNING: invalid host type: --enable-parse-clocks config

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-19 Thread Eric Siegerman
On Mon, Mar 19, 2001 at 11:23:10AM +0100, Akim Demaille wrote: > Anyway, we really need to know how to create portably empty files, we > use it at other places IIRC. echo a | grep b >empty-file Yucky, but it works :-) I was going to suggest echo "" | grep . >empty-file but if grep ever

RE: autoconf test ': >emtpy' problem under Ultrix

2001-03-19 Thread Tim Van Holder
> | Jim> Not everyone has a /dev/null. > | > | Huh??? What for instance? > > Er... Well, I guess you've caught me :-) > I can't name one offhand. And even if I could, it probably > wouldn't be worth making a change. DJGPP doesn't (didn't) have one; its bash does support it, so the current redi

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-19 Thread Akim Demaille
> "Nicolas" == Nicolas Joly <[EMAIL PROTECTED]> writes: Nicolas> Ok, but the `diff -u' test is broken (see below). Thanks Nicolas!

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-19 Thread Nicolas Joly
On Mon, Mar 19, 2001 at 11:23:10AM +0100, Akim Demaille wrote: > > "Jim" == Jim Meyering <[EMAIL PROTECTED]> writes: > > Jim> I vaguely recall that we couldn't depend on touch. However, it > Jim> *is* in the list of programs mentioned in standards.info's > Jim> `Utilities in Makefiles' secti

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-19 Thread Akim Demaille
> "Paul" == Paul Eggert <[EMAIL PROTECTED]> writes: Paul> Sorry, I lost context. Why can't we ask people to invoke Paul> 'CONFIG_SHELL=/bin/sh5 /bin/sh5 configure' on Ultrix? If Paul> there's only a few dozen users, we can just inform them of this Paul> by hand. I was considering the optio

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-19 Thread Paul Eggert
> From: Jim Meyering <[EMAIL PROTECTED]> > Date: 19 Mar 2001 10:52:15 +0100 > > I'm beginning to think that Ultrix's /bin/sh is so bad that polluting, > er... changing autoconf to accommodate is may not be worthwhile. > Do any of you have an idea of how big the Ultrix-using community is? I don't

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-19 Thread Akim Demaille
> "Jim" == Jim Meyering <[EMAIL PROTECTED]> writes: Jim> | Jim> Not everyone has a /dev/null. Jim> | Jim> | Huh??? What for instance? Jim> Er... Well, I guess you've caught me :-) I can't name one Jim> offhand. And even if I could, it probably wouldn't be worth Jim> making a change. Well

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-19 Thread Jim Meyering
Akim Demaille <[EMAIL PROTECTED]> wrote: | > "Jim" == Jim Meyering <[EMAIL PROTECTED]> writes: | | Jim> I vaguely recall that we couldn't depend on touch. However, it | Jim> *is* in the list of programs mentioned in standards.info's | Jim> `Utilities in Makefiles' section. | | Anyway, we rea

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-19 Thread Akim Demaille
> "Jim" == Jim Meyering <[EMAIL PROTECTED]> writes: Jim> I vaguely recall that we couldn't depend on touch. However, it Jim> *is* in the list of programs mentioned in standards.info's Jim> `Utilities in Makefiles' section. Anyway, we really need to know how to create portably empty files, w

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-19 Thread Jim Meyering
Harlan Stenn <[EMAIL PROTECTED]> wrote: | I haven't seen any response to this report yet. I'm beginning to think that Ultrix's /bin/sh is so bad that polluting, er... changing autoconf to accommodate is may not be worthwhile. Do any of you have an idea of how big the Ultrix-using community is? |

Re: autoconf test ': >emtpy' problem under Ultrix

2001-03-19 Thread Akim Demaille
> "Harlan" == Harlan Stenn <[EMAIL PROTECTED]> writes: Harlan> I haven't seen any response to this report yet. Is there a Harlan> good reason the "empty" file can't be done using: Harlan> rm -f empty && touch empty (just to be thorough) Up to now we avoided touch. Harlan> or Harlan> cp