On the issue of Solaris: I checked with Don Cragun
(one of Sun's standards guys). I can provide chapter
and verse, but in short he says that all Solaris 2.0
and later releases have both /usr/bin/sh and /usr/bin/ksh,
with ksh being the posix shell. Here's an excerpt:
> I believe all Solaris 2.0 and later releases have provided both
>/usr/bin/sh and /usr/bin/ksh. (/bin is a symbolic link pointing to
>/usr/bin.) The issue with making /usr/bin/sh be a POSIX conforming
>shell is that doing so would break ancient customer shell scripts that
>depend on features that POSIX changed 8 years ago. Our development
>rules don't allow us to make incompatible changes like that, except in
>a major release. (Our last major release was Solaris 2.0 when we moved
>from a default BSD source base environment to a default SVR4 source
>base environment. It isn't obvious to me when, or if, we will have
>another major release.) (DT: then some very cogent reasoning about
the business decisions concerning customer expectation of
compatability with old scripts.)
I think we can infer two things: 1) that waiting for (at
the very least) Solaris to have /bin/sh be POSIX conformant may
be a long wait, and 2) that we needn't wait, we just need to be
able to point at a shell other than /bin/sh. (This also addresses
the issue that started all this, functions.)
I can't say in detail about all systems, but most have a similar
problem to some degree or other. (I actually can find some more
details if anyone cares... ask and I'll do the digging).
I think we can infer that if we say it must be the LCD of all
/bin/sh implementations (and recognizing that there's no requirement
that /bin/sh even exist in the POSIX standards) we'll continue
fighting problems that how have modern, easy, solutions.
(Don and I are both involved in the POSIX revision, and the
same topic just came up (not accidentally :-) ) and yes, /bin/sh
is NOT required to exist; just something called sh.)
The other alternative is to deal with multiple shell locations
to get a more modern shell; the issue here is are there enough
systems we care about that don't have a modern shell (I suspect
not). It's a matter of finding it. I'm working that issue thru
the revision; we'll see what happens.
Donn
> -----Original Message-----
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
> Sent: Wednesday, May 10, 2000 10:16 AM
> To: Donn Terry
> Cc: 'Felix Lee'; 'Robert Lipe'; 'Russ Allbery '; '[EMAIL PROTECTED] ';
> '[EMAIL PROTECTED] '
> Subject: Re: rfc: new libgcc build mechanism
>
>
> %% Donn Terry <[EMAIL PROTECTED]> writes:
>
> dt> Please take a look at the Makefile in that area; the
> dt> current code is using nested `` to create further command
> dt> lines (not data files). I'll admit I haven't tried, but
> dt> I'm not at all sure it can be done because the *creation*
> dt> of the string occurs in a different directory than the
> dt> *use* of it, and the purpose of the substitution is to
> dt> deal with that problem! Here's what I currently use
> dt> locally (with $()). If you can figure out a decent
> dt> (and generally acceptable) solution to the problem that
> dt> works with ``, I'll be more than happy to see it.
>
> dt> PREPEND_DOTDOT_TO_RELATIVE_PATHS = sed \
> dt> -e 's|^ *[^ /][^ /]*/|%&|' \
> dt> -e 's| -B| -B%|g' \
> dt> -e 's|% *[^- /]|%&|g' \
> dt> -e 's|%% *|../|g' \
> dt> -e 's|%||g'
>
> I'm a little concerned that the & token isn't supported by
> all versions
> of sed, either, but I couldn't find any examples... anyone
> have thoughts
> on the portability of that?
>
> dt> "CC=$$(echo $$(case '$(CC)' in (stage*) echo
> '$(CC)' | sed -e \
> dt> 's|stage|../stage|g';; (*) echo '$(CC)';; esac) | \
> dt> $(PREPEND_DOTDOT_TO_RELATIVE_PATHS) )"
>
> Hmm. Why can't you do this with just one ``? There's no need for the
> outer one and the outer echo at all as far as I can see, just pipe the
> results of the case statement directly into the sed script, like this:
>
> "CC=`case '$(CC)' in stage*) echo '$(CC)' | sed -e
> 's|stage|../stage|g';; \
> *) echo '$(CC)';; \
> esac | $(PREPEND_DOTDOT_TO_RELATIVE_PATHS)`"
>
> ?
>
> --
> --------------------------------------------------------------
> -----------------
> Paul D. Smith <[EMAIL PROTECTED]> Network
> Management Development
> "Please remain calm...I may be mad, but I am a
> professional." --Mad Scientist
> --------------------------------------------------------------
> -----------------
> These are my opinions---Nortel Networks takes no
> responsibility for them.
>
