zsh 4.0.2 (powerpc-apple-darwin1.4)

zsh 4.0.2 (powerpc-apple-darwin1.4)

[Chris Gehlker]

I'm a rank newbie but I got zsh 4.0.2 to build on Darwin1.4/Mac OSX 10.1 by
commenting out two lines in the makefile that configure generated. The two
lines were:

# LIBLDFLAGS      =  -s
# EXELDFLAGS      =  -s

In the compilation section.

Make install didn't seem to to be set up for Darwin so I installed the
binaries and man pages by copping them over the zsh 3.0.8 equivalents which
comes installed with Darwin. This was basically putting the binaries in /bin
and the man pages in /usr/share/man/man1/

I don't know when, if ever, I'll understand configure well enough to
automate the build/install process but I would be glad to make binaries
available for manual installation.
-- 
The secret of life is honesty and fair dealing. If you can fake that,
you've got it made. -Groucho Marx

Re: zsh 4.0.2 (powerpc-apple-darwin1.4)

[Zefram]

Chris Gehlker wrote:
>I'm a rank newbie but I got zsh 4.0.2 to build on Darwin1.4/Mac OSX 10.1 by
>commenting out two lines in the makefile that configure generated. The two
>lines were:
>
># LIBLDFLAGS      =  -s
># EXELDFLAGS      =  -s
>
>In the compilation section.

configure goes to some effort to avoid using the -s flag where that
wouldn't work.  Could you send us the complete output from the configure
script?

>Make install didn't seem to to be set up for Darwin so I installed the
>binaries and man pages by copping them over the zsh 3.0.8 equivalents which
>comes installed with Darwin. This was basically putting the binaries in /bin
>and the man pages in /usr/share/man/man1/

Are these locations the only thing that was wrong in the generated
Makefiles?  If so, you can arrange for "make install" to do what you
want by invoking configure as

	./configure --prefix=/usr --mandir=/usr/share/man --exec-prefix=

-zefram

Re: zsh 4.0.2 (powerpc-apple-darwin1.4)

[Chris Gehlker]

[-- Attachment #1: Type: text/plain, Size: 1936 bytes --]

On 10/7/01 4:50 AM, "Zefram" <zefram@fysh.org> wrote:

> Chris Gehlker wrote:
>> I'm a rank newbie but I got zsh 4.0.2 to build on Darwin1.4/Mac OSX 10.1 by
>> commenting out two lines in the makefile that configure generated. The two
>> lines were:
>> 
>> # LIBLDFLAGS      =  -s
>> # EXELDFLAGS      =  -s
>> 
>> In the compilation section.
> 
> configure goes to some effort to avoid using the -s flag where that
> wouldn't work.  Could you send us the complete output from the configure
> script?

I reran configure and piped the output to configure.out. I'm attaching that
file. I hope this is what you want. If it isn't, just let me know.

I doubt it's relevant, but I'm running Darwin on HFS+ which is case
insensitive.

>> Make install didn't seem to to be set up for Darwin so I installed the
>> binaries and man pages by copping them over the zsh 3.0.8 equivalents which
>> comes installed with Darwin. This was basically putting the binaries in /bin
>> and the man pages in /usr/share/man/man1/
> 
> Are these locations the only thing that was wrong in the generated
> Makefiles?  If so, you can arrange for "make install" to do what you
> want by invoking configure as
> 
> ./configure --prefix=/usr --mandir=/usr/share/man --exec-prefix=
> 
> -zefram

That's really good to know. (And it  was probably documented) I just forged
ahead until I noticed that install was creating a bunch of new directories.
I don't want to give the impression that install didn't work. It just put
things in different directories than where the old zsh files were. Zsh
would work when I pointed terminal at it. Of course man was still showing
the version 3 files.

I'm surprised that Apple is still including zsh 3.0.8 with their brand new
Darwin 1.4

I hope this helps. Please write back if I can do any more testing.
-- 
Tact is the ability to describe others as they see themselves. -Abraham
Lincoln, 16th president of the U.S (1809-1865)


[-- Attachment #2: configure.out --]
[-- Type: text/html, Size: 8659 bytes --]

Re: zsh 4.0.2 (powerpc-apple-darwin1.4)

[Zefram]

>> configure goes to some effort to avoid using the -s flag where that
>> wouldn't work.

But it only makes that effort when it's likely to need it, which is when
it's doing dynamic module loading.  The configure output you provided
showed that it didn't find any dynamic linking facilities that it
could use.

Why did you have to remove the -s flags?  With -s left in, did it fail
to build, or did you get an executable that didn't work?  What error
messages were produced?

>I doubt it's relevant, but I'm running Darwin on HFS+ which is case
>insensitive.

That's interesting but shouldn't affect this.  I don't think we ever
intended zsh to build on a case-insensitive filesystem, but I can't
offhand think of any part of it that will actually have a problem.

-zefram

Re: zsh 4.0.2 (powerpc-apple-darwin1.4)

[Chris Gehlker]

On 10/7/01 4:15 PM, "Zefram" <zefram@fysh.org> wrote:

>>> configure goes to some effort to avoid using the -s flag where that
>>> wouldn't work.
> 
> But it only makes that effort when it's likely to need it, which is when
> it's doing dynamic module loading.  The configure output you provided
> showed that it didn't find any dynamic linking facilities that it
> could use.
> 
> Why did you have to remove the -s flags?  With -s left in, did it fail
> to build, or did you get an executable that didn't work?  What error
> messages were produced?

The compiler barfed on the -s. It produced pages of output without so much
as a warning. Then, as you  can see below, it hit the -s and stopped with an
error message. I'm pasting the last  few lines of make's output below. The
rest looks uninteresting but I'll send it if you want.


cc -c -I.  -DHAVE_CONFIG_H -Wall -Wno-implicit -Wmissing-prototypes -O2 -o
main.o ./main.c
( echo '#!'; cat builtin.syms compat.syms cond.syms exec.syms glob.syms
hashtable.syms hist.syms init.syms input.syms jobs.syms lex.syms
linklist.syms loop.syms math.syms mem.syms module.syms options.syms
params.syms parse.syms pattern.syms prompt.syms signals.syms signames.syms
string.syms subst.syms text.syms utils.syms watch.syms  | sed -n
'/^X/{s/^X//;p;}' | sort -u ) > zsh.export
rm -f zsh
cc  -s  -o zsh main.o  `cat stamp-modobjs`   -lm  -lc
/usr/bin/ld: warning unused multiple definitions of symbol _glob
glob.o definition of _glob in section (__TEXT,__text)
/usr/lib/libm.dylib(glob.o) unused definition of _glob
/usr/bin/ld: can't use -s with input files containg indirect symbols (output
file must contain at least global symbols, for maximum stripping use -x)
make[1]: *** [zsh] Error 1
make: *** [all] Error 1

The compiler docs say it's GNU 2.7.2 with certain extensions to support
Objective C and AltiVec instructions.

As always, feel free to ask if I can provide more help.

-Chris
-- 
Adults are obsolete children.
-Dr. Seuss, humorist, illustrator, and author (1904-1991)

RE: zsh 4.0.2 (powerpc-apple-darwin1.4)

[Borsenkow Andrej]

> cc  -s  -o zsh main.o  `cat stamp-modobjs`   -lm  -lc
> /usr/bin/ld: warning unused multiple definitions of symbol _glob
> glob.o definition of _glob in section (__TEXT,__text)

On the second thought, it is conflict of glob() in Src/glob.c with
system glob(). Such conflict is always bad (and I wonder, is it possible
that mysterious globbing failures that were reported are related).

At first glance it is used just in two places:

  File    Function Line
0 exec.c  execcmd  1760 glob(args, firstnode(args), 0);
1 subst.c globlist  236 glob(list, node, nountok);

so what about just renaming it into zglob or like?

-andrej

Re: zsh 4.0.2 (powerpc-apple-darwin1.4)

[Peter Stephenson]

Borsenkow Andrej wrote:
> > cc  -s  -o zsh main.o  `cat stamp-modobjs`   -lm  -lc
> > /usr/bin/ld: warning unused multiple definitions of symbol _glob
> > glob.o definition of _glob in section (__TEXT,__text)
> 
> On the second thought, it is conflict of glob() in Src/glob.c with
> system glob(). Such conflict is always bad (and I wonder, is it possible
> that mysterious globbing failures that were reported are related).
> 
> At first glance it is used just in two places:
> 
>   File    Function Line
> 0 exec.c  execcmd  1760 glob(args, firstnode(args), 0);
> 1 subst.c globlist  236 glob(list, node, nountok);
> 
> so what about just renaming it into zglob or like?

We've had to avoid a few library functions with generic names, and it's
certainly more future-proof.

I'll put this on both branches.

Is anyone keeping track of 4.0 patches or should I go through the log
myself?  (`Anyone' means Bart, obviously.)

Index: Src/exec.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/exec.c,v
retrieving revision 1.37
diff -u -r1.37 exec.c
--- Src/exec.c	2001/09/24 10:12:51	1.37
+++ Src/exec.c	2001/10/08 08:03:38
@@ -1757,7 +1757,7 @@
 	    if (!(cflags & BINF_NOGLOB))
 		while (!checked && !errflag && args && nonempty(args) &&
 		       has_token((char *) peekfirst(args)))
-		    glob(args, firstnode(args), 0);
+		    zglob(args, firstnode(args), 0);
 	    else if (!unglobbed) {
 		for (node = firstnode(args); node; incnode(node))
 		    untokenize((char *) getdata(node));
Index: Src/glob.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/glob.c,v
retrieving revision 1.21
diff -u -r1.21 glob.c
--- Src/glob.c	2001/07/09 16:05:14	1.21
+++ Src/glob.c	2001/10/08 08:03:38
@@ -920,7 +920,7 @@
 
 /**/
 void
-glob(LinkList list, LinkNode np, int nountok)
+zglob(LinkList list, LinkNode np, int nountok)
 {
     struct qual *qo, *qn, *ql;
     LinkNode node = prevnode(np);
Index: Src/subst.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/subst.c,v
retrieving revision 1.23
diff -u -r1.23 subst.c
--- Src/subst.c	2001/09/25 17:37:55	1.23
+++ Src/subst.c	2001/10/08 08:03:38
@@ -233,7 +233,7 @@
     badcshglob = 0;
     for (node = firstnode(list); !errflag && node; node = next) {
 	next = nextnode(node);
-	glob(list, node, nountok);
+	zglob(list, node, nountok);
     }
     if (badcshglob == 1)
 	zerr("no match", NULL, 0);

-- 
Peter Stephenson <pws@csr.com>                  Software Engineer
CSR Ltd., Science Park, Milton Road,
Cambridge, CB4 0WH, UK                          Tel: +44 (0)1223 392070


**********************************************************************
The information transmitted is intended only for the person or
entity to which it is addressed and may contain confidential 
and/or privileged material. 
Any review, retransmission, dissemination or other use of, or
taking of any action in reliance upon, this information by 
persons or entities other than the intended recipient is 
prohibited.  
If you received this in error, please contact the sender and 
delete the material from any computer.
**********************************************************************

RE: zsh 4.0.2 (powerpc-apple-darwin1.4)

[Borsenkow Andrej]

> I'll put this on both branches.
> 


Are you going to update ChangeLog (no entry in 4.1 at least).

-andrej

Re: zsh 4.0.2 (powerpc-apple-darwin1.4)

[Bart Schaefer]

On Oct 8,  9:07am, Peter Stephenson wrote:
}
} Is anyone keeping track of 4.0 patches or should I go through the log
} myself?  (`Anyone' means Bart, obviously.)

Unfortunately, no, this time I've only been keeping track of the patches
that I put on the branch myself.  Nobody seems to have agreed with (or
at least, not stuck with) my suggestion that ChangeLog entries on the
branch be kept in sync with the original appearance of the patch on the
trunk, which would have made diffing the changes simpler.  I was planning
to just diff the entire trees to attempt to determine what else might
need to be copied, but I've been pretty busy.  (I shouldn't even be out
of bed now ...)

-- 
Bart Schaefer                                 Brass Lantern Enterprises
http://www.well.com/user/barts              http://www.brasslantern.com

Zsh: http://www.zsh.org | PHPerl Project: http://phperl.sourceforge.net   

RE: zsh 4.0.2 (powerpc-apple-darwin1.4)

[Borsenkow Andrej]

> >> configure goes to some effort to avoid using the -s flag where that
> >> wouldn't work.
> 
> But it only makes that effort when it's likely to need it, which is
when
> it's doing dynamic module loading.

Obviously we hit the case when we need it even without dynamic loading.
Now we have to come up with good test for it.

-andrej

Re: zsh 4.0.2 (powerpc-apple-darwin1.4)

[Geoff Wing]

Borsenkow Andrej <Andrej.Borsenkow@mow.siemens.ru> typed:
:> >> configure goes to some effort to avoid using the -s flag where that
:> >> wouldn't work.
:> But it only makes that effort when it's likely to need it, which is when
:> it's doing dynamic module loading.
: Obviously we hit the case when we need it even without dynamic loading.
: Now we have to come up with good test for it.

Maybe this bit of zshconfig.ac needs to be pulled into zsh-4 branch.

-------------------- snip from zshconfig.ac --------------------
if test -n "$auto_ldflags"; then
  case "${enable_zsh_debug}$host_os" in
    yesaix*|yeshpux*|yesnetbsd*|yesopenbsd*) ;;  # "ld -g" is not valid on these systems
    darwin*) LDFLAGS=-Wl,-x ;;
    yes*)    LDFLAGS=-g ;;
    *)       LDFLAGS=-s ;;
  esac
fi
----------------------------------------------------------------

Regards,
-- 
Geoff Wing  <gcw@pobox.com>
rxvt stuff  <gcw@rxvt.org>
zsh stuff   <gcw@zsh.org>

Re: zsh 4.0.2 (powerpc-apple-darwin1.4)

[Geoff Wing]

Geoff Wing <mason@primenet.com.au> typed:
: Maybe this bit of zshconfig.ac needs to be pulled into zsh-4 branch.

Umm, "zsh-4_0-patches" branch.

Re: zsh 4.0.2 (powerpc-apple-darwin1.4)

[Oliver Kiddle]

Geoff Wing wrote:
> 
> Maybe this bit of zshconfig.ac needs to be pulled into zsh-4 branch.
> 
> -------------------- snip from zshconfig.ac --------------------
> if test -n "$auto_ldflags"; then
>   case "${enable_zsh_debug}$host_os" in
>     yesaix*|yeshpux*|yesnetbsd*|yesopenbsd*) ;;  # "ld -g" is not valid on these systems
>     darwin*) LDFLAGS=-Wl,-x ;;
>     yes*)    LDFLAGS=-g ;;
>     *)       LDFLAGS=-s ;;
>   esac
> fi
> ----------------------------------------------------------------

Yes it does. Sorry, that was me just being slow to update it on the
stable branch.

It isn't the case here but with some of the completion functions for
example, the files are still identical on both branches. Is it possible
to commit it as normal on the 4.1 branch and then just update the tag
for the stable branch to the new file instead of copying and committing
on the stable branch.

Oliver

_____________________________________________________________________
This message has been checked for all known viruses by the 
MessageLabs Virus Scanning Service. For further information visit
http://www.messagelabs.com/stats.asp

Re: zsh 4.0.2 (powerpc-apple-darwin1.4)

[Bart Schaefer]

On Oct 8, 10:41am, Oliver Kiddle wrote:
}
} It isn't the case here but with some of the completion functions for
} example, the files are still identical on both branches. Is it possible
} to commit it as normal on the 4.1 branch and then just update the tag
} for the stable branch to the new file instead of copying and committing
} on the stable branch.

Unfortunately, no, CVS doesn't work like that.

If all you do is retag the trunk revision with an ordinary (non-branch)
tag, then checkouts of exactly that tag will pick up exactly that rev of
the file, but checkouts/updates of the branch tag will still get the
previous one.

You could manually delete and recreate the magic "branch tag", but that's
rather dangerous.

-- 
Bart Schaefer                                 Brass Lantern Enterprises
http://www.well.com/user/barts              http://www.brasslantern.com

Zsh: http://www.zsh.org | PHPerl Project: http://phperl.sourceforge.net   

Re: zsh 4.0.2 (powerpc-apple-darwin1.4)

[Geoff Wing]

Oliver Kiddle <okiddle@yahoo.co.uk> typed:
:Geoff Wing wrote:
:> Maybe this bit of zshconfig.ac needs to be pulled into zsh-4 branch.
:> -------------------- snip from zshconfig.ac --------------------
:Yes it does. Sorry, that was me just being slow to update it on the
:stable branch.

Actually, I don't have a problem with that.  Sometimes I think that
some people are too fast to update on the "stable" (bug-fix-type-patches-
only-please) branch without sufficient testing on the development branch.
This is possibly less relevant w.r.t. */functions/   where it may be
possible to consider all possible permutations, still . . .  Of course,
I'm probably guilty of that too:  there was only a three week rethink
period before I pulled up my last set of patches into the stable branch
so I guess I shouldn't speak too loudly.

Regards,
-- 
Geoff Wing  <gcw@pobox.com>
rxvt stuff  <gcw@rxvt.org>
zsh stuff   <gcw@zsh.org>