For the MACHINES file: OpenStep/NeXTStep 4.2 and zsh-3.1.6-dev-19

For the MACHINES file: OpenStep/NeXTStep 4.2 and zsh-3.1.6-dev-19

[Brian Boonstra]

Hi Folks

	After a long period of inability to build development versions on  
OpenStep 4.2, I can now say that zsh-3.1.6-dev-19 builds and works.  Here are  
the details:
	
	
WORKING
-------
Architecture:  Apple/NeXT OpenStep 4.2 for i386

gcc version:	2.8.1
gawk version:	2.15 patchlevel 6


DOES NOT WORK
-------------

Apple/NeXT-supplied cc (version 2.7.2.1)
Apple/NeXT-supplied awk (version unknown)


Since it is unusual to have a compiler other than the usual cc with these  
machines, gcc should be mentioned in the special instructions.  I have  
uploaded a binary to the common (for OpenStep users) site at ftp.peak.org.



	Regards,
	
		Brian

Re: For the MACHINES file: OpenStep/NeXTStep 4.2 and zsh-3.1.6-dev-19

[Zefram]

Brian Boonstra wrote:
>DOES NOT WORK
>-------------
>
>Apple/NeXT-supplied cc (version 2.7.2.1)

Is this actually a gcc?  (gcc 2.7.2.1 is a relatively recent version.)
In what way does it fail?

>Apple/NeXT-supplied awk (version unknown)

This is interesting.  Again, in what way does it fail?  The awk scripts
we use are intended to be as widely portable as possible.

-zefram

Re: For the MACHINES file: OpenStep/NeXTStep 4.2 and zsh-3.1.6-dev-19

[Brian Boonstra]

Zefram wrote:
> >Apple/NeXT-supplied cc (version 2.7.2.1)
>
> Is this actually a gcc?  (gcc 2.7.2.1 is a relatively recent version.)
> In what way does it fail?

	Well, yes and no.  NeXT modified gcc with some Objective C goodies  
and shipped it with their development environment.  I always assumed they  
maintained approximate version number compatibility, but I don't know for  
sure.

	The failure is a memory problem of some kind.  Issue a command, like  
"ls", and you get the output and then a hang.
	
	I spent a fair bit of time (with Bart's help) trying to diagnose it  
with 3.1.6-bart7, but I'm no expert and I ultimately got nowhere -- you can  
read about it in the mailing list archives from Oct 17-25.  It was made  
especially hard by the fact that turning on the memory debugging killed the  
compile completely.
	
	One interesting question I can't answer is whether Apple OS X  
machines (which are closely related) will successfully compile 3.1.6.


> >Apple/NeXT-supplied awk (version unknown)
>
> This is interesting.  Again, in what way does it fail?  The awk scripts
> we use are intended to be as widely portable as possible.


I too was surprised they did not work, but I did not try debugging them.   
Here is what happens (after a successful ./configure):

...
creating Src/Zle/Makefile
awk -f ./signames1.awk /usr/include/bsd/sys/signal.h >sigtmp.c
cc -E -traditional-cpp sigtmp.c >sigtmp.out
awk -f ./signames2.awk sigtmp.out > signames.c
rm -f sigtmp.c sigtmp.out
grep 'define.*SIGCOUNT' signames.c > sigcount.h
Updated `zsh.mdh'.
echo 'timestamp for zsh.mdh against zsh.mdd' > zsh.mdhs
awk -f ../Src/makepro.awk builtin.c Src > builtin.syms
awk: syntax error near line 15
awk: illegal statement near line 15
...
<Lots more syntax errors deleted>
...
awk: bailing out near line 107
awk: newline in regular expression near line 107
*** Exit 2
*** `builtin.syms' removed
Stop.





		Best Regards,
		
			Brian

Re: For the MACHINES file: OpenStep/NeXTStep 4.2 and zsh-3.1.6-dev-19

[Zefram]

Brian Boonstra wrote:
>awk -f ../Src/makepro.awk builtin.c Src > builtin.syms
>awk: syntax error near line 15
>awk: illegal statement near line 15

gsub.  That's interesting.  What's the correct syntax for a global
substitution in this awk?

-zefram

Re: For the MACHINES file: OpenStep/NeXTStep 4.2 and zsh-3.1.6-dev-19

[Brian Boonstra]

Zefram wrote:
> gsub.  That's interesting.  What's the correct syntax for a global
> substitution in this awk?

Well, this awk doesn't have a lot of keywords.  The closest relevant thing  
from the manpage is:

     Files are read in order; if there are no files, the standard
     input is read.  The file name `-' means the standard input.
     Each line is matched against the pattern portion of every
     pattern-action statement; the associated action is performed
     for each matched pattern.

And here is what I have taken to be the complete list of functions:

     The print statement prints its arguments on the standard
     output (or on a file if >file is present), separated by the
     current output field separator, and terminated by the output
     record separator.  The printf statement formats its expres-
     sion list according to the format (see printf(3S)).

     The built-in function length returns the length of its argu-
     ment taken as a string, or of the whole line if no argument.
     There are also built-in functions exp, log, sqrt, and int.
     The last truncates its argument to an integer.
     substr(s, m, n) returns the n-character substring of s that
     begins at position m.  The function
     sprintf(fmt, expr, expr, ...) formats the expressions
     according to the printf(3S) format given by fmt and returns
     the resulting string.


I assume it matches the spec in the reference below:

    SEE ALSO
         lex(1), sed(1)
         A. V. Aho, B. W. Kernighan, P. J. Weinberger, Awk - a pat-
         tern scanning and processing language
	
	
	
	 		Regards,
			
				Brian