zsh-3.1.1 on NetBSD patch & problems

zsh-3.1.1 on NetBSD patch & problems

[gwing]

Heyla,
I've set up zsh to use modules on NetBSD
This is on NetBSD 1.2C .  Similar systems like OpenBSD and FreeBSD may
work with the same things.

A couple of things:

1) What's happened to default compctls?
% zsh-3.0.0 -f
% compctl
bindkey -b
builtin -c
export -v
setopt -o
typeset -v
unset -v
unsetopt -o
vared -v
which -c
COMMAND -c
DEFAULT -f
FIRST

% zsh-3.1.1 -f
% compctl
COMMAND -c
DEFAULT -f
FIRST


2) NetBSD make has lots of problems with the Makefile:
 2a) Zle/* stuff doesn't get built the first time you ``make'' in Src.
 2b) if you've made everything, then ``make'' without changing everything,
     everything is rebuilt.
 Of course, GNU make works fine.

3) why is -pedantic set for --enable-debug ?

4) How about `important' changes like
      * Src/options.c: set ALWAYSLASTPROMPT, APPENDHISTORY, AUTOLIST,
	AUTOMENU, AUTOPARAMKEYS, AUTOPARAMSLASH, AUTOREMOVESLASH,
	LISTAMBIGUOUS, LISTTYPES options by default
   going into Etc/NEWS

5) On NetBSD:
 ``ld -x -Bshareable -o foo.so foo..o''
seems to be doing the same thing as
 ``ar r foo.a foo..o bar..o
   ranlib foo.a
   ld -x -Bshareable -Bforcearchive -o foo.so foo.a''
which is the standard way NetBSD does shared libraries. 
The first form is compatible with the current zsh Makefile style.  The second
would require several changes.  Hopefully, the first form will always work 
like the second.

6) I had something else, but I can't think of it now, so here's the patch.


*** configure.in	1997/01/25 19:00:07	3.1.1.11
--- configure.in	1997/02/17 05:53:09
***************
*** 208,214 ****
  fi
  if test -n "$auto_ldflags"; then
    if test "${enable_zsh_debug}" = yes; then
!     LDFLAGS=-g
    fi
  fi
  
--- 208,217 ----
  fi
  if test -n "$auto_ldflags"; then
    if test "${enable_zsh_debug}" = yes; then
!     case "$host_os" in
!       netbsd*)    LDFLAGS="" ;;
!       *)          LDFLAGS=-g ;;
!     esac
    fi
  fi
  
***************
*** 821,826 ****
--- 824,830 ----
      solaris*)     DLLDFLAGS="${DLLDFLAGS=-G}" ;;
      sunos*)       DLLDFLAGS="${DLLDFLAGS=-assert nodefinitions}" ;;
      sysv4*|esix*) DLLDFLAGS="${DLLDFLAGS=-G $ldflags}" ;;
+     netbsd*)      DLLDFLAGS="${DLLDFLAGS=-x -Bshareable}" ;;
    esac
    case "$host_os" in
      hpux*)  EXTRA_LDFLAGS="${EXTRA_LDFLAGS=-Wl,-E}" ;;


-- 
Geoff Wing [gwing@primenet.com.au]   Technical Manager
  Phone    : +61-3-9818 2977         PrimeNet - Internet Consultancy
  Facsimile: +61-3-9819 3788         Web : <URL:http://www.primenet.com.au/>
  Mobile   : 0412 162 441

Re: zsh-3.1.1 on NetBSD patch & problems

[Zefram]

gwing@primenet.com.au wrote:
>1) What's happened to default compctls?

Who cares?  They were a bad idea anyway.

>3) why is -pedantic set for --enable-debug ?

So we don't accidentally introduce non-conformant code that gcc accepts.

-zefram

Re: zsh-3.1.1 on NetBSD patch & problems

[Vinnie Shelton]

zefram@dcs.warwick.ac.uk said:
> gwing@primenet.com.au wrote:
>1) What's happened to default compctls?
> Who cares?  They were a bad idea anyway. 
  ^^^^^^^^^^
Is that some kind of joke?  Does the phrase "backward compatibility" mean
anything to you?  Can we all agree a permanent moratorium on breaking
old code/scripts/login environments?

vin

Re: zsh-3.1.1 on NetBSD patch & problems

[Hrvoje Niksic]

Vinnie Shelton  <shelton@icd.teradyne.com> writes:

> zefram@dcs.warwick.ac.uk said:
> > gwing@primenet.com.au wrote:
> >1) What's happened to default compctls?
> > Who cares?  They were a bad idea anyway. 
>   ^^^^^^^^^^
> Is that some kind of joke?  Does the phrase "backward compatibility" mean
> anything to you?  Can we all agree a permanent moratorium on breaking
> old code/scripts/login environments?

Well, this change certainly breaks code and scripts, but I do agree
that the old defaults were very convenient, and *not* a bad idea.

When we are at work, why don't we kick out `bye' (we have exit, after
all), history (use `fc -l'), `r' (use `fc -r'), etc.  Bah!

-- 
Hrvoje Niksic <hniksic@srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
The IDIOT.  Usually a cretin, morphodite, or old COBOL programmer
selected to be the system administrator by a committee of cretins,
morphodites, and old COBOL programmers.

Re: zsh-3.1.1 on NetBSD patch & problems

[Zefram]

Vinnie Shelton wrote:
>Is that some kind of joke?  Does the phrase "backward compatibility" mean
>anything to you?  Can we all agree a permanent moratorium on breaking
>old code/scripts/login environments?

compctl is purely an interactive feature, so only the very freakiest
scripts could possibly be affected by this.  Also, all the commands
that had default compctls also have compctls set up in the example
compctl list distributed with zsh -- in some cases, these are actually
better compctls than the default.

I think that since backward compatibility is not a significant issue in
this case, it is wise to use the cleaner behaviour, of no default
compctls.

-zefram

Re: zsh-3.1.1 on NetBSD patch & problems

[Richard Coleman]

> >Is that some kind of joke?  Does the phrase "backward compatibility" mean
> >anything to you?  Can we all agree a permanent moratorium on breaking
> >old code/scripts/login environments?
> 
> compctl is purely an interactive feature, so only the very freakiest
> scripts could possibly be affected by this.  Also, all the commands
> that had default compctls also have compctls set up in the example
> compctl list distributed with zsh -- in some cases, these are actually
> better compctls than the default.
> 
> I think that since backward compatibility is not a significant issue in
> this case, it is wise to use the cleaner behaviour, of no default
> compctls.

I don't think this is a matter of backward compatibility, but
more of a matter of creating a decent environment for new users.

I realize most people on this list probably have a large list
of compctl's, and override all the default ones.  But this
matter concerns new users rather than power users.  What do
we save by removing the default compctl's?  Maybe 100 bytes, and
a few micro-seconds of startup time.

Zsh should (by default) start in a usable state.  I hate tools that
require extensive customization before they are usable.  Let's not
take zsh down that path.

rc

Re: zsh-3.1.1 on NetBSD patch & problems

[Vinnie Shelton]

RC wrote:
> I don't think this is a matter of backward compatibility, but
> more of a matter of creating a decent environment for new users.

Well, I hate to disagree with some who's trying to find common ground with 
me, but I think it goes beyond the issue of creating a usable environment 
for new users.  Why should I have to add new cruft to my startup scripts 
everytime a new release comes out?  Why do we have to abandon things which 
are already working?  Why did I have to start using this function for pushd:

pushd () {
        setopt localoptions globsubst
        unsetopt ksharrays
        case $1 in
                +*) setopt pushdignoredups
                        builtin pushd ${${=$(dirs)}[$1+1]};;
                -*) setopt pushdignoredups
                        builtin pushd ${${=$(dirs)}[$1-1]};;
                *) builtin pushd $*;;
 
        esac
}

[Thank you Peter and Bart.]   Because people don't care about backward 
compatibility.  As Peter so aptly put it in article 2230:
    It's yet another case of running very hard to stay in the same place.

So, Zefram, in answer to your rhetorical question: "I care".

> I realize most people on this list probably have a large list
> of compctl's, and override all the default ones.  But this
> matter concerns new users rather than power users.  What do
> we save by removing the default compctl's?  Maybe 100 bytes, and
> a few micro-seconds of startup time.
> 
> Zsh should (by default) start in a usable state.  I hate tools that
> require extensive customization before they are usable.  Let's not
> take zsh down that path.

RC, on this we do agree.

I'll get down off my soapbox now.

vin

Re: zsh-3.1.1 on NetBSD patch & problems

[gwing]

Zefram wrote:
:gwing@primenet.com.au wrote:
:>1) What's happened to default compctls?
:Who cares?  They were a bad idea anyway.

How did I find that they weren't there any more?  How about:
% zsh -f
% setopt <TAB>
<... file list ...>

I'm not of the opinion that they're a bad idea. 

:>3) why is -pedantic set for --enable-debug ?
:So we don't accidentally introduce non-conformant code that gcc accepts.

What sort of things?  Non-conforming to what?  Pedantic checks ANSI C.
Things like variable length arrays are (configure) supported.
Quad types aren't supported in printf("%q...

Anyway, my gcc man page says:
``without this option, certain GNU extensions and traditional C features
are supported as well.  With this option, they are rejected.  There is no
reason to use this option; it exists only to satisfy pedants.''

Oh well, if it makes you happy...
-- 
Geoff Wing [gwing@primenet.com.au]   Technical Manager
  Phone    : +61-3-9818 2977         PrimeNet - Internet Consultancy
  Facsimile: +61-3-9819 3788         Web : <URL:http://www.primenet.com.au/>
  Mobile   : 0412 162 441

Re: zsh-3.1.1 on NetBSD patch & problems

[Zoltan Hidvegi]

gwing@primenet.com.au wrote:
> Anyway, my gcc man page says:
> ``without this option, certain GNU extensions and traditional C features
> are supported as well.  With this option, they are rejected.  There is no
> reason to use this option; it exists only to satisfy pedants.''
> 
> Oh well, if it makes you happy...

Withous -pedantic, gcc does not warn about missing (char *) <->
(unsigned char *) casts.  Some compilers reject the code  and some
give you warnings on that.

And several people submitted patches which later had to be changed to add
these missing casts.

Re: zsh-3.1.1 on NetBSD patch & problems

[Zefram]

gwing@primenet.com.au wrote:
>:>3) why is -pedantic set for --enable-debug ?
>:So we don't accidentally introduce non-conformant code that gcc accepts.
>
>What sort of things?  Non-conforming to what?  Pedantic checks ANSI C.

Yes.  zsh doesn't require gcc to build; it can be compiled by any ANSI
compiler.  It's very easy to accidentally write illegal code that gcc
won't even warn about normally.

>Things like variable length arrays are (configure) supported.

Yes, we use variable length arrays if they are available.  That's a
special case.

>Quad types aren't supported in printf("%q...

We use quad types where they are available and needed.

>Anyway, my gcc man page says:
>``without this option, certain GNU extensions and traditional C features
>are supported as well.  With this option, they are rejected.  There is no
>reason to use this option; it exists only to satisfy pedants.''

Do you really believe all that GNU propaganda?  They seem to be
encouraging the creation of an "all the world's a GNU" syndrome.  The
truth is that there is no reason to use -pedantic if one only wants to
compile some code with gcc, and is not concerned about any other
compilers.  We *do* care about other compilers, so -pedantic is very
useful.

-zefram