MacOS X patches

MacOS X patches

[Pedro Giffuni]

Hello;

FWIW, I noticed that Apple has some patches for zsh-5.0.5 available here:

http://www.opensource.apple.com/source/zsh/zsh-61/patches/

The changes are rather minor, but perhaps there’s something you may want to take. 

Regards,

Pedro.

ps. IANAL but I don’t see nay license disclaimer that could prohibit taking them.

Re: MacOS X patches

[Mikael Magnusson]

On Sat, Nov 1, 2014 at 5:48 AM, Pedro Giffuni <pfg@freebsd.org> wrote:
> Hello;
>
> FWIW, I noticed that Apple has some patches for zsh-5.0.5 available here:
>
> http://www.opensource.apple.com/source/zsh/zsh-61/patches/
>
> The changes are rather minor, but perhaps there’s something you may want to take.
>
> Regards,
>
> Pedro.
>
> ps. IANAL but I don’t see nay license disclaimer that could prohibit taking them.

I wonder what they have against the calendar system documentation. The
only "interesting" patch I can see is the utmpx one, but all the other
ones are mostly pointless. I'm not exactly sure what the arg_zero
patch does though. (One wonders why they want to modify the
install-strip target instead of just calling the install target). On a
side note and to save us answering the same problems over and over in
#zsh, I wish they would drop this too,
http://www.opensource.apple.com/source/zsh/zsh-61/zshenv

-- 
Mikael Magnusson

Re: MacOS X patches

[Bart Schaefer]

On Nov 1,  8:57am, Mikael Magnusson wrote:
} Subject: Re: MacOS X patches
}
} On Sat, Nov 1, 2014 at 5:48 AM, Pedro Giffuni <pfg@freebsd.org> wrote:
} >
} > http://www.opensource.apple.com/source/zsh/zsh-61/patches/
} 
} I wonder what they have against the calendar system documentation.

I suspect they removed it because they don't install the corresponding
functions ... because they want everyone to use their calendar manager?

"ls $fpath/calendar(N)" finds nothing.

} The only "interesting" patch I can see is the utmpx one 

Yeah, I'm a bit curious about that one.  If that's really necessary, how
does zsh even compile without it?  Does something get "configure"d away?

} I'm not exactly sure what the arg_zero patch does though.

Apparently it's somehow possible to invoke the program with argc == 0.
The call to zerr() eventually goes through zwarning() which passes
argzero to nicezputs(), so it has to have a value for that to work.

But isn't the real bug that something runs the shell with argv empty?

} (One wonders why they want to modify the
} install-strip target instead of just calling the install target).

Probably just preventing someone from trying something that's going
to fail anyway.

The no_auto.patch and config.h.ed bits appear to have something to do
with compiling from the tarball.  I'm guessing config.h.ed is run to
undo something that configure gets "wrong," and then that makes the
default build want to re-run autoconf/autoheader for dependencies.

Re: MacOS X patches

[Bart Schaefer]

In case anyone from Apple is reading this thread ...

} } > http://www.opensource.apple.com/source/zsh/zsh-61/patches/

The svn-zsh-complete.patch is both obsolete and consequently wrong.
It's based on files from 2007-2008 but there was an equivalent and
now conflicting patch in 2011.

} } The only "interesting" patch I can see is the utmpx one 
} 
} Yeah, I'm a bit curious about that one.

This sure looks to me to be redundant with the existing lines

     95 # ifndef ut_name
     96 #  define ut_name ut_user
     97 # endif

which are inside the same #if as the addition in utmpx_ut_user.patch.
I can only guess that there was an older version of that patch and
they simply updated it without checking whether it was still needed.

} } I'm not exactly sure what the arg_zero patch does though.
} 
} Apparently it's somehow possible to invoke the program with argc == 0.
} The call to zerr() eventually goes through zwarning() which passes
} argzero to nicezputs(), so it has to have a value for that to work.

I guess we could pick this one up, though it seems bizarre.