zrecompile

zrecompile

[Andrej Borsenkow]

This is the second time ... not sure what's going wrong:

bor@itsrm2% l .zsh.d
completion/      completion.zwc   functions/
bor@itsrm2% zrecompile
zrecompile:zcompile:144: invalid dump file:
/home/bor/.zsh.d/completion.zwc
re-compiling /home/bor/.zsh.d/completion.zwc: failed
zrecompile:zcompile:144: invalid dump file:
/home/bor/.zsh.d/completion.zwc
re-compiling /home/bor/.zsh.d/completion.zwc: mv: Cannot access
/home/bor/.zsh.d/completion.zwc: No such file or directory
failed
bor@itsrm2% l .zsh.d
completion/          completion.zwc.old   functions/
bor@itsrm2% print $fpath
/home/bor/.zsh.d/completion /home/bor/.zsh.d/completion.zwc
/home/bor/.zsh.d/functions /tools/share/zsh/site-functions
/tools/share/zsh/3.1.6-dev-20/functions/Base
/tools/share/zsh/3.1.6-dev-20/functions/Builtins
/tools/share/zsh/3.1.6-dev-20/functions/Commands
/tools/share/zsh/3.1.6-dev-20/functions/Core
/tools/share/zsh/3.1.6-dev-20/functions/Debian
/tools/share/zsh/3.1.6-dev-20/functions/Linux
/tools/share/zsh/3.1.6-dev-20/functions/Misc
/tools/share/zsh/3.1.6-dev-20/functions/Prompts
/tools/share/zsh/3.1.6-dev-20/functions/User
/tools/share/zsh/3.1.6-dev-20/functions/X
/tools/share/zsh/3.1.6-dev-20/functions/Zftp
/tools/share/zsh/3.1.6-dev-20/functions/Zle

I suspect, that zrecompile is confused by _directory_
~/.zsh.d/completion and wordcode file ~/.zsh.d/completion.zwc. The file
was produced by

zcompile ~/.zsh.d/completion.zwc /tools/share/zsh/$ZSH_VERSION/**/*(^\)

-andrej

Have a nice DOS!
B >>

Re: zrecompile

[Bart Schaefer]

On Apr 2,  2:01pm, Andrej Borsenkow wrote:
} Subject: zrecompile
}
} This is the second time ... not sure what's going wrong:
} 
} bor@itsrm2% zrecompile
} zrecompile:zcompile:144: invalid dump file:
} /home/bor/.zsh.d/completion.zwc

I get that error when I try to use -dev-20 to recompile a file that was
written with -dev-19.  I think the -k/-z option changes introduced this
incompatibility.  You'll just have to recompile "by hand."

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

Re: zrecompile

[Sven Wischnowsky]

Bart Schaefer wrote:

> On Apr 2,  2:01pm, Andrej Borsenkow wrote:
> } Subject: zrecompile
> }
> } This is the second time ... not sure what's going wrong:
> } 
> } bor@itsrm2% zrecompile
> } zrecompile:zcompile:144: invalid dump file:
> } /home/bor/.zsh.d/completion.zwc
> 
> I get that error when I try to use -dev-20 to recompile a file that was
> written with -dev-19.  I think the -k/-z option changes introduced this
> incompatibility.  You'll just have to recompile "by hand."

Right. zrecompile should be a bit more clever to handle error
conditions, though.

And another thing: the zwc files till use $ZSH_VERSION in the header
to test for compatibility -- somehow I didn't like to add an additional 
version number scheme for them, but it would be better, I think (the
format will certainly change less often than $ZSH_VERSION).

No patch, I'm still compiling the thing from CVS...

Bye
 Sven

P.S.: Sourceforge Bug #103797 can be closed, it seems. It also seems
      that only administrators can do that.

--
Sven Wischnowsky                         wischnow@informatik.hu-berlin.de

Re: zrecompile

[Sven Wischnowsky]

I forgot:

zrecompile probably makes a whole lot more sense when used with the -p 
option, because then it also catches new functions in directories.

Maybe we should make that the default and the automatic zwc file
searching only be done when given some option.

Bye
 Sven


--
Sven Wischnowsky                         wischnow@informatik.hu-berlin.de

Re: zrecompile

[Bart Schaefer]

On Apr 3, 10:39am, Sven Wischnowsky wrote:
} Subject: Re: zrecompile
}
} And another thing: the zwc files till use $ZSH_VERSION in the header
} to test for compatibility -- somehow I didn't like to add an additional 
} version number scheme for them, but it would be better, I think (the
} format will certainly change less often than $ZSH_VERSION).

Yea, but it'll be much less recognizable in the -t output.  I think the
$ZSH_VERSION test is fine.

On the other hand, that should be put in some immutable part of the file
header so that *any* version of zsh can be guaranteed to be able to read
it back -- to rephrase, we set the requirement NOW that changes to the
format of the header must only come *after* the version string (or some
equivalent requirement), so that it's always possible to display the
version mismatch (or at least know that was the reason for the error).

On Apr 3, 11:45am, Sven Wischnowsky wrote:
}
} zrecompile probably makes a whole lot more sense when used with the -p 
} option, because then it also catches new functions in directories.
} 
} Maybe we should make that the default and the automatic zwc file
} searching only be done when given some option.

My instinct is to say that it's OK the way it is; but I haven't used it
very much, so ...

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