Zsh Mailing List Archive
Messages sorted by: Reverse Date, Date, Thread, Author

Re: did the way arrays get indexed change between 4.2 and 4.3?



On Jul 25,  9:47am, Peter Stephenson wrote:
} Subject: Re: did the way arrays get indexed change between 4.2 and 4.3?
}
} Bart Schaefer wrote:
} > The flavor of this that works is very recent.  What you're seeing in
} > 4.2.3 is what you'll find pretty much universally in "stable" versions
} > of the shell shipped with various OSs and Linux versions.
} 
} Are you sure?  I get the same result in 4.2.6 and 4.3.4.

OK, it's not quite as recent as I thought -- it was fixed between 4.2.5
and 4.2.6:

2005-09-06  Peter Stephenson  <pws@xxxxxxx>

        * 21701: Src/params.c, Test/D06subscript.ztst: negative offsets
        that indexed before the start of a scalar weren't handled properly.

RedHat EL 4.5 still ships zsh 4.2.0, though, so there are still going to
be a lot of stock systems out there (RHEL and otherwise) with versions
of zsh old enough to have this bug.

} When it gets it wrong it's presumably related to the other bug you
} pointed out.

Seems not; 21701 fixes range-starting indices in a similar way to 23696
fixing range-ending indices.



Messages sorted by: Reverse Date, Date, Thread, Author