Front page | perl.perl5.porters |
Postings from February 2001
On Sun, Feb 25, 2001 at 03:48:02PM +0100, Karsten Sperling wrote:
> >> > This does look like a problem:
> >>
> >> > $ uname -a
> >> > OS/390 LPR28 08.00 02 9672
> >> > $ perl -e '$_=qq/\x8E/; print/[\x89-\x91]/, q/,/ ,tr/\x89-\x91//d'
> >> > ,0$
> >>
> >> > (that was perl 5.005_03 FWIW)
> >>
> >> Well it means that character ranges and even transliterations
> >> are essentially broken when you try to use them for working on
> >> binary data.
>
> > Yes it does. This BTW looks like a good candidate for a regression test.
>
> Here's a patch for tr.t that checks for this one and another
> thing i just found. The last 3 cause segfaults / out of memory errors
> during compilation.
Thanks, I extended the test somewhat (to check also for the uppercase
gaps), made the tests conditional on being on an EBCDIC box, and added
similar // tests to pat.t. Now I have to figure out what's the core
dump about...
--
$jhi++; # http://www.iki.fi/jhi/
# There is this special biologist word we use for 'stable'.
# It is 'dead'. -- Jack Cohen
Thread Previous
|
Thread Next