From hgin2088@mail.kolej.mff.cuni.cz

  There should be

  The `while' statement executes ... or `"0"').

istaed of ending or `"0")' in perlsyn.pod.

  In perlre.pod in the paragraph starting
  `On simple groups, ...'
is somenthing also wrong with apostrophes or re's.

Subject​: Getting Locales - General Protection Fault

  The first example in the section
  `The localeconv function'
of
  perllocale.pod
makes General Protection Fault.

OS version​: Microsoft Windows 95. [Version 4.00.1111] Czech

Bash version​: 1.14.7 for DJGPP

Perl​: This is perl, version 5.005_02 built for dos-djgpp
  MS-DOS port Copyright (c) 1989, 1990, Diomidis Spinellis
  djgpp v2 port (jpl5003c) by Hirofumi Watanabe, 1996
  djgpp v2 port (perl5004+) by Laszlo Molnar, 1997-1998

  Last thing.

  Why @​a=('\x80'..'\xFF') does not create an array?

  Best regards,

  _ Hans
  ,| |/ /~~~, ,
  | /| / . /
  |/ |/// // \~~ | _ / // / // /
  /| /| / /_ / / , \ \ / / / / / /~~~ /
  ~ ~ ~~~ ' ' ~~~ ~~~|' ' ' ' ~~~ '
  Hans GINZEL /| Room​: B509
  Faculty of Mathematics / / Kolej 17. listopadu
  and Physics ~~ Pa'tkova 3
  Charles University Prague 182 00 Praha 8
  MailTo​:Hans@​kolej.mff.cuni.cz CallTo​:++420+2+8551040-49/761

From [Unknown Contact. See original ticket]

On Fri, 20 Aug 1999 at 19​:00​:05 +0200, Hans GINZEL wrote​:

There should be
The `while' statement executes ... or `"0"').
istaed of ending or `"0")' in perlsyn.pod.

This one's easy​:

--- pod/perlsyn.pod.orig	Tue Jul 20 18:18:10 1999
+++ pod/perlsyn.pod	Sat Aug 21 11:05:29 1999
@@ -156,5 +156,5 @@
 
 The C<while> statement executes the block as long as the expression is
-true (does not evaluate to the null string (C<"">) or C<0> or C<"0")>.  The LABEL is
+true (does not evaluate to the null string (C<""> or C<0> or C<"0">).  The LABEL is
 optional, and if present, consists of an identifier followed by a colon.
 The LABEL identifies the loop for the loop control statements C<next>,

What do you think is wrong? The source code 'C<a>s' to indicate multiple
examples of 'a' isn't very clear; after pod2man->man it just reads
'as', which could certainly confuse non-native English speakers.

Subject​: Getting Locales - General Protection Fault
OS version​: Microsoft Windows 95. [Version 4.00.1111] Czech

Sorry, I personally can't help here - I neither have any interesting
Locales nor Windows 95 perl. If you've built this perl from source, you
should be able to run 'perlbug' which will help you to submit an email
with more info. You imply, but don't state that you have the Czech
locale set​: please be explicit, try again with other locales and submit
the bug again with 'perlbug'. If you didn't build the perl binary
yourself, you need to talk to whoever did.

Last thing.
Why @​a=('\x80'..'\xFF') does not create an array?

Well, it did; it's just that the array only has one element :-)

The auto-increment only works on numbers, and ascii letters+numbers (see
perlop). Hence the initial value is taken as 0, incremented to 1 and
compared with the '\xFF', which is also 0. Hence one array element.

What you need is​:
  @​a = map { chr } 128..255;

There's an issue here​:
  Should not ++ warn before converting junk into numbers?

At the moment { $x='?'; $x++; } silently leaves $x as '1'.
Ilya (Cced) was arguing against my idea of "123\0foo" being recognised
as a number without warning. Here we find '?' being regarded as the
number 0 without warning.

Ian

From [Unknown Contact. See original ticket]

On Tue, 24 Aug 1999 at 17​:00​:08 +0200, Hans GINZEL wrote​:

Hallo\.

Thank you for quick answer.

On Sat, 21 Aug 1999, Ian Phillipps wrote​:

On Fri, 20 Aug 1999 at 19​:00​:05 +0200, Hans GINZEL wrote​:

In perlre.pod in the paragraph starting
`On simple groups, ...'
is somenthing also wrong with apostrophes or re's.

What do you think is wrong? The source code 'C<a>s' to indicate multiple
examples of 'a' isn't very clear; after pod2man->man it just reads
'as', which could certainly confuse non-native English speakers.

Not, that's not the problem. I've now looked in the source code
(perlre.pod) and there is​:

On simple groups, such as the pattern C<(?> [^()]+ )>, a comparable
effect may be achieved by negative lookahead, as in C<[^()]+ (?! [^()] )>.

I think problem is in the first greather then symbol. It should be
escaped.

This is fixed in 5.005_61.

I am working on Linux with Czech locales. It seems they do work in
perl. I only wanted to test what about that on Microsoft Windows. I have
expected, that would be all English, but not an crasch (protection
fault). I had should write it direct to DJGPP's team. I noticed You have
sent copy to perl5-porters. Thank you. I hope that was.

Messages from Perl are always in English, sorry.
If you compiled Perl from DJGPP yourself, then it could be Perl, or the
compiler, or your operating system at fault. Welcome to the perl
bug-fixing fraternity :-)

There's an issue here​:
Should not ++ warn before converting junk into numbers?

I think not. That is the definition of ++ and conversion of scalar to
number. I also understand the magical function of ++​:

It was the conversion to numbers by '++' that I was querying. I don't
think it's right that '?'->'1' is a valid '++' result. The alphabetic
case of 'Z'->'AA' is quite separate.

that range
"<one 8-bit character>" .. "<one 8-bit character>"

Sorry, no. Suppose you are running Perl on an EBCDIC machine, and
specify 'A'..'Z'; what results? Perl does have DWIM (do what I mean)
functionality in a lot of places, but this one would add too much
ambiguity.

Currently, the '..' operator uses increment to get from one bound to the
other; it's the only reasonable interpretation. Increment is not defined
on characters, since Perl doesn't work with characters like that.
Definition of 'characters' is itself a minefield, with Unicode and
Locales interfering.

is well defined and should work same way as in Pascal.

Well, if you want Pascal, you know where to find it :-)

Ian