<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
<title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
<br>
<blockquote cite="mid:20071220054442.GA8107@taz.net.au" type="cite">
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">if it bothers you, then strip out the extras yourself:
my $full_path = "$dirname/$entry";
$full_path =~ s=//=/=g;
</pre>
</blockquote>
</blockquote>
<pre wrap=""><!---->
it's wrong (at least, sub-optimal), anyway. should be:
$full_path =~ s=//+=/=g;
</pre>
<blockquote type="cite">
<pre wrap="">Except, don't use = as your separator. It breaks people's brains.
</pre>
</blockquote>
<pre wrap=""><!---->
only if they can't read perl.
</pre>
</blockquote>
nobody can read Perl... :)<br>
<blockquote cite="mid:20071220054442.GA8107@taz.net.au" type="cite">
<pre wrap="">
i use = (when i don't use /) because it's one of the characters least
likely to appear as a character literal in a regexp also containing /
character literals.
e.g. any regexp containing a literal / (which is why i'm using another
char as the separator, to avoid the unreadable ugliness of lots of \
escaping) is fairly likely to also have : characters (<a class="moz-txt-link-rfc2396E" href="http://.....">"http://....."</a>)
'=' also has the significant advantage of being visually distinct from
most other characters (no highs or lows, just a horizontally centred
double-bar so it highlights the separation almost like whitespace).
'-' works reasonably well too.
if the regexp has literal = characters too, i'll use something else.
the whole point is to make the regexp more readable.
</pre>
</blockquote>
except that using your http example, you can't parse url arguments
(without escaping) and the ugliness-factor would but much worse. Perl
POD often uses '!'...
maybe that is a good character to use?<br>
<br>
<blockquote cite="mid:20071220054442.GA8107@taz.net.au" type="cite">
<blockquote type="cite">
<pre wrap="">Use curly braces instead:
</pre>
</blockquote>
<pre wrap=""><!---->
i find that ugly inside regexps, and don't/won't do it. curly braces are
for hashes or code blocks etc, not for uglifying regexps.
i'll use pretty nearly anything else before i'll use them.
</pre>
<blockquote type="cite">
<pre wrap=""> $full_path =~ s{//}{/}g;
</pre>
</blockquote>
<pre wrap=""><!---->
the point of using a different separator is to make the regexp more
readable, not less.
YMMV, but to me, that is significantly less readable.
</pre>
</blockquote>
The POD contains many references where using balanced brackets is a
good thing - it would be only you that considers them ugly.<br>
<br>
Mathew<br>
<br>
</body>
</html>