On October 12, 1998 at 10:58, Emery Ford wrote:
> The way it currently hyperlinks urls includes includes often used
> punctuation in the context of a sentence.
Things like '?' and '!' at the end were unintentionally left out.
> Also, query strings aren't
> usually hyperlinked appropriately.
This is a known limitation.
> All of the changes mentioned here only afffect the file mhtxtplain.pl
>
> $HUrlExp = $Url . q%[^\s\(\)\|<>"']*[^\.;,"'\|\[\]\(\)\s<>\?\!]%;
>
> This allows query stings such as
> http://www.infoseek.com/Titles?qt=hello+world&col=WW&sv=IS&lk=noframes,
> which are currently broken off at the first ampersand.
And broke with what you have. "Htmlization" is done first before
URL detection, so '&' will actually become '&'. If you include
'&' in the expression, then you are actually generating an incorrect
link. ('&' was a horrible choice as an argument separator)
> $str =~ s/A\ \ HREF/A HREF/g; # ADDED BY EMERY
Not good if "A HREF" is actual message data. Also, does
not deal with the case when the target option is specified.
I have code that does a better job with URL detection (and even handles
the <URL:...> notation with embedded newlines), but supporting
"keepspace" causes a problem.
--ewh
----
Earl Hood | University of California: Irvine
[EMAIL PROTECTED] | Electronic Loiterer
http://www.oac.uci.edu/indiv/ehood/ | Dabbler of SGML/WWW/Perl/MIME
