Philip Hazel wrote:
> I have placed release candidate tarballs here:
>
> ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre/Testing/pcre-7.3-RC4.tar.gz
> ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre/Testing/pcre-7.3-RC4.tar.bz2
> ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre/Testing/pcre-7.3-RC4.zip
>
> These are "true" release candidates in the sense that the only changes
> I'll make before the proper release are to fix problems that show up in
> this code. Please test if you can. I am about to take a short vacation.
> Back next Wednesday.
>
> Philip
>
>
Hi, just back from vacation myself (Alaska, beautiful!) and I've just
built and briefly tried 7.3 RC4.
Think possibly you overlooked the following issue I previously posted here.
Regards,
Sheri
Subject: matching problem w default newline of anycrlf
> pcre default newline is anycrlf
>
> This matches:
>
> re> /(\n|\r)/
> data> \r\na
> 0: \x0d
> 1: \x0d
>
> and this matches:
>
> re> /\n/
> data> \r\na
> 0: \x0a
>
> but this fails:
>
> re> /(\n|\r)a/
> data> \r\na
> No match
>
> Matches same data with either of these:
>
> re> /(\n|\r)a/<lf>
> re> /(\n|\r)a/<cr>
>
> fails with any of these:
>
> re> /(\n|\r)a/<crlf>
> re> /(\n|\r)a/<any>
> re> /(\n|\r)a/<anycrlf>
>
> I expected only the behavior of a few metacharacters to be affected by
> the newline in effect. In our application update project we added new
> functions that allow the user to specify external options, but we also
> maintained the existing posix based functions which allow for only
> internal options. Since the libary default is the only choice for posix,
> and (as part of our application update) we changed from LF to ANYCRLF as
> default, we are now hearing about user conversion difficulties.
>
> Regards,
> Sheri
>
>
>
