My suggestion was to change it conditionally, for EBCDIC only, with clear #if defined, etc. and I volunteered to code it, but I do understand why you do not want to do it. If we ultimately decide not to do it, we should publish that it is not fully supported for EBCDIC.
I will try the suggestions about the escape sequences, but as opposing to \c, the rest of them should actually work! Ze'ev Atlas
From: "ph10@???" <ph10@???>
To: Ze'ev Atlas <zatlas1@???>
Cc: Pcre Exim <pcre-dev@???>
Sent: Wednesday, June 10, 2015 12:38 PM
Subject: Re: issues with EBCDIC and pcretest
On Wed, 10 Jun 2015, Ze'ev Atlas wrote:
> In my opinion, for the \c sequences, pcre_internal.h should have a
> section to explicitly define the below characters (taken from
> perlebcdic). Something like:
Hmm. The code that handles \c for EBCDIC has been in PCRE for a long
time. I don't know if changing it will affect anybody, but in my
experience there is always *somebody* who is relying on current
behaviour. I am loath to change it now, especially as \c is such an
obscure feature.
Philip
--
Philip Hazel
