On Tue, Feb 26, 2008 at 12:44 AM, Danek Duvall <[EMAIL PROTECTED]> wrote:
> On Mon, Feb 25, 2008 at 09:39:58PM -0600, Shawn Walker wrote:
>
>  > I'm working on bug 244 ( 
> http://defect.opensolaris.org/bz/show_bug.cgi?id=244 ).
>
>  I want to fix this by introducing a real parser for the filters, rather
>  than relying on this bastardized version of python.  I'm using PLY for the
>  solaris.py replacement, but the OSR for that hasn't gone through yet, so
>  I haven't bothered integrating its use that deeply in the IPS codebase.

Can I still make an interim fix then or should I just abandon this change?

>  > For example, suppose that we have something tagged as such:
>  >
>  >                 file path=/usr/bin/hostname arch=386 version=1.0
>  >
>  > Should arch=386 & version=1 match the first line?
>  >
>  > Or, should the first line only match if arch=386 & version=1.0?
>
>  The latter.  All attribute values are considered strings; the reason that
>  numerics don't work right now is because Python identifiers can't start
>  with a number.

Okay, well I have it working already and I have a few test cases for
it working as simple strings.

Is it worth posting for review and integration then?

>  This means that parsing an attribute like "3ufw355" will be tough, since
>  it's neither a number nor an identifier.  Which is one reason that we may
>  not be able to use the Python parser to do this.

That would make sense.

>  I'm planning on extending the language to allow for substring matches, or
>  something similar, but again, that'll be tons easier with a lower-level
>  parser.

I definitely can agree with that.

Cheers,
-- 
Shawn Walker, Software and Systems Analyst
http://binarycrusader.blogspot.com/

"To err is human -- and to blame it on a computer is even more so." -
Robert Orben
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss

Reply via email to