Re: Expect pattern match in printcmds.exp and setvar.exp

[Syd Polk <spolk at redhat dot com>]

At 09:09 PM 9/28/00 +0200, Orjan Friberg wrote:
>Orjan Friberg wrote:
> >
> > Obviously I haven't understood expect's pattern matching, so a URL would
> > be greatly appreciated. I've been digging through the expect info pages
> > without finding a list of the meaning of special characters. I have
> > patches for both test cases ready, but I would like to understand why it
> > doesn't work in the first place before submitting them.
> >
>
>Just to save other people from the trouble; I educated myself with
>dev.scriptics.com:
>
>{
>      when followed by a character other than a digit, matches the
>left-brace character `{';
>      when followed by a digit, it is the beginning of a bound (see
>above)
>
>The cases where I had to put \\ the { was followed by a digit, which
>sort of explains it. Still, I'm a bit surprised that these testcases
>should fail for native gdb without anybody else noticing, so I will
>defer from posting the patches until someone else confirms that they
>actually do.
>
>Built from the very latest cvs sources, for a native i686-pc-linux-gnu I
>get:
>
>printcmds.exp:
>
># of expected passes            658
># of expected failures          1
># of unresolved testcases       11
>
>setvar.exp:
>
># of expected passes            98
># of unexpected failures        20
># of unresolved testcases       22
>
>All of the unexpected failures and unresolved testcases are due to this
>pattern matching problem. runtest --version on my system says:
>
>Expect version is       5.31.2
>Tcl version is          8.2
>Framework version is    1.3.1

We do not support Tcl 8.2 when running our tests; as you have seen the 
regexp symantics changed. You must run the test suite with the version of 
dejagnu, expect and tcl that get checked out when you check out gdb.

If your change still runs with Tcl 8.0 as the base, then it will be ok.


Syd Polk		spolk@redhat.com
Engineering Manager	+1 415 777 9810 x 241
Red Hat, Inc.