At 01:04 PM 10/8/2001 -0400, Charles Lane wrote:
>> Does this mean we will always have six zeroes in any absolute path spec?  Yuck.
>
>No, it gets *rid* of extra 000000's , because if both the input-path and the
>cwd "chunks" match, they're eliminated from both.

OK, sounds good then. 

>> We also need to make the debugger clean up PERL_DB_PIDS from the environment when 
>it 
>> exits; otherwise a second run of the debugger always thinks it's a daughter 
>> db session.
>
>I have a patch (following) for this, but it's kinda ugly...for
>whatever reason, it seems like the "END" block isn't being executed,
>so I had to patch into the logic for the 'q' command.  It's not 
>VMS-specific, so it should probably be vetted by the Unix crowd.
>
>BTW, before using the "Term::Cap" entry = undef, I tried grabbing a 
>VT100 termcap entry and just forcing it's use...there were problems in
>the test suite, but the Perl debugger had some *nice* underlining and
>highlighting in the meanwhile.

Interesting, might be worth pursuing further.

>diff -uBb lib/perl5db.pl-orig lib/perl5db.pl
>--- lib/perl5db.pl-orig Fri Oct  5 11:17:50 2001
>+++ lib/perl5db.pl      Fri Oct  5 11:40:29 2001
>@@ -721,7 +721,7 @@
>                            next CMD;
>                        } 
>                    }
>-                   $cmd =~ /^q$/ && ($fall_off_end = 1) && exit $?;
>+                   $cmd =~ /^q$/ && ($fall_off_end = 1) && clean_ENV() && exit $?;



Exiting with $? is probably wrong for us 99% of the time as well.  This 
looks like a big step in the right direction, though.

Reply via email to