geda-developers team mailing list archive

[Peter Clifton <pcjc2@xxxxxxxxx>]

This commit broke one of my simulations which had an existing "file=?"
attribute for a model statement.

Why is "foo=unknown" better than "foo=?", and isn't changing this
against our usual  requirement of not breaking existing designs?



commit d5d1a98f0e95589c33bff881198cca01a1b08c7e
Author: Clif Cox <clif@xxxxxxxxxxxxx>
Date:   Mon Jun 13 13:38:39 2011 +0100

    Updated the problematic name attributes in some symbols
    
    In some gnetlist backends the name=? attributes cause problems because
    they look like valid attribute values. eg file=? may be mistaken for a
    valid filepath, and value=? for an actual value. Changing them to
    name=unknown is one way to solve this problem in the interim. It's now
    safe to remove the special checks for these "?" like the one in
    spice-sdb. This patch updates these problematic name=? attributes in
    some symbols to name=unknown, and removes FIXME code snipit that
    checked for the ? value in spice-sdb.
    
    It also updates gnetlist tests to reflect these changes. Turns out
    that the test input file TwoStageAmp.sch had a ? in it, which of couse
    caused problems.  Then when it was removed vams's output changed so it
    had to be fixed, which wasn't a very clever fix but seems to work.
    
    Closes-bug: lp-698741
    Reviewed-by: Peter TB Brett <peter@xxxxxxxxxxxxx>


-- 
Peter Clifton <pcjc2@xxxxxxxxx>

Attachment: signature.asc
Description: This is a digitally signed message part