Aha - this is why I didn't use ->can originally

Punt for the time being, will investigate why later. The error looks like:

rabbit@Ahasver:~/devel/dbic$ git rev-parse HEAD
32497625ce5699576c8a63edd16401d4f741c282

rabbit@Ahasver:~/devel/dbic$ PERL5LIB=/home/rabbit/devel/sqla/lib prove -lT t/100populate.t
t/100populate.t .. 1/? Stub found while resolving method "???" overloading """" in package "Path::Class::File" at /home/rabbit/devel/sqla/lib/SQL/Abstract.pm line 118, <> line 1.
t/100populate.t .. Dubious, test returned 255 (wstat 65280, 0xff00)
All 43 subtests passed

Test Summary Report
-------------------
t/100populate.t (Wstat: 65280 Tests: 43 Failed: 0)
  Non-zero exit status: 255
  Parse errors: No plan found in TAP output
Files=1, Tests=43,  1 wallclock secs ( 0.04 usr  0.00 sys +  0.90 cusr  0.03 csys =  0.97 CPU)
Result: FAIL

diff --git a/lib/SQL/Abstract.pm b/lib/SQL/Abstract.pm
index ce4ead9..d8251ce 100644 (file)
--- a/lib/SQL/Abstract.pm
+++ b/lib/SQL/Abstract.pm
@@ -99,13 +99,18 @@ sub is_plain_value ($) {
       (
         # FIXME - DBI needs fixing to stringify regardless of DBD
         #
+        # FIXME - simply using ->can('(""') trips up Path::Class in
+        # inexplicable ways under -T (likely other modules too)
+        #
         # either has stringification which DBI SHOULD prefer out of the box
-        $_[0]->can( '(""' )
+        grep { *{ (qq[${_}::(""]) }{CODE} } @{ mro::get_linear_isa( ref $_[0] ) }
           or
         # has nummification and fallback is *not* disabled
         # reuse @_ for even moar speedz
         (
-          $_[0]->can('(0+')
+          # FIXME - simply using ->can('(0+') trips up Path::Class in
+          # inexplicable ways under -T (likely other modules too)
+          grep { *{"${_}::(0+"}{CODE} } @{ mro::get_linear_isa( ref $_[0] ) }
             and
           (
             # no fallback specified at all