On Fri 28 Nov 2003, DH <[EMAIL PROTECTED]> wrote:
> C:\new\CGI-Wiki-0.50_02>nmake test

DH, thanks for your patience.  We *will* get this working eventually.
Part of the reason I rewrote the entire test suite was to make it
easier to narrow down what is breaking on your system.

> t\012_blank_data............ok 4/7Use of uninitialized value in subroutine entry at
> C:\new\CGI-Wiki-0.50_02\blib\lib/CGI
> /Wiki/Store/Database.pm line 191.

Now this is a bit odd; that value should be the empty string, not
undef.  I wonder where it's getting lost.  This may be related to the
test that is dying.  Could you send me the verbose output of test 012
please?

> C:\new\CGI-Wiki-0.50_02>nmake test TEST_FILES=t\011_recent_changes.t TEST_VERBOSE=1
> [...]
> No content parameter supplied for writing at t\011_recent_changes.t line 137

This is a real puzzle.  The content parameter passed in on line 137 is
the empty string; that croak is only triggered when undef is passed.
And on line 46 of the very same test, the very same node is written
with the very same content, and all is OK!

I don't suppose I could get you to run that test through the debugger
and find out exactly when that empty string turns into undef?  I can't
reproduce your bug on any of my systems, so I can't do it myself.

Kake

Reply via email to