Re: Building XML-Xerces-2.3.0-3 on Windows

Jason E. Stewart wrote:

>>t\Schema..................ok 3/3
>>
>>Handling Unknown exception, <IN> chunk 1.
>># Looks like your test died just after 3.
>>t\Schema..................dubious
>>         Test returned status 255 (wstat 65280, 0xff00)
>>         after all the subtests completed successfully
>>
>>Test 3 says: "Resolver: Couldn't create input source for entity:
>>No matching function for overloaded 'new_LocalFileInputSource' at
>>Xerces.pm line 1786."
> 
> 
> Once again this message happens because the constructor gets called
> without a filename - which means that the EntityResolver actually
> failed to resolve things.
> 
> Could you activate the print statements in the test code and run it
> again? 

If I put a "print $@" after each eval and run the test separately then 
it passes ok ...

D:\Src\XML-Xerces-2.3.0-3> perl t/schema.t

1..3
ok 1 - module loaded
ok 2
ok 3
FATAL ERROR:
FILE:    FAKE_SYSTEM_ID
LINE:    41
COLUMN:  6
MESSAGE: Expected end of tag 'personnel'
  at t/schema.t line 50
Handling Unknown exception, <IN> chunk 1.
# Looks like your test died just after 3.

(I'm not sure if the last two lines are correct.)

>>t\URLInputSource..........dubious
>>         Test returned status 2 (wstat 512, 0x200)
>>DIED. FAILED tests 6, 8
>>         Failed 2/8 tests, 75.00% okay
>>
>>In Test 6 $error->GetCode() returns 0 (correct: 105).
>>In Test 8 $error->GetCode() returns 101 (correct) if called
>>immediately, later 67114035.
> 
> 
> sounds like serious problems in the exception handling code.
> 
> 
>>It seems the error code somehow gets lost or overwritten when the
>>$error object is referenced again. If you reverse the arguments in the
>>"ok" clause as follows it passes the test:
>>
>>ok($error->getCode() == $XML::Xerces::XMLExcepts::URL_NoProtocolPresent
>>&&
>>UNIVERSAL::isa($error,'XML::Xerces::XMLException')
>>&&
>>$error);

> That will be a big problem for handling exceptions - the above code
> will die if the error isn't an exception object...

Yes, I just wanted to show that the code is returned correctly if called 
first. May be the reason is that the exception handling code itself 
raises an exception.

Martin