On Sat Jul 10 04:27:37 2010, ADAMK wrote:
Show quoted text> Running the benchmark on Windows, I get "MISMATCHED TEMPLATE OUTPUT" for
> most of the templates.
>
> I suspect it may have something to do with newline types.
This sounds likely, I'm going to have to get a windows install of perl
up and running before I can investigate though.
About time I investigated strawberry...
Show quoted text> Should the benchmark ignore newline differences?
There's two quick "fixes" that I see and one slower one depending on
what the root cause turns out to be.
I'll work on making the template output be normalized before comparison,
and while I'm at it I'll add an option that lets you bypass the
comparison if you think the failure is bogus.
That should fix the immediate problem and provide a work-around for
people should it crop up in another form in the future.
Third fix will be to find out if the difference is in the templates
being executed rather than just the output, if so then that will need
some tweaking to ensure that like-for-like execution is correct on
windows. (Although I doubt it'll impact the results significantly.)
I hope that'll address your bug, and please let me know if you think
I've missed something obvious in this. I've not used perl on windows in
over a decade, so I'm pretty rusty.
Thanks for the bug report.