On Fri Jul 29 10:21:14 2011, NINE wrote:
Show quoted text> Please upgrade the Digest::MD5 requirements to at least version 2.50.
> After a year of
> trying to find out why a certain test of Inline::Python failes on many
> machines, I finally
> tracked it down to different Digest::MD5 versions being installed.
>
> Versions < 2.50 do not preserve UTF-8-ness of the arguments but
> downgrade them. This
> is why for me UTF-8 encoded Python code never gave a problem, because
> Python never
> saw it. When I found out that I had to tell the Python interpreter
> that the code was in UTF-
> 8, it suddenly broke for me until I moved down the "use utf8;" in the
> test file to below the
> Python code.
>
> I just added the requirement of Digest::MD5 >= 2.50 to Inline::Python,
> but since I'm not
> the one using this module and there may be other Inline modules having
> similar problems,
> it should be added to Inline itself.
Hi,
cc'ing Ingy in case he has any thoughts about this.
Strictly speaking, if Inline modules (other than Inline itself or
Inline::C) have similar problems then *they* should take the same
action as you have taken wrt Inline::Python.
If neither Inline nor Inline::C are affected, then there's no need
(again, "strictly speaking") for Inline to do anything about it.
However, if you could provide a simple test case that shows that either
Inline or Inline::C is affected, then that might force my hand :-)
In any case, I'll consider doing as you suggest. (It's the flow-on
affects that deter me - eg since Digest::MD5-2.50 requires perl-5.8,
Inline could no longer support perl-5.6, so there are some additional
amendments that might need to be made. I don't really want to be doing
stuff that's unnecessary ... otoh it's probably time that Inline
stopped supporting 5.6 anyway.)
Thanks for raising this.
Cheers,
Rob