Configure I32df, U32uf, etc. in config.h rather than in perl.h

[t-a-k]

I32df, U32uf, and similar macros had been defined in perl.h using preprocessor conditionals, but determining wheter I32 is int or long is pretty hard with preprocessor when INTSIZE == LONGSIZE.
I think it is better to make Configure script determine whether %d or %ld shall be used to format I32 value, as it should know exact underlying type of I32.

Without this change, U32 is typedef'ed to unsigned long and U32uf is defined to "u" for -Duse64bitint build on 32-bit host, which leads a lot of build warnings about mismatches between LINE_Tf (= U32uf) format and line_t (= U32) argument.

[khwilliamson]

I think these should be squashed into a single commit; but otherwise LGTM

[t-a-k]

I'm sorry that there is a mistake in uconfig64.sh. I'll fix it (and rebase to current blead).

[khwilliamson]

@t-a-k Please squash the commits into one, and it will get merged

[demerphq]

Doesnt some of this need to be coordinated with @Tux ?

[jkeenan]

Do we have any way of testing the changes to the two files in this p.r. pertaining to the plan9 OS?

[tonycoz]

Doesnt some of this need to be coordinated with @Tux ?

As long as you let him know, I believe we can just commit, and he'll backport.

Do we have any way of testing the changes to the two files in this p.r. pertaining to the plan9 OS?

If anyone is interested in setting up a VM http://9legacy.org/download.html but I'm not.

[Tux]

Doesnt some of this need to be coordinated with @Tux ?

Yes it does. The change to Configure and related files looks fine, except for the omission of the entries in the Glossary.
If you(plural) add those, go ahead and I'll backport

[t-a-k]

If you(plural) add those, go ahead and I'll backport

I've pushed a commit (2694c7d) to add Porting/Glossary entries.
(Another push is to update perldelta.)

If this is okay, I will squash the commits into one.