Re: Inline-0.55 released to CPAN

On 04/29/2014 03:24 AM, [email protected] wrote:
> -----Original Message----- From: Reini Urban
>
>> Yes, we at rperl are quite happy how it turned out. Thanks Rob!
>
> You're welcome. Many thanks to you and the other contributors.
>
>> There are still some minor issues I wanted to work on, but had no time
>> yet.
>
> No rush - I can deal with them as they get fixed.
>
>> - need to exec make with tainted path in t/08taint.t
>
> There's another variant of that failure just cropped up:
> http://www.cpantesters.org/cpan/report/ba327e3a-7235-1014-b072-7ddc5cf4dae8
> I don't recall having seen that one before. (The "require
> './t/08taint_1.p';" fails.)

You usually have make in a non-tainted path. But on some smokers you 
cannot rely on that. So we need to untaint /usr/bin or wherever the make 
in found. Nothing serious, just annoying.

>> - t/10callback.t error on windows in one report
>
> That one's a bit of an odd smoker - x64 ActivePerl (built using Platform
> SDK) being used with mingw64 gcc-4.5.4 compiler.
> I doubt that it's an Inline::C bug - more likely something wrong with
> the hacking of some flag(s).
> I'd like to see a "-s" in its $Config{optimize} and I'd like to see
> $Config{ivtype} set to "long long" instead of "__int64".
> I seem to recall experiencing problems in relation to both of those (in
> the past) .... though I can't really see why either *should* pose any
> problems.
> It would certainly be nice to understand what's going wrong there.

Agree

>
>> - missing LICENSE (in meta and as file)
>
> What needs doing as regards that ?
> (I can then immediately fix that in readiness for next release.)

The cpan "nazis" are requiring it in their meta tests.
Guess it's for the debian folks.


-- 
Reini