[Distutils] Distutils roadmap of refactoring for 2.7

[Ronald Oussoren]

On 3 Feb, 2010, at 0:25, David Lyon wrote:

>> Chris wrote:
>>>> On Sat, Jan 23, 2010 at 7:11 PM, P.J. Eby <pje at telecommunity.com>
>>>> wrote:
>>>> Notice that other changes are coming up once
>>>> the series of PEP we worked on are accepted (which should be before
>>>> Pycon - at least for 345 and 386)
>>> 
>>> ok - so are you going to sort out the confusing (i386 == mac)
>>> and (x86 == pc) or (i586 == pc) or (i686 == pc) ?
>>> 
>>> in PEP-345? or leave it as downright confusing?
>> 
>> David,
>> 
>> If you're going to hijack a thread to bang on about your pet peeve,
>> please at least attribute the mail you're quote to the right person...
> 
> ok. I agree.
> 
> But it isn't a pet peave..
> 
> Introducing 'code' into the metadata will mean that PKG_INFO
> files will become hated just as much by developers as what
> .PTH have been.
> 
> Let me remind you what happened, .PTH files were supposed to
> be data files (like metadata files).
> 
> Then a bright spark said, 'Let them contain code..'. And then
> they did..
> 
> Then another bright spark said 'I shalt introduce code that
> nobody else cannot understand or change'.
> 
> Then regular users found the whole thing such a nightmare
> they came to loath the entire concept of a .PTH file.
> 
> So now, the exact same dreaded mechanism, is going to
> be introduced into PKG_INFO files and I can't see why
> it won't be also hijacked to make system administrators
> lives a misery.
> 
> Giving a normal system administrator the job of knowing
> that an 'i386' is actually a mac is going to drive them
> bananas figuring out why the package isn't installing
> properly on a pc.

Could you please explain what the problem is you're complaining about instead of assuming everyone knows the details by heart? The only reference I saw to 'i386' in PEP 345 is in the section about Environment markers and that only refers to the CPU architecture in a test for CPU architecture (platform.machine == 'i386') which seems reasonable to me.  

You can use "sys.platform" to detect various OS types.

BTW. I guess you mean "Windows" where you write "pc".

> 
> Nine out of ten system administrators would guess that
> 'i386' is actually a pc. It should be but it isn't.

'i386' is the commonly used abbreviation for an instruction set also known as 'x86' or 'ia32' and doesn't suggest any particular OS to me, all of Windows, Linux and MacOSX use or can use 'i386' capable CPUs (to just name 3 commonly used operating systems).

Ronald

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 3567 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/distutils-sig/attachments/20100203/67938bac/attachment.bin>