Need help with distutils

[Andy Todd]

Andy Todd <andy47 at halfcooked.com> wrote in 
news:Xns91EB605BFDDADandy47halfcookedcom at 203.109.252.31:

> "A.M. Kuchling" <akuchlin at crystal.mems-exchange.org> wrote in 
> news:slrnab3gei.1i0.akuchlin at crystal.mems-exchange.org:
> 
>> In article <Xns91EA988AA8260andy47halfcookedcom at 203.109.252.31>, 
>>    Andy Todd wrote:
>>> notice that the pattern you pass to global-exclude has to be a Unix-
>>> style "glob" pattern. In that notation '.' is a special character so
>>> using  '.cvsignore' doesn't actually match your files ;-) 
>> 
>> No, '.' is not special in glob notation.
> 
> Which means that this is a bug? I can verify that ".cvsignore" doesn't 
work 
> and "*cvsignore" does.
>   

Apologies, I spoke too soon. I had "*cvsignore" excluded for another reason 
and thought it was because of glob. So I ended up combining two 
misunderstandings and confused them for a bug. I have changed my 
MANIFEST.in to read

global-exclude .cvsignore

I can verify that this does work. The problem the op had (I believe) was 
that he tried

recursive-exclude . .cvsignore

Which by my reading of the documentation *should* do the same thing, but I 
can confirm doesn't (at least on Windows2000 with distutils running under 
Cygwin). I observed the same problem - namely a message stating;

warning: no previously-included files matching '.cvsignore' found under 
directory '.'

Ah, the joys of distutils. I don't know if this is a feature or a bug, but 
at least we can workaround it.

>> 
[snip]


Regards,
Andy
-- 
Contents free posts a speciality