python-odf

python-odf

[Marco Atzeri]

Hi Jon,
while updating python-odf for the 3.9 round I noticed that for 1.4.1-1
I wrongly set a arch package instead of noarch as was for 0.9.5-1.

To solve the issue I copied the content of

/sourceware/ftp/anonftp/pub/cygwin/x86_64/release/python-odf
on
/sourceware/ftp/anonftp/pub/cygwin/noarch/release/python-odf


and removed the

/sourceware/ftp/anonftp/pub/cygwin/x86_64/release/python-odf
/sourceware/ftp/anonftp/pub/cygwin/x86/release/python-odf

I assume the modification should propagate at the next full rebuild 
(that I can not trigger)

I plan tomorrow to deploy almost all python packages



Regards
Marco

Re: python-odf

[Marco Atzeri]

On 20.12.2021 23:14, Marco Atzeri wrote:
> Hi Jon,
> while updating python-odf for the 3.9 round I noticed that for 1.4.1-1
> I wrongly set a arch package instead of noarch as was for 0.9.5-1.
> 
> To solve the issue I copied the content of
> 
> /sourceware/ftp/anonftp/pub/cygwin/x86_64/release/python-odf
> on
> /sourceware/ftp/anonftp/pub/cygwin/noarch/release/python-odf
> 
> 
> and removed the
> 
> /sourceware/ftp/anonftp/pub/cygwin/x86_64/release/python-odf
> /sourceware/ftp/anonftp/pub/cygwin/x86/release/python-odf
> 
> I assume the modification should propagate at the next full rebuild 
> (that I can not trigger)
> 
> I plan tomorrow to deploy almost all python packages
> 
> 

it seems to have worked

Regards
Marco

Re: python-odf

[Jon Turney]

On 22/12/2021 13:17, Marco Atzeri wrote:
> On 20.12.2021 23:14, Marco Atzeri wrote:
>> Hi Jon,
>> while updating python-odf for the 3.9 round I noticed that for 1.4.1-1
>> I wrongly set a arch package instead of noarch as was for 0.9.5-1.
>>
>> To solve the issue I copied the content of
>>
>> /sourceware/ftp/anonftp/pub/cygwin/x86_64/release/python-odf
>> on
>> /sourceware/ftp/anonftp/pub/cygwin/noarch/release/python-odf
>>
>>
>> and removed the
>>
>> /sourceware/ftp/anonftp/pub/cygwin/x86_64/release/python-odf
>> /sourceware/ftp/anonftp/pub/cygwin/x86/release/python-odf
>>
>> I assume the modification should propagate at the next full rebuild 
>> (that I can not trigger)
>>
>> I plan tomorrow to deploy almost all python packages
>>
> it seems to have worked
Yes, this is fine.

It's perhaps slightly non-ideal because it changes the published hash 
for the x86 python-odf 0.9.5-1 package, but I don't think this actually 
causes any problems.

(setup doesn't know the package has changed, so doesn't reinstall it, 
but that's not a problem as the contents are the same. There's perhaps 
some room for a setup feature where it records the hash of installed 
packages so it can notice this situation.)

(There used to be issues where setup would notice the a package in the 
package download cache with a mismatched hash and stop, but I think that 
is no longer the case and setup now evicts such packages from the 
package download cache.  But I don't think that applies in this case 
because a different download path will cause this package to be located 
in a different place in the package download cache)