[Debian-science-sagemath] 7.4 with the latest shiny new packages: 4000 failed tests

[Tobias Hansen]

The idea would be that we don't want to spend our time to make sage work
with a new singular version, so we ship 4.8.3-p3 in stretch.

Actually, the epoch "1:" is pretty ugly, because once introduced, one
needs to keep it for all future versions of the package. In the case of
singular there is actually a better way.

$ dpkg --compare-versions 4.8.3p3 gt 4.8.3-p4; then echo 1; else echo 0; fi
1

So you could use the version 4.8.3p3 and it's bigger than 4.8.3-p4.
No need for the epoch here.


On 11/06/2016 10:28 PM, Jerome BENOIT wrote:
>> Yes, it's with singular 4.0.3-p3. That tiny change of version number 
>> hides quite a few changes.
> 
> Sorry, I did not realized it. As p stands for patch, I assumed only micro changes.
> Note the a new reload of 4.0.3-p3 needs to depends on brial . This i a minor
> change any. I guess that the idea idea is to fix first the 75 or error before
> making this micro jump. How can I reload singular 4.8.3p3 as 1:4.0.3-p3 ,
> and what would be the appropriate suffix for 4.8.3p4 within this scheme ?
> 
>  The easiest way to deal with this would
>> be reuploading 4.0.3-p3 as version 1:4.0.3-p3 and sticking with this 
>> version for stretch.
>