[Debian-science-sagemath] pari-sage or pari ?

[Ximin Luo]

jdemeyer at cage.ugent.be:
> 
>> If you wanted to make it easier for us (and other OS distributions), you can generally put more emphasis on upstreaming your patches more quickly
> 
> You seem to think that we have any control over upstream. We can only propose a patch to upstream andere hope they accept it...
> 

No, I don't think you "control upstream", and the fact you're saying this, shows that you don't understand what it means to work in the broader FOSS community. The whole point is that you *do not* have control over many of your dependencies. Gradually one learns how to work *with* this fact instead of *against* it.

As a concrete example: you *do* have control over what code Sage accepts. So, as I suggested below, you can delay accepting certain patches into Sage, until upstream patches are applied - especially minor ones (3 lines) that have a major outside effect (~15 failed tests in Sage) - and encourage or require Sage contributors to first upstream their patches. Or, you write your code against the current bad upstream API, and add a TODO and a description of how to change it to use the good one, when your patch finally gets accepted.

These types of things are all what FOSS projects are expected to do, what they are supposed to do. Yes they carry a slight cost - that is the cost of interoperation, as I mentioned before. And once you practice doing this workflow, the cost reduces further.

I can appreciate that perhaps releases of academic software do not follow the same standards as other FOSS software. But in this case, it's even more important to push your patches upstream, to benefit their users.

>> For example, I don't understand why this patch [1] has not yet been accepted. All you have to do is write some unit tests.
> 
> That is maybe the only example where Sage is to blame that a patch is not accepted upstream.
> 

Don't be so sure. I found this one yesterday evening and it took our test failures down from 319 to 303. There are still plenty left.

X

-- 
GPG: ed25519/56034877E1F87C35
GPG: rsa4096/1318EFAC5FBBDBCE
https://github.com/infinity0/pubkeys.git