Too many issues

[nhartland]

We have a lot of duplicate issues or issues about things that may have been solved.
We could do with going through them.

For example, I'm pretty sure
#34 = #44
#178 = #176 = #14
#26 = #21 = #186

Just from a quick glance, there are probably more

[nhartland]

#38 = #25

#77 = #22

[scarrazza]

I agree, thanks for collecting this.

[nhartland]

I think #37 and #24 may have been solved already

[Zaharid]

Thanks.

• postfit.py check_replica_files not robust/error message not helpful #176 could use a patch in postfit to make things slightly better.
• Improve handling of random seeds #77 and Random states (C++) #22 are for python and cpp respectively and somewhat different in what they are trying to achieve. The python one is about having random states that are uncorrelated, used controllable and reproducible. The cpp one is about having basic infrastructure such as numpy.random.RandomState
• I haven't seen a patch for Segfault in destrucor of copy constructor in lhapdf #24.

Ok for the rest. Reading the covmat business.

[nhartland]

I had guessed #24 had been solved by that LHAPDF config bug. Is it a separate issue then?

[Zaharid]

That's a bug in the nnpdf side. The code example still segfaults on master.

[Zaharid]

In fact it even says what the problem is

Yes, sine we don't define it, it's copying fLHA by pointer, and then freeing it two times.

[nhartland]

There is a bit of overlap between #29 and #95, in that coverage reports are also mentioned in #95.
Other than that I don't see much else to be removed. Which is a bit distressing.