updated template asv.conf.json to Python 3.8

[orbeckst]

Fixes #3430

Changes made in this Pull Request:

• updated matrix from 2.7 to 3.8
• automated benchmarks were updated (see update benchmark env to > 3.7 benchmarks#11)

PR Checklist

• n/a Tests?
• n/a Docs?
• n/a CHANGELOG updated?
• Issue raised/referenced?

[hmacdope]

LGTM pending CI :)

[orbeckst]

Thanks for the quick review. Could you merge it when ready, please? Thanks!

[IAlibay]

Cancelling CI and restarting - something off with the docs build, seeing if it's intermittent.

[codecov]

Codecov Report

Merging #3521 (215df1b) into develop (01c9dec) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff            @@
##           develop    #3521   +/-   ##
========================================
  Coverage    94.06%   94.06%           
========================================
  Files          176      176           
  Lines        23288    23288           
  Branches      3305     3305           
========================================
  Hits         21905    21905           
  Misses        1337     1337           
  Partials        46       46           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 01c9dec...215df1b. Read the comment docs.

[IAlibay]

oof something is going wrong with CI... looks like runners are having a field day but gh status is being fine. Hold for re-run in a few hours?

[hmacdope]

re-triggered

[IAlibay]

Still failing on docs - let me check if there's been a recent update to sphinx related things.

[IAlibay]

This seems to be the major difference between the last passing CI run and the current failure:

"intersphinx inventory 'https://docs.scipy.org/doc/numpy/objects.inv' not readable due to ValueError: unknown or unsupported inventory version: ValueError('invalid inventory header: ')"

looks like an issue with the scipy website (https://docs.scipy.org/ - none of the numpy links resolve)

[IAlibay]

Possibly linked to scipy/scipy#15545

@tylerjereddy seems to be dealing with it scipy/scipy#15545 (comment)

[hmacdope]

Can we safely say its upstream? Or shall we wait for it to be fixed, re-run then merge.

[IAlibay]

Can we safely say its upstream? Or shall we wait for it to be fixed, re-run then merge.

Is this PR in a rush to make it to develop? If not I'd say hold until upstream fixes, we don't have any other simple PRs hanging around to check if it's still failing.

[IAlibay]

Can we safely say its upstream? Or shall we wait for it to be fixed, re-run then merge.

Is this PR in a rush to make it to develop? If not I'd say hold until upstream fixes, we don't have any other simple PRs hanging around to check if it's still failing.

[IAlibay]

sorry - hit close instead of comment for some reason...

[hmacdope]

Nah no rush

[IAlibay]

Let's give it a day or so - if Tyler comes back and says it's a long term issue then we can either fix or ignore.

[tylerjereddy]

Well, I tried to help fix it upstream, but the intersphinx stuff is tricky. I got a few suggestions, but if we have solid web server/developer folks see: tylerjereddy/htaccess_backup#1

[tylerjereddy]

The root of the problem is that SciPy has a new docs theme, but it was never tested in a release deployment, so for a while we just didn't publish the release docs b/c of the damage to intersphinx links, but at this point we either try to fix it or we'll never fix it..

[IAlibay]

should have been fixed with #3522, let's see what CI returns like.