erfcinv(::BigFloat): switch tolerance to eps(y) instead of eps(x)

[lrnv]

Fixes #557

It took me few attempts but i managed to do it. I do not know why this tol was set as eps(2abs(x)), and whether my proposal is sound, but at least it solves the issue.

[lrnv]

I am closing because it worked locally on my machine, but it looks like it hangs and hangs on the CI, i do not want to consume 10* 6h of CI for JuliaMath and cannot stop teh Ci myself.

[giordano]

Closing a PR doesn't cancel CI jobs (that's just how GitHub Actions works). Perhaps it'd be a good start to open a PR to set a reasonable timeout these workflows

[lrnv]

Did not know that, thanks. I added a temporary timeout to the workflows.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 94.67%. Comparing base (b2a7190) to head (f684fc4).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #558   +/-   ##
=======================================
  Coverage   94.67%   94.67%           
=======================================
  Files          14       14           
  Lines        3023     3023           
=======================================
  Hits         2862     2862           
  Misses        161      161           

Flag	Coverage Δ
unittests	94.67% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[devmotion]

I opened #559 with a different fix based on my understanding of the issue.

[lrnv]

I knew someone will know more than me ;) Thanks @devmotion