Extend HTTPError to satisfy the Go 1.13 error wrapper interface

[flimzy]

I don't see that this has been discussed before in the issues, so my apologies if this has been discussed and rejected for some reason.

[codecov]

Codecov Report

Merging #1559 into master will increase coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #1559      +/-   ##
==========================================
+ Coverage   84.94%   84.96%   +0.01%     
==========================================
  Files          28       28              
  Lines        2166     2168       +2     
==========================================
+ Hits         1840     1842       +2     
  Misses        211      211              
  Partials      115      115              

Impacted Files	Coverage Δ
echo.go	85.93% <100.00%> (+0.08%)	⬆️

---

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 c08f303...ea34bf9. Read the comment docs.

[flimzy]

- Misses        211      213       +2     

I'm happy to add a test for this if it's deemed appropriate.

[lammel]

@flimzy Could you please add a test for unwrap. This will make coverage happy and ensure it will be kept in the future.

Thanks for the contribution.

[flimzy]

@flimzy Could you please add a test for unwrap. This will make coverage happy and ensure it will be kept in the future.

Yep, no problem.

I've pushed an additional commit with tests. Let me know if I can change anything else here.

Note: The tests only run in Go 1.13 or later. I could run tests on older versions of Go using the golang.org/x/xerrors package, but unless you think that's especially valuable, I didn't want to add another dependency just for a simple test.

[lammel]

Referencing #1472 as a duplicate.

This PR looks clean for building also on older versions of go

[flimzy]

That PR's test doesn't directly validate that the error works with Go 1.13, rather it infers it. But I think that's acceptable, since the correctness can be easily validated visually.

[ffenix113]

Is there any ETA on when(if) this would be merged?