fix: handle trailing slashes in block explorer URLs

[bguiz]

What

• modify ui/app/helpers/utils/transactions.util.js and
  ui/lib/account-link.js to strip trailing slashes
  if they are present.
• added relevant tests not just for the new scenario,
  but also the general scenarios for these functions,
  as there previously was no test coverage for these
  two functions.

Why

• Current behaviour, when user enters a block explorer URL
  when configuring a custom RPC, and that block explorer URL
  contains a trailing /.
  • e.g. https://block.explorer/
  • this results in a double-slash (//) in the transaction
    and account URLs generated by MetaMask.
  • e.g. https://block.explorer/tx/0xabcd...,
    https://block.explorer/account/0xabcd...
  • This needs to be handled using a router redirect
    on the server of the block explorer,
    and this changes would avoid that requirement.

[rekmarks]

Neat work, thanks!

Please resolve the lint failures and push again. yarn lint:fix or yarn lint:changed:fix (quicker, if you do a soft reset), should do the trick.

[bguiz]

Neat work, thanks!

Please resolve the lint failures and push again. yarn lint:fix or yarn lint:changed:fix (quicker, if you do a soft reset), should do the trick.

@rekmarks done, looks like the various CI lint tasks have all passed now. Earlier on ci/circleci: test-e2e-firefox appears to have failed as well, but a forced-push without any changes appears to have resolved it, so perhaps that test is fragile/ non-deterministic.

https://circleci.com/gh/MetaMask/metamask-extension/164993

[2020-05-14 02:31:56.620][e2e]   1) MetaMask
[2020-05-14 02:31:56.620][e2e]        Going through the first time flow
[2020-05-14 02:31:56.620][e2e]          clicks the continue button on the welcome screen:
[2020-05-14 02:31:56.620][e2e]      ElementClickInterceptedError: Element <button class="button btn-primary first-time-flow__button"> is not clickable at point (154,509) because another element <div class="loading-overlay"> obscures it

[whymarrh]

Yeah, our e2e tests aren’t especially reliable, currently 🙃

[bguiz]

Yeah, our e2e tests aren’t especially reliable, currently

@whymarrh yeah, well one more failed test-e2e-Firefox, same error, but I manually triggered the build by doing another forced push. All checks are passing now!

[rekmarks]

LGTM!

@bguiz, this is a lovely example of an external PR. I really appreciate the care you put into it.

[bguiz]

LGTM!

@bguiz, this is a lovely example of an external PR. I really appreciate the care you put into it.

You're welcome @rekmarks - and thanks for the comment!