[java] Copy SM binary to cache folder and use it from there (#11359)

[bonigarcia]

Description

This PR copies the SM binary from the distribution (jar) to the cache folder (`~/.cache/selenium), which is invoked from then.

This PR has two commits:

1. The first commit assumes that the cache path is always the path (`~/.cache/selenium), which makes things easier.
2. The "problem" is that the cache folder can be configured in Selenium Manager with other mechanisms: the configuration file (se-config.toml -if we finally renamed that way-) or environmental variables (in particular, SE_CACHE_PATH is the one for the cache path). The second commit allows to read also the cache path from the configuration file (first) and the env (second).

Motivation and Context

Resolves #11359.

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)

Checklist

• I have read the contributing document.
• My change requires a change to the documentation.
• I have updated the documentation accordingly.
• I have added tests to cover my changes.
• All new and existing tests passed.

[codecov-commenter]

Codecov Report

All modified lines are covered by tests ✅

Comparison is base (d7dd881) 56.49% compared to head (9cb0041) 56.49%.

❗ Current head 9cb0041 differs from pull request most recent head e6f018e. Consider uploading reports for the commit e6f018e to get more accurate results

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@           Coverage Diff           @@
##            trunk   #12539   +/-   ##
=======================================
  Coverage   56.49%   56.49%           
=======================================
  Files          86       86           
  Lines        5255     5255           
  Branches      187      187           
=======================================
  Hits         2969     2969           
  Misses       2099     2099           
  Partials      187      187           

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

[titusfortner]

I really don't want the bindings to process the toml file. The goal is to move the extra processing out of the bindings, not create more.

[bonigarcia]

For that reason, I sent two commits in this PR. If the bindings do not want to process the config file (or the envs), the first commit should be enough.

[diemol]

I agree, only env vars I would say. No TOML config read for now.

Also, this class should give you the Selenium version

selenium/java/src/org/openqa/selenium/BuildInfo.java

Line 26 in 43eb1e5

public class BuildInfo {

[bonigarcia]

I have just sent two additional commits to this PR with the requested changes (read env but not config file and use of BuildInfo to get the Selenium version instead of a constant).

[titusfortner]

The Java & Ruby implementations are slightly different and we need to make sure we have what we want, and then implement in the others.

Let's prioritize this as primary goal for 4.13

[titusfortner]

I have no idea what we did that fixed it, but thanks for letting us know.

[diemol]

It seems that browsers are left open, which could be the consequence of a bug in your code.

Can you please open a new issue and fill out all the information required in the template? With that, we can reproduce the issue.

[diemol]

We can only troubleshoot your issue if there is a way to reproduce it.

[titusfortner]

How does this work with when the cache location is not writeable? Do we need to fall back to streaming it?

[bonigarcia]

Yes, that case was not considered, so it will fail when the cache is not writable in the current implementation. Let me review this PR to consider that case.

[bonigarcia]

I've added a new patch to this PR to consider when the cache path is not writable. In that case, the Selenium Manager binary will be extracted in a temporal folder and used from there. That binary is not deleted using the original shutdown hook to avoid the problem related to grid (see #12802).

[titusfortner]

Will that potentially result in a bunch of Selenium Manager binaries in temp files? (but only if cache directory isn't writeable)

[bonigarcia]

No problem. I added a new commit to restore the shutdown hook to delete the SM binary only when it is in the temporal folder.