#991 more robust terminal width detection for international Windows versions

[S-Gaertner]

• changed terminal width detection regex pattern for Windows
• check for process termination in order to fail faster

[codecov-io]

Codecov Report

Merging #993 into master will decrease coverage by 0.03%.
The diff coverage is 50.00%.

@@             Coverage Diff              @@
##             master     #993      +/-   ##
============================================
- Coverage     94.68%   94.65%   -0.04%     
  Complexity      415      415              
============================================
  Files             2        2              
  Lines          6136     6138       +2     
  Branches       1637     1638       +1     
============================================
  Hits           5810     5810              
  Misses           89       89              
- Partials        237      239       +2     

Impacted Files	Coverage Δ	Complexity Δ
src/main/java/picocli/CommandLine.java	94.50% <50.00%> (-0.04%)	281.00 <0.00> (ø)

---

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 bb0019f...281156a. Read the comment docs.

[remkop]

Looks good. Were you able to test that this prevents the 2 second timeout when the pattern cannot match anything (e.g. with the old pattern on a German language environment)?

[S-Gaertner]

Yes, I manually tested that on my system with the old pattern. The 2 second delay was gone as soon as I introduced the alive test for the thread.

[remkop]

Great, thank you!

[remkop]

Overall looks good. I have some other work in progress but I will try to merge this soon.

Thank you for the contribution!

[S-Gaertner]

You're welcome! Thank you for this great project :)

[remkop]

Can I ask:

• What is it that you like about picocli?
• What would you do if picocli did not exist?

[S-Gaertner]

What I like about picocli:

• It allows complex use cases, yet is pretty easy to use - especially with the convenience methods.
• It has zero dependencies - I am reluctant to include dependencies that themselves entail more dependencies.
• It has superb documentation.
• It is open source, mature, active, feature-rich and has a friendly community.

What I would do without picocli:

• I already have rolled my own command line parser for simple uses cases. Now my use cases grew more complex, so I probably would have adapted it to some degree and also simplified my use cases (and at the same time would provide more user training for the probably more complicated command line invocation).
• I also would have looked into alternatives like JCommander.

[remkop]

@S-Gaertner Great feedback, thank you! 👍