Revert usage of cmd on Windows

[axel3rd]

cf. #17 (comments from May 2018): Revert the usage of cmd.exe (on Windows), because prevents the destroy/kill launched by this way when CTRL+C.

For history: This Windows shell specific is removed since plexus-utils 3.0.15, except for 3.1.0.

On Windows, the usage of cmd builtin (like echo) or .cmd / .bat on PATH should be implemented by the client.

[axel3rd]

Travis oraclejdk7 support seems removed (see here).
Could be done in // PR if not relevant here.

[michael-o]

Please move the OpenJDK 7 commit to a separate PR.

[axel3rd]

About openjdk7: already in issue-fix branch (commit: 54f6ca0d9c7d0940f84610788f3e55c2548dc5c4).

[axel3rd]

Other PR #42, to propose an approach for cmd usage if wanted (consensus ? hardcode a cmd.exe /X /C for a simple echo disturb me 😜).

[mkarg]

@axel3rd Please fix conflicts. Thanks.

[axel3rd]

@axel3rd Please fix the conflicts. Thanks.

@mkarg : Sorry, not able todo that in this PR, because is >2 years old so I have deleted the fork since this time.

[michael-o]

@axel3rd Please fix the conflicts. Thanks.

@mkarg : Sorry, not able todo that in this PR, because is >2 years old so I have deleted the fork since this time.

That's ok. @mkarg Honestly, I don't see a reason to have a shell wrapper at all, but use ProcessBuilder. This worked for me for years. Those wrappers only cause confusion and issues which I have to fix in Maven SCM.

[mkarg]

@axel3rd Please fix the conflicts. Thanks.
@mkarg : Sorry, not able todo that in this PR, because is >2 years old so I have deleted the fork since this time.
That's ok. @mkarg Honestly, I don't see a reason to have a shell wrapper at all, but use ProcessBuilder. This worked for me for years. Those wrappers only cause confusion and issues which I have to fix in Maven SCM.

My target is to cleaup the open PRs, so either we merge or drop them or turn them into drafts, but not keep them open for
review for more years in an unclear state. To reach that goal, first the original authors need to resolve conflicts to allow reviews ontop master. I am looking into the actual change once the code at least will compile again.

[axel3rd]

My target is to cleaup the open PRs, [...] not keep them open for review for more years in an unclear state

Fully agree with that 👍

I don't see a reason to have a shell wrapper at all, but use ProcessBuilder

I understand the content of this PR is not really relevant, so closing.

[michael-o]

My target is to cleaup the open PRs, [...] not keep them open for review for more years in an unclear state

Fully agree with that 👍

I don't see a reason to have a shell wrapper at all, but use ProcessBuilder

I understand the content of this PR is not really relevant, so closing.

I didn't say that. What I am saying is that our concept is wrong from begin with.

[axel3rd]

I didn't say that. What I am saying is that our concept is wrong from begin with.

Ok, sorry, misunderstanding from my side.

I will provide some new compiling PR for #41 & #42 and not delete the fork until status, allowing review.

[michael-o]

I didn't say that. What I am saying is that our concept is wrong from begin with.

Ok, sorry, misunderstanding from my side.

I will provide some new compiling PR for #41 & #42 and not delete the fork until status, allowing review.

Thanks! Take your time! No need to hurry.

[axel3rd]

See #109 for following.