WindowsProcess printStackTrace

[benor1470]

Hey, the WindowsProcess uses printStackTrace (lines 173,220,249,303,332,392) and not using logger, can you please use a logger so i can turn in off when they aren't needed? My log shows those errors on tests and without a way to turn them off they fill my log

Thank you

[brettwooldridge]

Pull requests welcomed.

[benor1470]

Then a pull request was created

[fanf]

I have the same problem with BasePosixProcess. I have two concerns:

• I can't adapt (and potentially recover at the application logic level) from the error that appears in the stacktrace, and which is not always an exception from the application logic
• I have non-contextualise stacktrace. Typically, when I get "Invocation of posix_spawn() failed, return code: 13, last error: 13", I would prefer to tell my user to look at the command permission (with the path to the command), but never ever display that message to them.

To not depend to a given log framework, and idea could be to add a PrinStream to the process builder, and use that one in place of System.err if provided. Users who care would be able to provide a byte array output stream to post process the stream in case of exit code Int.MinValue, and chose what to actually disaply to users using their application logging framework.

Do you think the idea can fly ?

[bturner]

In the years since this was last updated, BasePosixProcess has already been changed to replace e.printStackTrace with use of the JDK's java.util.logging.Logger, and 1.2.x and 2.0.x releases include that change. I've recently opened pull request #118 which applies a similar treatment to WindowsProcess. Hopefully that can be merged/released soon.

[bturner]

My changes from pull request #118 were included in NuProcess 2.0.2, eliminating the last e.printStackTrace calls from the codebase.