The published version of the package 0.3.3 I believe doesn't fail if I have stderr and stdout which is great for me.
However, some recent changes in the package cause that when stderr is present, an error is shown even if the command succeeded.
This is a problem because I run a command that outputs "warning messages" to stderr, and then the command succeeds but shows an error instead. I believe this behavior is incorrect and I would like this behavior to be reverted back, otherwise I think this should be an option.
I guess one could argue that my command shouldn't be using stderr and I should fix that, but I can also argue that your program should throw an error if it fails rather than you having to rely on stderr. I guess making it an option would be the best for both cases.
The published version of the package 0.3.3 I believe doesn't fail if I have
stderrandstdoutwhich is great for me. However, some recent changes in the package cause that whenstderris present, an error is shown even if the command succeeded.This is a problem because I run a command that outputs "warning messages" to
stderr, and then the command succeeds but shows an error instead. I believe this behavior is incorrect and I would like this behavior to be reverted back, otherwise I think this should be an option.I guess one could argue that my command shouldn't be using stderr and I should fix that, but I can also argue that your program should throw an error if it fails rather than you having to rely on stderr. I guess making it an option would be the best for both cases.