Optionally propagate the exit code

[hnykda]

Hi,

I added an --exit-code option which tries to propagate the exit code of the underlying program execution. mprof is just silently eating it, causing some unexpected behavior when used in e.g. scripts, containers, ...

It is backward compatible in a way that you need to pass --exit-code|-E option in the run option.

One disruptive change is that I added logger from the standard logging library + added the basicConfig. This is not necessary but IMO is a very good standard in the python world. print are not respecting formatters etc., I consider their usage wrong.

[astrojuanlu]

Left only one comment. Before merging this though, I would like to know what specific use case do you have in mind, because measuring the memory of a script that fails does not make much sense to me.

Regarding prints: I think they are common practice when displaying help messages for CLI interfaces, but I did not check with other projects and I might be wrong. What about pip, twine, sphinx-quickstart?

[fabianp]

thanks @hnykda for the contribution. Changes look good to me

[hnykda]

@Juanlu001 We are using mprof for long-running processing job, where failures are common (e.g. dataset couldn't be processed or the dynamic data got from a 3rd party malformed). So it's part of a longer pipeline. We collect these log files and can historically compare if we are getting better or worse with memory consumption.

Wrt. prints - I didn't thought about that this way 🤔 . It probably make sense then. Should I remove the logger?

[fabianp]

Regarding printing vs logging, I think logging adds some flexibility. I don't have a use for it but also don't see any reason not to use it. Hence I'm fine with it.

[fabianp]

@Juanlu001 : are you OK with the changes in this PR?

[astrojuanlu]

LGTM!

[hnykda]

Thanks! Is there any planed release?

[fabianp]

I just pushed a new release (0.55) to pypi. Thanks!

…

On Fri, Dec 14, 2018 at 3:26 AM Daniel Hnyk ***@***.***> wrote: Thanks! Is there any planed release? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#220 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAQ8h1x7cCQnGUc33me6uTEEkoxWVqhOks5u42C4gaJpZM4ZRW_m> .