Avoid silently overflowing (signed) ints to unsigned int values in args

[alexsnaps]

Follow up to #3303

Also, I think I should add a commit to handle the formatting around { on the same line as their matching… if statements? All statements? What is the formatting preference? Should I bring consistency to the entire file? See #3303 (comment)

Finally, I think c429d54 probably warrants a CHANGELOG entry, wdyt?

$ ./helloworld --ponythreads=0
--ponythreads can't be less than 1
$ echo $?
255

[alexsnaps]

Should this be 0 or 1 as the min value? 0 (or unspecified) used to mean physical core count… but did people really use 0 explicitly to mean that?

I like to hope not, and I think this is a more explicit behavior.

[SeanTAllen]

So we can scale down to 0 scheduler threads which means that minthreads is ok to be set to 0.

ponythreads with scaling on is really a maxthreads.

So, I think 1 as probably because, at 0, nothing can ever happen. we need to be able to do at least 1.

[SeanTAllen]

This does raise an interesting question. Should --ponythreads be renamed to --ponymaxthreads given that we now do scaling. I think there's a reasonable argument to be made for that change.

[SeanTAllen]

@alexsnaps I believe this is correct and ready to be merged, yes?

[alexsnaps]

Two questions:

• Do we want to add a changelog entry?
• What about the formatting? See Codify the pony(min)threads behavior #3303 (comment)

[SeanTAllen]

Don't worry about formatting for now.

CHANGELOG.... yeah I think an entry would be good.

[alexsnaps]

Added — I think this is good for merge.

[SeanTAllen]

Thanks, @alexsnaps. Can you add release notes to the issue for the next release?