git-missing: do proper argument parsing. Fix #562

[nicolaiskogheim]

Ref #562

This is a rewrite of git-missing to make it behave like the docs says it will.

I have tested by running these commands, which should be enough:

$ git missing feature
$ git missing master feature
$ git missing --no-merges feature
$ git missing --no-merges master feature
$ git missing --no-merges master feature --no-merges

That last one may seem stupid, but argument parsing, and especially passing arguments along to other commands in bash can be very tricky, so that one is important. git log doesn't care about duplicate arguments, so that's alright.

Only tested on OSX, with bash 4.3.46(1)-release.
I don't think there is anything that would make this crash on other systems, but I will test it on FreeBsd and Arch tomorrow.

---

Possible future enhancements:

git-missing will get confused with a call looking like this: git missing --option value firstbranch secondbranch. I don't know if git log accepts any argument on the form --arg value. I scanned man git-log and it doesn't look like it, but git log understands hundreds of different arguments it looks like, so I may be wrong.

[nuragic]

Thanks so much for this one!

[qw3rtman]

Tested and working on FreeBSD 10.3 with GNU bash, version 4.3.46(1)-release (amd64-portbld-freebsd10.1) and git version 2.9.0. I also tested git missing master feature --no-merges and git missing feature --no-merges.

[nicolaiskogheim]

Thanks, @qw3rtman.

---

I found something that will break this, and that is the -L <start>,<end>:<file>, -L :<funcname>:<file> argument. I'll consider redoing the argument loop to tackle that later.

[nuragic]

@nicolaiskogheim hi! now that this change is on master, it will be automatically available in the latest release... ?

[nicolaiskogheim]

No, sorry. We will have to do a release first. When that will be I don't know.
@hemanth When do you think we'll be able to cut a new release?

[hemanth]

After #534 ?

[nicolaiskogheim]

Thanks, forgot about that one.

@nuragic We are waiting to resolve #532, which is partly blocked while waiting for a reply by the author of that issue. But after that, it shouldn't be too long.

[nuragic]

Hi @nicolaiskogheim,

Sorry in advance if I sound pedantic but, looking at that PR (#532) there are no news since June..

Wondering if you guys can just do a patch release...? (v4.1.1) 😉

[nicolaiskogheim]

Yeah, you're right, it has been a long time. We do have a lot of stuff that should be released.

@spacewander @hemanth @qw3rtman

[qw3rtman]

@nicolaiskogheim, made a new issue for further discussion: #579.