Removing old_run module.

[dhermes]

Has been deprecated since 700f13f which occurred on July 2, 2013.

@craigcitro Do you see any reason not to remove this module?

(I debated getting 100% coverage for it in #212 and this seemed like an easier path forward)

[craigcitro]

I suspect a number of people still use it.

[dhermes]

OK @nathanielmanistaatgoogle shall we

• Remove it anyway and instruct those people to upgrade to using run_flow (we could even provide a snippet that would show them how to use run_flow with gflags without having to provide the old_run module)
• Un-deprecate the old_run module
• Do nothing but make sure it gets 100% test coverage for Get to 100% code coverage #212

[dhermes]

@craigcitro I assume you mean internal to Google yes?

[craigcitro]

i'm sure there are some in google -- but actually, i suspect there are plenty outside. i wager some folks just silenced or ignore the deprecation warning and have been happy their code is working with no work on their part.

i don't see "we need 100% test coverage" as a good reason to potentially break them?

[dhermes]

@craigcitro That isn't the reason for breaking them. It just came up because I was looking at the module and realized it's been deprecated for 2 years 1 month (this library is only 5 years 4 months old: ed13252)

[nathanielmanistaatgoogle]

I approve this commit and am fine with "we no longer wish to be responsible for this code and have given more than enough advanced warning" as a reason for its removal. @craigcitro would that work for you? If I were to offer "a developer who depends on more-than-two-years-deprecated code is already broken and merely doesn't yet know it" as a philosophical proposition, how would you receive it?

[dhermes]

@nathanielmanistaatgoogle I'll let you pull the trigger on the merge if/when this discussion sorts itself out.

[craigcitro]

i don't have strong feelings either way -- i was just answering danny's question about "safe to delete".

i'd be hard-pressed to believe any philosophical propositions about an auth library.