Upgrade to time 0.3.2

[tyhi]

Type tests for both postgres and mysql completed without any errors.

Feedback is very much appreciated.

Closes #1277

[paolobarbolini]

Looks good, just a few minor tweaks

[abonander]

I would prefer to keep support for time 0.2 as an optional feature, since it usually takes the whole ecosystem a while to upgrade with these kinds of things. We can use the renaming scheme as shown here so we can have both versions available to users: #1277 (comment)

Now the question becomes, which version keeps the name time and which gets renamed to time-02/time-03? I guess it makes more sense to have the old version be time-02 and the current version be time, but what happens when time upgrades to 0.4? Do we just do this whole dance all over again?

I think if we have both time-02 and time-03 then there's no ambiguity and we can add time-04 backwards-compatibly later on, and drop support for old versions as the ecosystem catches up.

When time drops a 1.0, I think that'll be the safe time to drop the version suffix.

Wouldn't mind having @mehcode weigh in here though.

[abonander]

Conferred a bit with @mehcode and @danielakhterov and we think we want to do this:

• Keep time = "0.2" but rename to time-02
• Add time = "0.3" as time-03
• Keep the time feature and have it enable the time-02 feature:

[dependencies]
time-02 = { package = "time", version = "0.2", optional = true }
time-03 = { package = "time", version = "0.3", optional = true, features = ["macros", "parsing"] }

[features]
# NOTE: deprecated; use `time-02` instead
time = ["time-02"]

And ideally, somehow add a deprecated warning if the time feature is enabled.

Then drop the unsuffixed form as a feature in sqlx 0.6, to be brought back iff time releases a 1.0 with a guarantee to support it for some time like with Tokio 1.0, because I don't want to have to do this again if we foresee having to concurrently support time = "1.0" and time = "2.0"; if that's the case then we'll probably just do time-1 and time-2 or something.

[abonander]

@tyhi did you want to make those changes in this PR? I saw you gave my comment a 👍 and then force pushed your branch but I don't see anything significantly different than before.

[tyhi]

I had force pushed to get ready to hard reset my pr and apply time-03 over it however I was getting some cargo errors once I had introduced the time-03 feature along with some import errors with macros and others. I looked into it for a little but haven't had time recently.

I agree with what you said above and looking at the crate.io numbers it would look like there's barely any use in 0.3 currently, and it may just end up being annoying at the least to maintain now and possibly in the future if something like this happens again like you stated.

I think it may be best to close this PR since I'm not sure when I'll be able to finish this currently, among other things.