Replace deprecated integer suffixes

[alfiedotwtf]

In the following example:

fn main() { 
    println!("{:?}", 42is);
    println!("{:?}", 42us);
} 

The output is:

42i
42u

This should be outputting the following instead:

42is
42us

[rust-highfive]

r? @nikomatsakis

(rust_highfive has picked a reviewer for you, use r? to override)

[alexcrichton]

I think that there's definitely a number of tests which exercise this behavior, can you be sure to run the test suite and make sure there are no test failures as well?

[alfiedotwtf]

Sure, will do once I'm connected to power.

(side note - I was always under the impression that Travis-CI ran the tests before allowing changes to be approved?)

[steveklabnik]

Travis only runs 'make tidy'

[alfiedotwtf]

@alexcrichton sorry. I'm an idiot - completely forgot about tests. I'll remember next time.

@steveklabnik ahh gotcha!

Tests have been updated and all pass, except for 4 Valgrind tests. @Kimundi on IRC said it could be something local on my machine doing this, so if anyone can check it out that would be great.

[alfiedotwtf]

Travis doesn't like this because "line longer than 100 chars". Is this a config bug or enforced rule in Travis rather than a bug in my commit?

[alfiedotwtf]

Got help from irc. I'm splitting the line since line limits are enforced (TIL). Just compiling and re-running make check now.

[alfiedotwtf]

Ignore that last commit. My local is screwy.

[alfiedotwtf]

Awesome...

Not sure that happened with that last commit. Anyhow, 5cc6a06 looks good apart from the 4 Valgrind errors. Also, to get around the 100 char limit, I cheated a bit with src/test/run-fail/assert-eq-macro-panic.rs

[alexcrichton]

@bors: r+ 5cc6a06

Thanks!

[alfiedotwtf]

Last night, it built from clean and all the tests passed (except Valgrind tests). This morning on a difference machine (same commit) it won't build for me as lots of assertions are failing. Can you cancel @bors including this change until I've sorted this out please.

Sorry for the inconvenience.

[alexcrichton]

@bors: r-

Feel free to just ping this whenever you've got the tests passing again.

[alfiedotwtf]

@alexcrichton Ok... so "make clean; make check" done. Everything passed except the same four Valgrind tests.

(I mustn't had done a make clean last night before submitting. I'll remember next time!)

[alfiedotwtf]

Had to redo because make tidy was failing (over 100 chars), but make check was happy? Maybe check should include tidy?

hrmm.. more phantom commits e.g. e7f2280.

Ok, wait till I give the thumbs up. Doing a make clean, check, then tidy now.

[alexcrichton]

Looks like there still may be some extra commits, perhaps a rebase is in order?

[alfiedotwtf]

@alexcrichton It was weird because I rebased onto master...

Anyway, make clean, make check and make tidy. All good besides Valgrind. Thumbs up to 13b800d.

This has been a good learning experience for me. Sorry about giving you guys the run around!

[alexcrichton]

@bors: r+ 13b800d

Thanks!

[tbu-]

@alexcrichton This moves into the wrong direction, a recently accepted RFC says that the fmt::Show output of integer types should not include the integer suffix. Could you r- this?

[alfiedotwtf]

@tbu- Sorry, didn't see that RFC.

So just to confirm, this is for all integer types and not just for is and us?

format!("{:?} {:?} {:?}", 42i32, 43is, 44us) == "42 43 44"

[tbu-]

@Alfie Yes, all integer suffixes are to be omitted.

[alfiedotwtf]

@tbu- cool, thanks for the clarification. Will update this PR.

[alexcrichton]

Ah indeed @tbu-, I ended up taking care of this as part of #21457, so I'll just close this in favor of that. Thanks again @Alfie!