demonstrate RHS != Self use cases for Add and Sub

[matthew-piziak]

No description provided.

[rust-highfive]

r? @aturon

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

[matthew-piziak]

r? @steveklabnik

[steveklabnik]

@bors: r+ rollup

thanks a ton, this is great!

[bors]

📌 Commit 96bd6cf has been approved by steveklabnik

[sophiajt]

@bors r-

Looks like this might be causing a possible issue on Windows. @steveklabnik any idea what might cause:

Linkcheck stage2 (i686-pc-windows-gnu)
core\ops\trait.Add.html:121: broken link - time\struct.SystemTime.html
core\ops\trait.Sub.html:95: broken link - time\struct.SystemTime.html
std\ops\trait.Add.html:121: broken link - time\struct.SystemTime.html
std\ops\trait.Sub.html:95: broken link - time\struct.SystemTime.html
command did not execute successfully: "C:\\bot\\slave\\auto-win-gnu-32-opt-rustbuild\\build\\obj\\build\\i686-pc-windows-gnu\\stage2-tools\\i686-pc-windows-gnu\\release\\linkchecker.exe" "C:\\bot\\slave\\auto-win-gnu-32-opt-rustbuild\\build\\obj\\build\\i686-pc-windows-gnu\\doc"
expected success, got: exit code: 101
Makefile:49: recipe for target 'check' failed
>> rustjob: found 0 remaining processes

[steveklabnik]

Does it need one .. and not two, maybe?

[sophiajt]

@steveklabnik - no idea. Who's a good contact for rustdoc?

[sophiajt]

I wonder if @brson or @alexcrichton might know...

[alexcrichton]

I think the links here may have just one too many ../?

[matthew-piziak]

@alexcrichton Is that the case? If I take https://doc.rust-lang.org/nightly/std/ops/trait.Add.html and add /../../time/struct.SystemTime.html. to the end of it, then it seems to navigate correctly. There's the question of the leading slash, of course. Now, a bigger issue for me is that I'm unable to test docs changes locally (rustdoc fails on multiple machines). If anybody wants to help me debug it on #rust-docs, that would be hugely helpful.

[alexcrichton]

Sometimes the traits are rendered in multiple locations due to inlining in rustdoc, so maybe that's what's happening here?

[matthew-piziak]

I'm having difficulty testing this. When I run cargo docs in src/libcore on nightly I get the following output: https://gist.github.com/matthew-piziak/052e4db5d7dd57fe4ccfc95d05832ef1

[ollie27]

If I take https://doc.rust-lang.org/nightly/std/ops/trait.Add.html and add /../../time/struct.SystemTime.html. to the end of it, then it seems to navigate correctly.

That's not what the browser does. It basically tries to load https://doc.rust-lang.org/nightly/std/ops/../../time/struct.SystemTime.html which of course fails. So yeah, there is one too many ../s.

As far as I know the most reliable way to build the docs for testing is to use make docs see here.

[matthew-piziak]

Thanks @ollie27! Fixed.

[matthew-piziak]

Looks like the build is still being weird here. I don't think it's my fault this time...

[steveklabnik]

Yes, this is not you. Let's give it a shot again on bors.

@bors: r+ rollup

[bors]

📌 Commit 89f7e92 has been approved by steveklabnik

[TimNN]

This still fails linkcheck for me:

core/ops/trait.Sub.html:121: broken link - core/time/struct.SystemTime.html
core/ops/trait.Add.html:121: broken link - core/time/struct.SystemTime.html

[steveklabnik]

@bors: r-

[matthew-piziak]

Anybody have advice on how to run linkcheck locally?

[matthew-piziak]

@matthew-piziak

cd src/tools/linkchecker
cargo run ../../lib_you_want_to_check

[matthew-piziak]

@TimNN I cannot reproduce.

Here is my repro attempt:

• make docs
• open file:///home/matt/rustfork/rust/doc/std/ops/trait.Sub.html in Chromium
• click on std::time::SystemTime link

This navigates to file:///home/matt/rustfork/rust/doc/std/time/struct.SystemTime.html as expected.

[TimNN]

The problem is that Add / Sub are document in both core and std, however time is only in std.

Try the links from file:///home/matt/rustfork/rust/doc/core/ops/trait.Sub.html and you should get a 404.

[matthew-piziak]

Thanks @TimNN!

[steveklabnik]

Travis looks spurious, let's try with bors.

@bors: r+ rollup

[bors]

📌 Commit 9a400f0 has been approved by steveklabnik