Fix GDB pretty-printer for tuples and pointers

[gentoo90]

Names of children should not be the same, because GDB uses them to distinguish the children.

Before	After

main.rs

enumTest{Zero,One(i32),Two(i32,String),Three(i32,String,Vec<String>),}fnmain(){let tuple = (1,2,"Asdfgh");let zero = Test::Zero;let one = Test::One(10);let two = Test::Two(42,"Qwerty".to_owned());let three = Test::Three(9000,"Zxcvbn".to_owned(),vec!["lorem".to_owned(),"ipsum".to_owned(),"dolor".to_owned()]);println!("");// breakpoint here}

launch.json

{
"version": "0.2.0",
"configurations": [
{
"type": "gdb",
"request": "launch",
"gdbpath": "rust-gdb",
"name": "Launch Program",
"valuesFormatting": "prettyPrinters", //this requires plugin Native Debug >= 0.20.0"target": "./target/debug/test_pretty_printers",
"cwd": "${workspaceRoot}"
}
]
}

[rust-highfive]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @alexcrichton (or someone else) soon.

If any changes to this PR are deemed necessary, please add them as extra commits. This ensures that the reviewer can see what has changed since they last reviewed the code. Due to the way GitHub handles out-of-date commits, this should also make it reasonably obvious what issues have or haven't been addressed. Large or tricky changes may require several passes of review and changes.

Please see the contribution instructions for more information.

[carols10cents]

Thanks for the PR! We’ll periodically check in on it to make sure that @alexcrichton or someone else from the team reviews it soon.

[gentoo90]

Fixed exception in pointers parsing

Before	After

Code:

fnmain(){let opt_s = Some("AAAA");let opt_i = Some(42);println!("");// breakpoint here}

[alexcrichton]

r? @michaelwoerister

( knows more about debuginfo and pretty printers than I )

This looks great though!

[gentoo90]

str() failed on things like let opt_s = Some("I \u{2665} Unicode");

[michaelwoerister]

Thanks a lot for the PR, @gentoo90! I would be great if you could add test cases to the src/test/debuginfo suite so those fixes don't break again in the future. Let me know if you need any help with that.

[gentoo90]

Hi @michaelwoerister, is there any way to run a single test using the nightly binary build from rustup, without fetching all the submodules or at least without compiling everything in the repository?

[michaelwoerister]

You could try ./x.py test --stage 0 src/test/debuginfo. That should spare you as much compiling as possible.

[gentoo90]

Most of the tests are ignored, accept gdb-pretty-struct-and-enums-pre-gdb-7-7.rs. Is this OK? Is this because I don't have lldb installed?

[michaelwoerister]

You should not need LLDB for the GDB based tests. I assume this is on Linux? What's the GDB version you are using?

[gentoo90]

Yes, gentoo linux, gdb 7.12

[gentoo90]

I've found another issue while trying to debug bootstrap. When it's built as a part of cargo workspace, it doesn't have info about its source files. So when I set a breakpoint in src/bootstrap/bin/main.rs, gdb says No source file named /home/user/rust/src/bootstrap/bin/main.rs.

But if I remove src/Cargo.toml and build bootstrap with cargo build, the breakpoint works just fine.

[michaelwoerister]

Most of the tests are ignored...

I just tested with GDB 7.12 on Ubuntu and tests are not ignored for me. If you remove all *.stamp files from build/x86_64-unknown-linux-gnu/test/debuginfo, are tests still ignored during the first run of x.py test --stage 0 src/test/debuginfo? I get 13 ignored tests in that case, which should be fine.

I've found another issue while trying to debug bootstrap.

Would you mind opening an issue about that?

[gentoo90]

A proper way to test tuple field names would be

// gdbr-command:interpreter-exec mi2 "-enable-pretty-printing"// gdbr-check:^done// gdbr-command:interpreter-exec mi2 "-var-create noPadding8 @ noPadding8"// gdbr-check:^done,name="noPadding8",numchild="2",value="{...}",type="(i8, u8)",thread-id="1",has_more="0"// gdbr-command:interpreter-exec mi2 "-var-list-children noPadding8"// gdbr-check:^done,numchild="2",displayhint="array",children=[child={name="noPadding8.0",exp="0",numchild="0",value="-100",type="i8",thread-id="1"},child={name="noPadding8.1",exp="1",numchild="0",value="100",type="u8",thread-id="1"}],has_more="0"

but -enable-pretty-printing doesn't do anything in batch mode and pretty-printers are not used :(. Looks like a bug in GDB.

[gentoo90]

Added separate pretty-printer for empty structs, which prints the struct type instead of {...} in variable objects mode.

Before	After

Testing against variable objects is impossible again without -enable-pretty-printing, and CLI interface has the tests already.

[arielb1]

r? @michaelwoerister

[michaelwoerister]

We should try making this gdb-check instead of just gdbr-check, so both kinds of GDB are tested.

[michaelwoerister]

Ping @gentoo90, regarding #42278 (comment). Could you make these regular gdb-commands and gdb-checks? Otherwise we'll only test versions of GDB that already support Rust natively, while the pretty printers should also work for older GDB versions.

[gentoo90]

Sorry for delay.

older GDB versions

ahh, so that's why gdbg-* stuff isn't runned on my system.

[michaelwoerister]

Sorry for delay.

No worries. Thanks for continuing to work on this!

ahh, so that's why gdbg-* stuff isn't runned on my system.

Yes, that seems like a likely reason. I'm actually not sure what version of GDB is on our test infrastructure.

[michaelwoerister]

@bors r+

Let's give it a try.

[bors]

📌 Commit 63076dd has been approved by michaelwoerister

[bors]

⌛ Testing commit 63076dd with merge 3d5b8c6...

[bors]

☀️ Test successful - status-appveyor, status-travis
Approved by: michaelwoerister
Pushing 3d5b8c6 to master...

[michaelwoerister]

🎉
Thanks, @gentoo90!