thread 'main' panicked at 'expected identifier'

[adetaylor]

Follow instructions in #468 now that #470 is fixed.

Get

  thread 'main' panicked at 'expected identifier', /Users/adetaylor/.cargo/registry/src/github.com-1ecc6299db9ec823/syn-1.0.68/src/parse_quote.rs:96:21
  stack backtrace:
     0: rust_begin_unwind
               at /rustc/42816d61ead7e46d462df997958ccfd514f8c21c/library/std/src/panicking.rs:493:5
     1: std::panicking::begin_panic_fmt
               at /rustc/42816d61ead7e46d462df997958ccfd514f8c21c/library/std/src/panicking.rs:435:5
     2: syn::parse_quote::parse
               at /Users/adetaylor/.cargo/registry/src/github.com-1ecc6299db9ec823/syn-1.0.68/src/parse_quote.rs:96:21
     3: autocxx_engine::conversion::codegen_rs::fun_codegen::gen_function
               at /Users/adetaylor/dev/autocxx/engine/src/conversion/codegen_rs/fun_codegen.rs:133:45
     4: autocxx_engine::conversion::codegen_rs::RsCodeGenerator::generate_rs_for_api
               at /Users/adetaylor/dev/autocxx/engine/src/conversion/codegen_rs/mod.rs:424:17
     5: autocxx_engine::conversion::codegen_rs::RsCodeGenerator::rs_codegen::{{closure}}
               at /Users/adetaylor/dev/autocxx/engine/src/conversion/codegen_rs/mod.rs:156:27
     6: core::iter::adapters::map::map_fold::{{closure}}
               at /Users/adetaylor/.rustup/toolchains/nightly-x86_64-apple-darwin/lib/rustlib/src/rust/library/core/src/iter/adapters/map.rs:82:28
     7: core::iter::traits::iterator::Iterator::fold
               at /Users/adetaylor/.rustup/toolchains/nightly-x86_64-apple-darwin/lib/rustlib/src/rust/library/core/src/iter/traits/iterator.rs:2147:21
     8: <core::iter::adapters::map::Map<I,F> as core::iter::traits::iterator::Iterator>::fold
               at /Users/adetaylor/.rustup/toolchains/nightly-x86_64-apple-darwin/lib/rustlib/src/rust/library/core/src/iter/adapters/map.rs:122:9
     9: core::iter::traits::iterator::Iterator::unzip
               at /Users/adetaylor/.rustup/toolchains/nightly-x86_64-apple-darwin/lib/rustlib/src/rust/library/core/src/iter/traits/iterator.rs:2821:9
    10: autocxx_engine::conversion::codegen_rs::RsCodeGenerator::rs_codegen
               at /Users/adetaylor/dev/autocxx/engine/src/conversion/codegen_rs/mod.rs:152:91
    11: autocxx_engine::conversion::codegen_rs::RsCodeGenerator::generate_rs_code
               at /Users/adetaylor/dev/autocxx/engine/src/conversion/codegen_rs/mod.rs:141:9
    12: autocxx_engine::conversion::BridgeConverter::convert
               at /Users/adetaylor/dev/autocxx/engine/src/conversion/mod.rs:150:26
    13: autocxx_engine::IncludeCppEngine::generate
               at /Users/adetaylor/dev/autocxx/engine/src/lib.rs:389:26
    14: autocxx_engine::parse_file::ParsedFile::resolve_all
               at /Users/adetaylor/dev/autocxx/engine/src/parse_file.rs:178:13
    15: autocxx_engine::builder::build_to_custom_directory
               at /Users/adetaylor/dev/autocxx/engine/src/builder.rs:141:5
    16: autocxx_engine::builder::build
               at /Users/adetaylor/dev/autocxx/engine/src/builder.rs:76:5
    17: autocxx_build::build
               at /Users/adetaylor/dev/autocxx/gen/build/src/lib.rs:37:5
    18: build_script_build::main
               at ./build.rs:21:17
    19: core::ops::function::FnOnce::call_once
               at /Users/adetaylor/.rustup/toolchains/nightly-x86_64-apple-darwin/lib/rustlib/src/rust/library/core/src/ops/function.rs:227:5
  note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.

[adetaylor]

This seems to be because there's a function called move

[adetaylor]

Reducing like this, for my records.

nohup ~/autocxx/target/release/autocxx-reduce -h sample.hpp -I . -p 'expected identifier' -d 'generate_all!()' -c-fblocks -k -- --n 64 

[adetaylor]

Minimized test case...

enum

Sigh.

[adetaylor]

It transpires that bindgen already tries hard to avoid this problem, but then we reintroduce it during subsequent autocxx processing.

For functions, we do the right thing. Bindgen generates a function named (for instance) move_ with an original_name attribute of move, which is perfect.

For methods, bindgen generates A_move, and then an impl called move_ which calls A_move. That's where things go wrong. So the fix in #498 should probably be discarded and replaced with one that is simply cleverer when dealing with method names, specifically.