r/f? @mbrubeck
No need to r+ urgently; I want to do a little bit more testing of the release build, but I'm hoping to land this bit (moving to a more sane build process) next week.
The new version of building an APK:
1) Removes the glutin-based APK builder from the link step
2) Adds a build.rs step to the build of the final Servo library that adds the native code required by glutin's android_rs_glue (e.g., `ANativeActivity_onCreate` definition)
3) Replaces the link step with a `fake-ld.sh` script that instead creates a libservo.so
4) Adds a new mach `package` step to build the APK that has some Rust code that builds the library from a set of in-tree build files
This setup fixes a number of problems:
1) We can use gdb, because we use `ndk-build`, which adds the .gdbserver info, plus we keep around all of the build files (also required by the ndk gdb)
2) We can add more Java code & hooks to handle Android intents
3) We no longer have any git submodules or the awkward two-step build with android-rs-glue
Many other setups were tried (and failed). The most obvious ones is building a libservo.so from a `dylib` target from the servo build on Android. This doesn't work because you can't have a different default lib target on one platform than others in Cargo, and you also can't pass it in from the commandline (e.g., --lib does not have a dylib arg). Additionally, if you don't go through the intermediate libservo.rlib step (which removes unused symbols), then you end up with a TON of missing symbols because our -sys crates are super sloppy about that. I spent a few weeks beginning to clean them up, but since it's something we can't easily enforce (and new -sys packages will have this problem, too, since it's only an issue with the Android loader), it made more sense to me to just have the build set up to discard those unused bits of code before they ever get to the linker, much less the loader.
Source-Repo: https://github.com/servo/servo
Source-Revision: 0699d38e80c029a384354da96596421f3a97ceef
This is a direct extract from my abandoned PR for a lint (#7546), along with some rather clumsy modifications (only on `components/script/dom/mod.rs` and `components/style/lib.rs`), because I had to sort some of the files again to make peace with tidy, which hasn't been educated about sorting yet!
Source-Repo: https://github.com/servo/servo
Source-Revision: a7208869f2903e36f9b2f540b55b50283d7df466
This PR adds Heartbeats capability to servo. Heartbeats are used for detailed performance and power/energy profiling. We will add the power/energy readings in the future.
New dependencies are introduced which need in-depth reviews. I'm the only one who has had eyes on any of this, and I have limited resources for testing cross-platform compatibility.
* https://github.com/libheartbeats/heartbeats-simple - provides native C libraries from a shared code base:
* hbs[-static] - performance monitoring
* hbs-acc[-static] - performance with accuracy monitoring
* hbs-pow[-static] - performance with power/energy monitoring (the one we're using)
* hbs-acc-pow[-static] - performance with accuracy and power/energy monitoring
* https://github.com/connorimes/heartbeats-simple-sys provides rust wrappers for the native C libraries above - one crate for each + a common crate. These link with the *-static versions of the heartbeats libraries.
* https://github.com/connorimes/heartbeats-simple-rust provides rust abstractions over the -sys crates above - one crate for each.
The new `heartbeats` module in the `profile` crate looks for environment variables telling it to use heartbeats for each ProfilerCategory and where to put log files. (Of course, if somebody knows how to iterate over the enum instead of hardcoding each one, that would be fantastic.) If the environment variables aren't set for particular categories, heartbeats aren't created or used.
An interface change is made in the `profile_traits` crate to pass both the start and end time in a `ProfilerMsg` instead of just the elapsed time. Later we will add energy readings as well.
Source-Repo: https://github.com/servo/servo
Source-Revision: d89e4f7991a4e43f16ea57587004e3616addcc09
Uses a couple of extra threads to work around the lack of cross-process
boxed trait objects.
r? @nnethercote
Source-Repo: https://github.com/servo/servo
Source-Revision: f778e0eecf7cd8a2b870d18c3c305ff10d6b1894
A rebuild after touching components/profile/mem.rs now takes 48 seconds (and
only rebuilds `profile` and `servo`) which is much lower than it used to be.
In comparison, a rebuild after touching components/profile_traits/mem.rs takes
294 seconds and rebuilds many more crates.
This change also removes some unnecessary crate dependencies in `net` and
`net_traits`.
Source-Repo: https://github.com/servo/servo
Source-Revision: 77f653da2c4120ea7ac1a946d97fc70059d513d4
- Most of util::memory has been moved into profile::mem, though the
`SizeOf` trait and related things remain in util::memory. The
`SystemMemoryReporter` code is now in a submodule
profile::mem::system_reporter.
- util::time has been moved entirely into profile::time.
Source-Repo: https://github.com/servo/servo
Source-Revision: d1268ec9c6633684270015e7b2619181aeb47b8b
