-
Couldn't load subscription status.
- Fork 24.9k
Define TracingTimeStamp alias and refactor #51454
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
|
This pull request was exported from Phabricator. Differential Revision: D74892330 |
Summary: # Changelog: [Internal] There are multiple changes: 1. `PerformanceTracer` class, `TraceEvent` struct are moved to `tracing` namespace. These are parts of the Tracing subsystems of the jsinspector, this should bring more clarity and make things more explicit. 2. Added `Timing.h` class which defines conversion logic from `HighResTimeStamp` to absolute units that are expected by CDP. 3. `PerformanceTracer` will receive timestamps for Performance Web API entries in `HighResTimeStamp`. Also, we will explicilty define a Tracking Clock time origin that will be epoch of the `steady_clock`. This aligns with the approach in Chromium and saves us from aligning custom DOMHighResTimeStamps that can be specified in performance.mark / performance.measure calls: these should not extend the timeline window. I've confirmed that this is the current behavior in Chromium. Differential Revision: D74892330
|
This pull request was exported from Phabricator. Differential Revision: D74892330 |
Summary: Pull Request resolved: facebook#51454 # Changelog: [Internal] There are multiple changes: 1. `PerformanceTracer` class, `TraceEvent` struct are moved to `tracing` namespace. These are parts of the Tracing subsystems of the jsinspector, this should bring more clarity and make things more explicit. 2. Added `Timing.h` class which defines conversion logic from `HighResTimeStamp` to absolute units that are expected by CDP. 3. `PerformanceTracer` will receive timestamps for Performance Web API entries in `HighResTimeStamp`. Also, we will explicilty define a Tracking Clock time origin that will be epoch of the `steady_clock`. This aligns with the approach in Chromium and saves us from aligning custom DOMHighResTimeStamps that can be specified in performance.mark / performance.measure calls: these should not extend the timeline window. I've confirmed that this is the current behavior in Chromium. Differential Revision: D74892330
Summary: Pull Request resolved: facebook#51454 # Changelog: [Internal] There are multiple changes: 1. `PerformanceTracer` class, `TraceEvent` struct are moved to `tracing` namespace. These are parts of the Tracing subsystems of the jsinspector, this should bring more clarity and make things more explicit. 2. Added `Timing.h` class which defines conversion logic from `HighResTimeStamp` to absolute units that are expected by CDP. 3. `PerformanceTracer` will receive timestamps for Performance Web API entries in `HighResTimeStamp`. Also, we will explicilty define a Tracking Clock time origin that will be epoch of the `steady_clock`. This aligns with the approach in Chromium and saves us from aligning custom DOMHighResTimeStamps that can be specified in performance.mark / performance.measure calls: these should not extend the timeline window. I've confirmed that this is the current behavior in Chromium. Differential Revision: D74892330
Summary: # Changelog: [Internal] There are multiple changes: 1. `PerformanceTracer` class, `TraceEvent` struct are moved to `tracing` namespace. These are parts of the Tracing subsystems of the jsinspector, this should bring more clarity and make things more explicit. 2. Added `Timing.h` class which defines conversion logic from `HighResTimeStamp` to absolute units that are expected by CDP. 3. `PerformanceTracer` will receive timestamps for Performance Web API entries in `HighResTimeStamp`. Also, we will explicilty define a Tracking Clock time origin that will be epoch of the `steady_clock`. This aligns with the approach in Chromium and saves us from aligning custom DOMHighResTimeStamps that can be specified in performance.mark / performance.measure calls: these should not extend the timeline window. I've confirmed that this is the current behavior in Chromium. Differential Revision: D74892330
|
This pull request was exported from Phabricator. Differential Revision: D74892330 |
Summary: Pull Request resolved: facebook#51454 # Changelog: [Internal] There are multiple changes: 1. `PerformanceTracer` class, `TraceEvent` struct are moved to `tracing` namespace. These are parts of the Tracing subsystems of the jsinspector, this should bring more clarity and make things more explicit. 2. Added `Timing.h` class which defines conversion logic from `HighResTimeStamp` to absolute units that are expected by CDP. 3. `PerformanceTracer` will receive timestamps for Performance Web API entries in `HighResTimeStamp`. Also, we will explicilty define a Tracking Clock time origin that will be epoch of the `steady_clock`. This aligns with the approach in Chromium and saves us from aligning custom DOMHighResTimeStamps that can be specified in performance.mark / performance.measure calls: these should not extend the timeline window. I've confirmed that this is the current behavior in Chromium. Differential Revision: D74892330
Summary: Pull Request resolved: facebook#51454 # Changelog: [Internal] There are multiple changes: 1. `PerformanceTracer` class, `TraceEvent` struct are moved to `tracing` namespace. These are parts of the Tracing subsystems of the jsinspector, this should bring more clarity and make things more explicit. 2. Added `Timing.h` class which defines conversion logic from `HighResTimeStamp` to absolute units that are expected by CDP. 3. `PerformanceTracer` will receive timestamps for Performance Web API entries in `HighResTimeStamp`. Also, we will explicilty define a Tracking Clock time origin that will be epoch of the `steady_clock`. This aligns with the approach in Chromium and saves us from aligning custom DOMHighResTimeStamps that can be specified in performance.mark / performance.measure calls: these should not extend the timeline window. I've confirmed that this is the current behavior in Chromium. Differential Revision: D74892330
…onverters (facebook#51455) Summary: Pull Request resolved: facebook#51455 # Changelog: [Internal] The main idea is that subsystems who might use a different time origin (the starting point of the whole timeline of events), can use `toChronoSteadyClockTimePoint` method to get raw `std::chrono::steady_clock::time_point` and then offset it by some arbitrary epoch: be it unix time origin or `std::chrono::steady_clock::epoch`. `fromChronoSteadyClockTimePoint` can be used to convert time stamps from external systems, like Hermes. Differential Revision: D74892329
Summary: # Changelog: [Internal] There are multiple changes: 1. `PerformanceTracer` class, `TraceEvent` struct are moved to `tracing` namespace. These are parts of the Tracing subsystems of the jsinspector, this should bring more clarity and make things more explicit. 2. Added `Timing.h` class which defines conversion logic from `HighResTimeStamp` to absolute units that are expected by CDP. 3. `PerformanceTracer` will receive timestamps for Performance Web API entries in `HighResTimeStamp`. Also, we will explicilty define a Tracking Clock time origin that will be epoch of the `steady_clock`. This aligns with the approach in Chromium and saves us from aligning custom DOMHighResTimeStamps that can be specified in performance.mark / performance.measure calls: these should not extend the timeline window. I've confirmed that this is the current behavior in Chromium. Reviewed By: rubennorte Differential Revision: D74892330
Summary: Pull Request resolved: facebook#51454 # Changelog: [Internal] There are multiple changes: 1. `PerformanceTracer` class, `TraceEvent` struct are moved to `tracing` namespace. These are parts of the Tracing subsystems of the jsinspector, this should bring more clarity and make things more explicit. 2. Added `Timing.h` class which defines conversion logic from `HighResTimeStamp` to absolute units that are expected by CDP. 3. `PerformanceTracer` will receive timestamps for Performance Web API entries in `HighResTimeStamp`. Also, we will explicilty define a Tracking Clock time origin that will be epoch of the `steady_clock`. This aligns with the approach in Chromium and saves us from aligning custom DOMHighResTimeStamps that can be specified in performance.mark / performance.measure calls: these should not extend the timeline window. I've confirmed that this is the current behavior in Chromium. Reviewed By: rubennorte Differential Revision: D74892330
|
This pull request was exported from Phabricator. Differential Revision: D74892330 |
Summary: Pull Request resolved: facebook#51454 # Changelog: [Internal] There are multiple changes: 1. `PerformanceTracer` class, `TraceEvent` struct are moved to `tracing` namespace. These are parts of the Tracing subsystems of the jsinspector, this should bring more clarity and make things more explicit. 2. Added `Timing.h` class which defines conversion logic from `HighResTimeStamp` to absolute units that are expected by CDP. 3. `PerformanceTracer` will receive timestamps for Performance Web API entries in `HighResTimeStamp`. Also, we will explicilty define a Tracking Clock time origin that will be epoch of the `steady_clock`. This aligns with the approach in Chromium and saves us from aligning custom DOMHighResTimeStamps that can be specified in performance.mark / performance.measure calls: these should not extend the timeline window. I've confirmed that this is the current behavior in Chromium. Differential Revision: D74892330
|
This pull request has been merged in 0ec36c4. |
|
This pull request has been reverted by 5489e18. |
Summary:
Changelog: [Internal]
There are multiple changes:
PerformanceTracerclass,TraceEventstruct are moved totracingnamespace. These are parts of the Tracing subsystems of the jsinspector, this should bring more clarity and make things more explicit.Timing.hclass which defines timing primitives for the Tracing domain. These are compliant with CDP requirements: microseconds granularity and all events inside one profile has to have a common time origin. In order to guarantee this, we are introducing these primitives,std::chrono::steady_clock::epochwas selected as the time origin. Tracing as a subsystem is responsible for receiving timestamps from multiple sources and then converting them to this common abstraction.PerformanceTracerwill receive timestamps for Performance Web API entries in domain-specific abstractions. Right now this isDOMHighResTimeStamp, which is an alias fordouble, but which will be replaced byHighResTimeStampin one of the diffs at the top of the stack.Differential Revision: D74892330