Navigara

github.com-facebook-hermes

all · 31 devs · built 2026-06-08

Repository snapshot

Performance · 90d

1.4ETV

−32.0% since Q2 2025

ETV / dev / mo

0.1ETV

2 devs · 30-day window

Last 7 days · per dev

0.1ETV

per developer per week

Work mix

Growth29%· 5.5Maintenance54%· 10.2Fixes17%· 3.1
18.8 ETV all-time210 commits31 all-time devsJan 2025 – Jun 2026

Monthly reports

  • Highlights

    • Enhanced *Hermes Promise polyfill* rejection tracking to be spec-compliant by passing the rejected Promise instance as a third argument to `onUnhandled` and `onHandled` callbacks, enabling correct `PromiseRejectionEvent` population in downstream consumers ([commit/3d71c698]).
    • Introduced new *JSI APIs* (`IEventLoopControl` and `ISetEventLoopControl`) to `hermes-interfaces.h`, providing the *Hermes VM* with the capability to manage its event loop and delegate asynchronous tasks, particularly crucial for *Hermes Workers* ([ae855414 · Chi Tsai]).
    • Improved *CPU sampling timestamp accuracy* within the *Hermes VM's profiling subsystem* on *Apple platforms* by utilizing `clock_gettime_nsec_np(CLOCK_UPTIME_RAW)`, ensuring continuous, non-sleeping timelines for more reliable performance analysis ([f4ac05d3 · Ruslan Lesiutin]).
    • Optimized *JSI layer performance* by refactoring the `dynamicFromValue` function to pre-allocate memory for `folly::dynamic` arrays when converting JSI Arrays, significantly reducing reallocations for larger data structures ([829dcd74 · generatedunixname1389709262956747]).
    • Addressed a `-Wunreachable-code-return` build error in the vendored *LLVM fork* within `Host.cpp`, providing a critical fix to unblock *Origami Studio* alpha and beta builds on CI ([7f186d96 · Rob Wood]).

    Observations

    • All measured scores for May 2026, including total output, grow, maintenance, and waste, were 0. This represents a significant decrease across the board compared to the 2-month average.
    • Total output for May 2026 was 0, a 67% decrease compared to the 2-month average of 1.
    • Grow score was 0, a 63% decrease compared to the 2-month average of 0.
    • Maintenance score was 0, a 77% decrease compared to the 2-month average of 0.
    • Waste score was 0, a 96% decrease compared to the 2-month average of 0.
    • Commit volume for May 2026 was 6, a 37% decrease compared to the 2-month average of 10 commits, consistent with the overall reduction in activity.
    • The commit [7f186d96 · Rob Wood] addressed a build fix for a `-Wunreachable-code-return` error in the vendored *LLVM fork*, highlighting the ongoing need to manage compiler warnings and external dependency compatibility.
    • The enhancement to *Promise polyfill* rejection tracking in [commit/3d71c698] included explicit updates to `test/hermes/promise.js`, indicating good practice for ensuring new features are adequately tested.

Performance over time

ETV stacked by Growth, Maintenance and Fixes — 90-day moving average, normalized to ETV / month.

Average performance per developer

ETV per active developer per month — 30-day moving average.

Active developers over time

Unique developers committing each day — 90-day moving average.

Knowledge concentration

How dependent is this repo on a small number of contributors? Higher top-1 share = higher key-person risk.

Top 1
20.0 %
Top 3
43.3 %
Top 5
58.1 %

Chi Tsai owns 20.0 % of commits.

Top contributors

Most impactful commits

Top 20 by ETV in the all-time window.