Overview - Nx

Nx

7 days30 days

Latest CI Pipeline Executions

Fuzzy

Succeeded
master
af07e75d feat(core): use jemalloc with tuned decay timers for native module (#34444) ## Current Behavior The Nx native module (Rust cdylib loaded by Node.js) uses the system allocator. The daemon process retains a large RSS footprint after the initial project graph build, even though most of that memory is no longer in use. On macOS and Linux, the system allocator doesn't aggressively return freed pages to the OS. ## Expected Behavior The daemon's steady-state RSS drops significantly after graph build by using jemalloc with tuned page purge timers. Peak RSS and wall time are unaffected. ## Changes Adds [tikv-jemallocator](https://github.com/tikv/jemallocator) as the global allocator on Linux and macOS, with two compile-time settings: - **`dirty_decay_ms:1000`** — returns freed pages to the OS after 1s instead of the default 10s. Tuned to Nx's phase-separated workload (graph build → idle → task execution), where transitions happen every ~30-60s. Benchmarked against 5s and 10s — both too slow to purge between phases. - **`muzzy_decay_ms:0`** — skips the lazy purge phase (`MADV_FREE`) and goes straight to `MADV_DONTNEED`. Required on macOS and Linux ≥ 4.5 where `MADV_FREE` doesn't actually reduce RSS. Windows and WASI continue using the system allocator. Windows is excluded because `tikv-jemalloc-sys` fails to build with MSVC (spaces in `cl.exe` path break the autoconf configure script). Tracked upstream in [tikv/jemallocator#99](https://github.com/tikv/jemallocator/pull/99). ### Other Settings Considered Tested narenas reduction, tcache_max, extent fit tuning, background threads, and decay timer values. Only the decay timer configuration improved steady-state RSS without wall time regression.
1 month ago by leosvelperez
Succeeded
master
af07e75d feat(core): use jemalloc with tuned decay timers for native module (#34444) ## Current Behavior The Nx native module (Rust cdylib loaded by Node.js) uses the system allocator. The daemon process retains a large RSS footprint after the initial project graph build, even though most of that memory is no longer in use. On macOS and Linux, the system allocator doesn't aggressively return freed pages to the OS. ## Expected Behavior The daemon's steady-state RSS drops significantly after graph build by using jemalloc with tuned page purge timers. Peak RSS and wall time are unaffected. ## Changes Adds [tikv-jemallocator](https://github.com/tikv/jemallocator) as the global allocator on Linux and macOS, with two compile-time settings: - **`dirty_decay_ms:1000`** — returns freed pages to the OS after 1s instead of the default 10s. Tuned to Nx's phase-separated workload (graph build → idle → task execution), where transitions happen every ~30-60s. Benchmarked against 5s and 10s — both too slow to purge between phases. - **`muzzy_decay_ms:0`** — skips the lazy purge phase (`MADV_FREE`) and goes straight to `MADV_DONTNEED`. Required on macOS and Linux ≥ 4.5 where `MADV_FREE` doesn't actually reduce RSS. Windows and WASI continue using the system allocator. Windows is excluded because `tikv-jemalloc-sys` fails to build with MSVC (spaces in `cl.exe` path break the autoconf configure script). Tracked upstream in [tikv/jemallocator#99](https://github.com/tikv/jemallocator/pull/99). ### Other Settings Considered Tested narenas reduction, tcache_max, extent fit tuning, background threads, and decay timer values. Only the decay timer configuration improved steady-state RSS without wall time regression.
1 month ago by leosvelperez
Succeeded
master
af07e75d feat(core): use jemalloc with tuned decay timers for native module (#34444) ## Current Behavior The Nx native module (Rust cdylib loaded by Node.js) uses the system allocator. The daemon process retains a large RSS footprint after the initial project graph build, even though most of that memory is no longer in use. On macOS and Linux, the system allocator doesn't aggressively return freed pages to the OS. ## Expected Behavior The daemon's steady-state RSS drops significantly after graph build by using jemalloc with tuned page purge timers. Peak RSS and wall time are unaffected. ## Changes Adds [tikv-jemallocator](https://github.com/tikv/jemallocator) as the global allocator on Linux and macOS, with two compile-time settings: - **`dirty_decay_ms:1000`** — returns freed pages to the OS after 1s instead of the default 10s. Tuned to Nx's phase-separated workload (graph build → idle → task execution), where transitions happen every ~30-60s. Benchmarked against 5s and 10s — both too slow to purge between phases. - **`muzzy_decay_ms:0`** — skips the lazy purge phase (`MADV_FREE`) and goes straight to `MADV_DONTNEED`. Required on macOS and Linux ≥ 4.5 where `MADV_FREE` doesn't actually reduce RSS. Windows and WASI continue using the system allocator. Windows is excluded because `tikv-jemalloc-sys` fails to build with MSVC (spaces in `cl.exe` path break the autoconf configure script). Tracked upstream in [tikv/jemallocator#99](https://github.com/tikv/jemallocator/pull/99). ### Other Settings Considered Tested narenas reduction, tcache_max, extent fit tuning, background threads, and decay timer values. Only the decay timer configuration improved steady-state RSS without wall time regression.
1 month ago by leosvelperez
Canceled
34647
Flaky task→
96361bfd cleanup(core): cache compiled regex in validate_outputs is_missing_prefix() compiled Regex::new() on every call (~1-5µs each). With hundreds of outputs per validation, this adds up to ~1-2ms of wasted work. Use LazyLock<Regex> to compile once and reuse.
1 month ago by leosvelperez
Succeeded
34491
2991056a fix(core): handle already-quoted values and word-split fragments in shell quoting Add isAlreadyQuoted guard to serializeOverridesIntoCommandLine and wrapArgIntoQuotesIfNeeded to skip re-quoting values already wrapped in matching quotes. Add rejoinQuotedFragments to reassemble shell word-split fragments of quoted values in __unparsed__ before applying quoting logic. Without this, fragments like args"}}' get incorrectly re-quoted, breaking commands that pass JSON config via escaped single quotes (e.g., cypress e2e --config).
1 month ago by leosvelperez
Succeeded
master
65d0f554 chore(repo): preserve x-prompt and requires in angular migrations upgrade script (#34641) When latest >= next, the build-migrations script now preserves the existing x-prompt and requires fields from the pre-release entry, rather than skipping them, and adjusts the requires upper bound to the stable version.
1 month ago by leosvelperez
Failed
34647
Fix rejected→
96a7a878 cleanup(core): cache compiled regex in validate_outputs is_missing_prefix() compiled Regex::new() on every call (~1-5µs each). With hundreds of outputs per validation, this adds up to ~1-2ms of wasted work. Use LazyLock<Regex> to compile once and reuse.
1 month ago by leosvelperez
Failed
34491
2ed5bb2f fix(core): skip shell quoting for already-quoted values in serializeOption
1 month ago by leosvelperez
Succeeded
master
af07e75d feat(core): use jemalloc with tuned decay timers for native module (#34444) ## Current Behavior The Nx native module (Rust cdylib loaded by Node.js) uses the system allocator. The daemon process retains a large RSS footprint after the initial project graph build, even though most of that memory is no longer in use. On macOS and Linux, the system allocator doesn't aggressively return freed pages to the OS. ## Expected Behavior The daemon's steady-state RSS drops significantly after graph build by using jemalloc with tuned page purge timers. Peak RSS and wall time are unaffected. ## Changes Adds [tikv-jemallocator](https://github.com/tikv/jemallocator) as the global allocator on Linux and macOS, with two compile-time settings: - **`dirty_decay_ms:1000`** — returns freed pages to the OS after 1s instead of the default 10s. Tuned to Nx's phase-separated workload (graph build → idle → task execution), where transitions happen every ~30-60s. Benchmarked against 5s and 10s — both too slow to purge between phases. - **`muzzy_decay_ms:0`** — skips the lazy purge phase (`MADV_FREE`) and goes straight to `MADV_DONTNEED`. Required on macOS and Linux ≥ 4.5 where `MADV_FREE` doesn't actually reduce RSS. Windows and WASI continue using the system allocator. Windows is excluded because `tikv-jemalloc-sys` fails to build with MSVC (spaces in `cl.exe` path break the autoconf configure script). Tracked upstream in [tikv/jemallocator#99](https://github.com/tikv/jemallocator/pull/99). ### Other Settings Considered Tested narenas reduction, tcache_max, extent fit tuning, background threads, and decay timer values. Only the decay timer configuration improved steady-state RSS without wall time regression.
1 month ago by leosvelperez
Succeeded
master
af07e75d feat(core): use jemalloc with tuned decay timers for native module (#34444) ## Current Behavior The Nx native module (Rust cdylib loaded by Node.js) uses the system allocator. The daemon process retains a large RSS footprint after the initial project graph build, even though most of that memory is no longer in use. On macOS and Linux, the system allocator doesn't aggressively return freed pages to the OS. ## Expected Behavior The daemon's steady-state RSS drops significantly after graph build by using jemalloc with tuned page purge timers. Peak RSS and wall time are unaffected. ## Changes Adds [tikv-jemallocator](https://github.com/tikv/jemallocator) as the global allocator on Linux and macOS, with two compile-time settings: - **`dirty_decay_ms:1000`** — returns freed pages to the OS after 1s instead of the default 10s. Tuned to Nx's phase-separated workload (graph build → idle → task execution), where transitions happen every ~30-60s. Benchmarked against 5s and 10s — both too slow to purge between phases. - **`muzzy_decay_ms:0`** — skips the lazy purge phase (`MADV_FREE`) and goes straight to `MADV_DONTNEED`. Required on macOS and Linux ≥ 4.5 where `MADV_FREE` doesn't actually reduce RSS. Windows and WASI continue using the system allocator. Windows is excluded because `tikv-jemalloc-sys` fails to build with MSVC (spaces in `cl.exe` path break the autoconf configure script). Tracked upstream in [tikv/jemallocator#99](https://github.com/tikv/jemallocator/pull/99). ### Other Settings Considered Tested narenas reduction, tcache_max, extent fit tuning, background threads, and decay timer values. Only the decay timer configuration improved steady-state RSS without wall time regression.
1 month ago by leosvelperez
Succeeded
34641
11b911af chore(repo): preserve x-prompt and requires in angular migrations upgrade script
1 month ago by leosvelperez
Failed
master
af07e75d feat(core): use jemalloc with tuned decay timers for native module (#34444) ## Current Behavior The Nx native module (Rust cdylib loaded by Node.js) uses the system allocator. The daemon process retains a large RSS footprint after the initial project graph build, even though most of that memory is no longer in use. On macOS and Linux, the system allocator doesn't aggressively return freed pages to the OS. ## Expected Behavior The daemon's steady-state RSS drops significantly after graph build by using jemalloc with tuned page purge timers. Peak RSS and wall time are unaffected. ## Changes Adds [tikv-jemallocator](https://github.com/tikv/jemallocator) as the global allocator on Linux and macOS, with two compile-time settings: - **`dirty_decay_ms:1000`** — returns freed pages to the OS after 1s instead of the default 10s. Tuned to Nx's phase-separated workload (graph build → idle → task execution), where transitions happen every ~30-60s. Benchmarked against 5s and 10s — both too slow to purge between phases. - **`muzzy_decay_ms:0`** — skips the lazy purge phase (`MADV_FREE`) and goes straight to `MADV_DONTNEED`. Required on macOS and Linux ≥ 4.5 where `MADV_FREE` doesn't actually reduce RSS. Windows and WASI continue using the system allocator. Windows is excluded because `tikv-jemalloc-sys` fails to build with MSVC (spaces in `cl.exe` path break the autoconf configure script). Tracked upstream in [tikv/jemallocator#99](https://github.com/tikv/jemallocator/pull/99). ### Other Settings Considered Tested narenas reduction, tcache_max, extent fit tuning, background threads, and decay timer values. Only the decay timer configuration improved steady-state RSS without wall time regression.
1 month ago by leosvelperez
Succeeded
33655
ae264808 fix(core): support numeric signal
1 month ago by leosvelperez
Succeeded
master
af07e75d feat(core): use jemalloc with tuned decay timers for native module (#34444) ## Current Behavior The Nx native module (Rust cdylib loaded by Node.js) uses the system allocator. The daemon process retains a large RSS footprint after the initial project graph build, even though most of that memory is no longer in use. On macOS and Linux, the system allocator doesn't aggressively return freed pages to the OS. ## Expected Behavior The daemon's steady-state RSS drops significantly after graph build by using jemalloc with tuned page purge timers. Peak RSS and wall time are unaffected. ## Changes Adds [tikv-jemallocator](https://github.com/tikv/jemallocator) as the global allocator on Linux and macOS, with two compile-time settings: - **`dirty_decay_ms:1000`** — returns freed pages to the OS after 1s instead of the default 10s. Tuned to Nx's phase-separated workload (graph build → idle → task execution), where transitions happen every ~30-60s. Benchmarked against 5s and 10s — both too slow to purge between phases. - **`muzzy_decay_ms:0`** — skips the lazy purge phase (`MADV_FREE`) and goes straight to `MADV_DONTNEED`. Required on macOS and Linux ≥ 4.5 where `MADV_FREE` doesn't actually reduce RSS. Windows and WASI continue using the system allocator. Windows is excluded because `tikv-jemalloc-sys` fails to build with MSVC (spaces in `cl.exe` path break the autoconf configure script). Tracked upstream in [tikv/jemallocator#99](https://github.com/tikv/jemallocator/pull/99). ### Other Settings Considered Tested narenas reduction, tcache_max, extent fit tuning, background threads, and decay timer values. Only the decay timer configuration improved steady-state RSS without wall time regression.
1 month ago by leosvelperez
Succeeded
master
af07e75d feat(core): use jemalloc with tuned decay timers for native module (#34444) ## Current Behavior The Nx native module (Rust cdylib loaded by Node.js) uses the system allocator. The daemon process retains a large RSS footprint after the initial project graph build, even though most of that memory is no longer in use. On macOS and Linux, the system allocator doesn't aggressively return freed pages to the OS. ## Expected Behavior The daemon's steady-state RSS drops significantly after graph build by using jemalloc with tuned page purge timers. Peak RSS and wall time are unaffected. ## Changes Adds [tikv-jemallocator](https://github.com/tikv/jemallocator) as the global allocator on Linux and macOS, with two compile-time settings: - **`dirty_decay_ms:1000`** — returns freed pages to the OS after 1s instead of the default 10s. Tuned to Nx's phase-separated workload (graph build → idle → task execution), where transitions happen every ~30-60s. Benchmarked against 5s and 10s — both too slow to purge between phases. - **`muzzy_decay_ms:0`** — skips the lazy purge phase (`MADV_FREE`) and goes straight to `MADV_DONTNEED`. Required on macOS and Linux ≥ 4.5 where `MADV_FREE` doesn't actually reduce RSS. Windows and WASI continue using the system allocator. Windows is excluded because `tikv-jemalloc-sys` fails to build with MSVC (spaces in `cl.exe` path break the autoconf configure script). Tracked upstream in [tikv/jemallocator#99](https://github.com/tikv/jemallocator/pull/99). ### Other Settings Considered Tested narenas reduction, tcache_max, extent fit tuning, background threads, and decay timer values. Only the decay timer configuration improved steady-state RSS without wall time regression.
1 month ago by leosvelperez
Succeeded
master
af07e75d feat(core): use jemalloc with tuned decay timers for native module (#34444) ## Current Behavior The Nx native module (Rust cdylib loaded by Node.js) uses the system allocator. The daemon process retains a large RSS footprint after the initial project graph build, even though most of that memory is no longer in use. On macOS and Linux, the system allocator doesn't aggressively return freed pages to the OS. ## Expected Behavior The daemon's steady-state RSS drops significantly after graph build by using jemalloc with tuned page purge timers. Peak RSS and wall time are unaffected. ## Changes Adds [tikv-jemallocator](https://github.com/tikv/jemallocator) as the global allocator on Linux and macOS, with two compile-time settings: - **`dirty_decay_ms:1000`** — returns freed pages to the OS after 1s instead of the default 10s. Tuned to Nx's phase-separated workload (graph build → idle → task execution), where transitions happen every ~30-60s. Benchmarked against 5s and 10s — both too slow to purge between phases. - **`muzzy_decay_ms:0`** — skips the lazy purge phase (`MADV_FREE`) and goes straight to `MADV_DONTNEED`. Required on macOS and Linux ≥ 4.5 where `MADV_FREE` doesn't actually reduce RSS. Windows and WASI continue using the system allocator. Windows is excluded because `tikv-jemalloc-sys` fails to build with MSVC (spaces in `cl.exe` path break the autoconf configure script). Tracked upstream in [tikv/jemallocator#99](https://github.com/tikv/jemallocator/pull/99). ### Other Settings Considered Tested narenas reduction, tcache_max, extent fit tuning, background threads, and decay timer values. Only the decay timer configuration improved steady-state RSS without wall time regression.
1 month ago by leosvelperez
Succeeded
master
af07e75d feat(core): use jemalloc with tuned decay timers for native module (#34444) ## Current Behavior The Nx native module (Rust cdylib loaded by Node.js) uses the system allocator. The daemon process retains a large RSS footprint after the initial project graph build, even though most of that memory is no longer in use. On macOS and Linux, the system allocator doesn't aggressively return freed pages to the OS. ## Expected Behavior The daemon's steady-state RSS drops significantly after graph build by using jemalloc with tuned page purge timers. Peak RSS and wall time are unaffected. ## Changes Adds [tikv-jemallocator](https://github.com/tikv/jemallocator) as the global allocator on Linux and macOS, with two compile-time settings: - **`dirty_decay_ms:1000`** — returns freed pages to the OS after 1s instead of the default 10s. Tuned to Nx's phase-separated workload (graph build → idle → task execution), where transitions happen every ~30-60s. Benchmarked against 5s and 10s — both too slow to purge between phases. - **`muzzy_decay_ms:0`** — skips the lazy purge phase (`MADV_FREE`) and goes straight to `MADV_DONTNEED`. Required on macOS and Linux ≥ 4.5 where `MADV_FREE` doesn't actually reduce RSS. Windows and WASI continue using the system allocator. Windows is excluded because `tikv-jemalloc-sys` fails to build with MSVC (spaces in `cl.exe` path break the autoconf configure script). Tracked upstream in [tikv/jemallocator#99](https://github.com/tikv/jemallocator/pull/99). ### Other Settings Considered Tested narenas reduction, tcache_max, extent fit tuning, background threads, and decay timer values. Only the decay timer configuration improved steady-state RSS without wall time regression.
1 month ago by leosvelperez
Succeeded
33862
5a8af0b0 fix(core): invalidate daemon cache singleton on nx.json changes
1 month ago by leosvelperez
Succeeded
master
b1614d75 feat(angular): add support for Angular v21.2 (#34592) ## Current Behavior Nx doesn't support Angular v21.2. ## Expected Behavior Nx should support Angular v21.2.
1 month ago by leosvelperez
Succeeded
master
b1614d75 feat(angular): add support for Angular v21.2 (#34592) ## Current Behavior Nx doesn't support Angular v21.2. ## Expected Behavior Nx should support Angular v21.2.
1 month ago by leosvelperez