nx_fmt.nx

buildroot/runtime/nx_fmt.nx

1529 B34 linesdepth 2pulls 2 transitivereach 32 importersview sourcekind library

docs dependencies structs consts functions

about

nx_fmt.nx -- shared LEAK-FREE integer/string formatters. Import this instead of copying the per-call sys_mmap putn idiom (the propagating mmap leak). Uses a FRAME-LOCAL STACK ARRAY `var buf: [64]u8` (the [N]T feature, live 2026-07-06): ZERO heap, per-call (NO data race even under threading), freed on return. This is the clean successor to the lazy-static version -- composes the language's own stack-array primitive, no mmap at all. KAT: nx_fmt_test (0 / 42 / -12345 / 1000000 / i64-max).

dependencies 1 imports · 32 importers

diagram shows first 10 each side; +0 more imports, +22 more importers in the complete lists below.

imports: nx_syscalls.nx

imported by: nx_batchscale_kat.nxnx_conv2d_mt_gate.nxnx_conv_speedup.nxnx_exp_ab.nxnx_f32_ffn_path_gate.nxnx_f32x8_range_gate.nxnx_fmt_test.nxnx_gguf_typecensus.nxnx_i8dot_ab.nxnx_lang_sota_census.nxnx_lw_cache_gate.nxnx_lwq24_probe.nxnx_matmul_t_pool_gate.nxnx_matmul_tile_kat.nxnx_membw.nxnx_moe_gate.nxnx_paged_kv_gate.nxnx_pool_ring_gate.nxnx_pteam_gate.nxnx_q4k_fused_vs_x4_gate.nxnx_q4k_matmul_mt_gate.nxnx_q4k_matmul_x4_gate.nxnx_q5_0_fused_dot_gate.nxnx_q5_0_threaded_gate.nxnx_q5q8_ab.nxnx_q8_0_simd_gate.nxnx_q8_coldwarm.nxnx_q8_matmul_micro.nxnx_q8_mix_ab.nxnx_q8_scaling.nxnx_q8_st_bw.nxnx_softfloat_cost_kat.nx

structs

none

consts

none

functions

9	func fmt_putn_fd(fd: i64, v: i64) -> i64 called by 1: fmt_putn calls 1: sys_write
30	func fmt_putn(v: i64) -> i64 { return fmt_putn_fd(1, v) } called by 31: main main main main main main+25 calls 1: fmt_putn_fd
33	func fmt_puts_fd(fd: i64, s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } sys_write(fd, s, n); return 0 } called by 1: fmt_puts calls 1: sys_write
34	func fmt_puts(s: *u8) -> i64 { return fmt_puts_fd(1, s) } called by 62: bs_nl main g_nl main main eab_nl+56 calls 1: fmt_puts_fd