bernardo/Nomarchy
1
1
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

removing the ISO eval from the gitea workflow
All checks were successful
Check / eval (push) Successful in 3m9s
Details

Browse Source
This commit is contained in:
Bernardo Magri
2026-07-12 20:07:13 +01:00
parent 67400b07dc
commit 53e3c32852
2 changed files with 18 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
docs/ROADMAP.md
View File

@@ -1007,8 +1007,11 @@ Design/decision records and a running log of shipped work (items marked
2026-07-12: both gates moved from one big `nix flake check --no-build` 2026-07-12: both gates moved from one big `nix flake check --no-build`
(~6.0 GB peak RSS — OOM'd the 4 GB VPS, then could never fit the (~6.0 GB peak RSS — OOM'd the 4 GB VPS, then could never fit the
runner's 2 GB container cap) to `tools/ci-eval.sh`, one output per nix runner's 2 GB container cap) to `tools/ci-eval.sh`, one output per nix
process (worst single output: nomarchy-live at 2.69 GB — needs the process. **nomarchy-live is excluded by decision** (Bernardo, same
container's default swap allowance; everything else ≤ 1 GB). Explicitly day): the ISO is built/distributed manually, so its guard rides that
manual flow — and its eval alone is 2.69 GB, the one output that can't
fit the cap (Actions run 315: everything else green, it OOM-killed).
With it out, the worst CI output is 0.99 GB — no server change needed. Explicitly
*not* a binary cache: compile-from-source is a deliberate values call *not* a binary cache: compile-from-source is a deliberate values call
(Gentoo-style), so this automates the *config/lock* channel, not (Gentoo-style), so this automates the *config/lock* channel, not
artifact distribution. artifact distribution.

14
tools/ci-eval.sh
View File

@@ -7,7 +7,9 @@
# process caps the peak at the largest single output (~0.93 GB, # process caps the peak at the largest single output (~0.93 GB,
# downstream-template-system): the heap is freed when each process # downstream-template-system): the heap is freed when each process
# exits. Wall time is worse (shared nixpkgs re-eval per process); # exits. Wall time is worse (shared nixpkgs re-eval per process);
# irrelevant next to the memory cap. Used by .gitea/workflows/check.yml # irrelevant next to the memory cap. nomarchy-live is skipped by
# decision (see the loop below); with it out, the worst output is
# hardware-toggles at 0.99 GB. Used by .gitea/workflows/check.yml
# and bump.yml; runs anywhere (`bash tools/ci-eval.sh`). # and bump.yml; runs anywhere (`bash tools/ci-eval.sh`).
set -euo pipefail set -euo pipefail
@@ -27,6 +29,16 @@ for a in $(list "checks.$sys"); do
evaluate "checks.$sys.$a.drvPath" evaluate "checks.$sys.$a.drvPath"
done done
for c in $(list nixosConfigurations); do for c in $(list nixosConfigurations); do
# nomarchy-live (the installer ISO) is deliberately NOT CI-gated
# (Bernardo, 2026-07-12): the ISO is built and distributed manually,
# so its guard lives with that manual flow (local `nix flake check` /
# the ISO build itself) — and it is the one output whose EVAL alone
# (~2.7 GB) cannot fit the runner's 2 GB cap (Actions run 315: every
# other output green, this one OOM-killed).
if [ "$c" = "nomarchy-live" ]; then
echo "skip: nixosConfigurations.$c (manual ISO flow — see comment)"
continue
fi
evaluate "nixosConfigurations.$c.config.system.build.toplevel.drvPath" evaluate "nixosConfigurations.$c.config.system.build.toplevel.drvPath"
done done
for h in $(list homeConfigurations); do for h in $(list homeConfigurations); do