Faster bootstrap runs without changing what a profile installs
Nobody has claimed this yet.
Assessment
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Newbie friendliness
- 42/100
- Issue type
- Feature
- Clarity
- Clearly specified
- Activity status
- Active
- Tech stack
- docker, shell
- Domain
- build-system
Research direction
Start by reading build/tart/vm run and install/brew, which the issue names as the VM launch and brew installation entry points. Check how profile sets are selected and how the checkout is mounted, then trace the bootstrap loop for mise, runtimes, and nix. Done means the allowed brew sets install in one bundle call, the host cache is reused, and a fresh run still proves the complete profile works.
Written by the indexing model from the issue text.
Description
A full headless run spends most of its time in the brew step. Almost all of that time goes to downloading bottles. The previous run already downloaded the same bottles into a VM that was then thrown away.
The profile cut is right and should not be trimmed further. The laptop needs gcc, llvm, pandoc, and tectonic, so the VM must install them too. The time can be recovered elsewhere.
-
Share Homebrew's download cache from the host.
build/tart/vm runalready mounts the checkout. A second mount, read-write, can hold the cache.HOMEBREW_CACHEinside the guest points at it. Bottles then download once per host instead of once per VM, and a reset VM pours from cache. This is the biggest win, and it changes nothing about the result. -
One bundle call per run.
install/brewbundles set by set so that a profile can omit some sets. Bundle fetches every download for one call in parallel. Twelve calls therefore mean twelve serial fetch phases. The installer can generate a temporary Brewfile that includes only the allowed sets and run onebrew bundleagainst it. The per-set omission survives, and the run gets one fetch phase. -
Homebrew environment for the run. The bootstrap can set
HOMEBREW_NO_AUTO_UPDATE,HOMEBREW_NO_INSTALL_CLEANUP, andHOMEBREW_NO_ENV_HINTSfor its own process. Those three stop the repeated update checks, the cleanup passes between calls, and the hint noise in the log. The installer still runsbrew updateonce, explicitly. -
A checkpoint VM for the loop, never for the proof.
tart cloneof a VM stopped right after the brew step gives a starting point for iterating onmise,runtimes, andnixin minutes. The fresh run stays the only evidence that the whole thing works. This checkpoint would have shortened the last four iterations considerably.
- Dominant language
- Shell
- Stars
- 1
- Forks
- 0
- Avg merge
- 14h 36m
- Merged PRs (30d)
- 3
Getting set up
- No Dockerfile or Docker Compose file
- Has a pull request template
- No contributing guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from indexzero/dotfiles
-
Difficulty 2/5 1-3 hours Newbie friendliness 74/100
-
Difficulty 4/5 3-5 days Newbie friendliness 56/100
-
Difficulty 4/5 3-5 days Newbie friendliness 39/100
-
Difficulty 1/5 Under an hour Newbie friendliness 48/100
-
Difficulty 5/5 Over a week Newbie friendliness 8/100
All issues in indexzero/dotfiles
Similar issues
-
数据界面的横向滚动条缺失Open
Difficulty 2/5 1-3 hours Newbie friendliness 62/100
zilliztech/attu#1037 ·
-
bug
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
Maintainers usually reply within 1 day
-
good first issue tests
Difficulty 2/5 1-3 hours Newbie friendliness 84/100
Maintainers usually reply within 1 day
-
documentation
Difficulty 2/5 1-3 hours Newbie friendliness 74/100
alunduil/alunduil-chezmoi#964 ·
Maintainers usually reply within 1 day
-
bug-unconfirmed
Difficulty 2/5 1-3 hours Newbie friendliness 64/100
Maintainers usually reply within 1 day