Repository navigation
Add Bitwarden's web vault as a real app; all results from one 4-core run - #1
Open
ChristopherPHolder wants to merge 3 commits into
Open
ChristopherPHolder wants to merge 3 commits into
ChristopherPHolder wants to merge 3 commits into
Conversation
Pins bitwarden/clients apps/web at 8139eb0 with a migration patch from Angular 21 / TypeScript 5.9 to Angular 22 / TypeScript 6, which the fetcher now applies. R skips apps not built with the Angular CLI. ngp's type-check config maps deep imports into the app's npm-workspace packages to ngp's output, so originals aren't loaded next to ngp's copies.
Every table now comes from CI run 39469262093 on GitHub's standard 4-core, 16 GB runners, replacing the 2-core runs 38888167669 and 38899025048 and the separate Bitwarden run 39463856365. Adds a section on what changed from 2 to 4 cores: the tsgo pipelines gained most (S4 at 5,000 components 0.75x -> 0.96x, S8 1.10x -> 1.50x), tsc 6 barely moved, and native didn't pull ahead of wasm. No pipeline was noisy (max MAD 2.7%), so no S3 rerun was needed.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds Bitwarden's web vault (
bitwarden/clientsapps/webat8139eb0: 598 components, 3,295 source files) as a third real app. The README's results now all come from one CI run on 4-core runners, run 39469262093.Bitwarden
corpus/real/bitwarden.jsonpins the app. Bitwarden is on Angular 21 and TypeScript 5.9, socorpus/real/bitwarden.patch(nine files) upgrades it to the Angular 22 and TypeScript 6 the toolchain is built from, and nothing more. It covers the CDK stepper and dialog API changes, the new NG1054 error ontoggledChange,baseUrl,esModuleInteropfor JSZip, and excluding Jest specs. After the patch, ngc's full type-check reports nothing, withstrictTemplateson.corpus/fetch-real.mjsapplies a pin'spatchand recordscliBuildin the manifest.harness/run.mjsskips theng buildreference (R) when an app isn't built with the Angular CLI. Bitwarden builds with webpack.@bitwarden/components/src/...) at ngp's output. Before, TypeScript resolved them through thenode_modulessymlink to the original sources, loaded every class twice, and reported a false error. Ghostfolio's and Super Productivity's diagnostics are unchanged.Results (run 39469262093, 4 cores): S1 2.46×, S2 2.21×, S3
tsc6 0.79×, S4tsgo7 1.64×, S8tsgo7 2.07×.New ngp finding: ngp's type-check blocks report 34 errors that ngc doesn't. Five come from a bug. When an input's
transformis declared in another file, ngp copies the transform's parameter type into the consumer's type-check block using offsets from the declaring file but reads the text from the consumer's own source. The blocks gettepper,asswor,nsealeands.chanwherestringbelongs. The other 29 are places where ngp's blocks check more strictly than ngc's. Details are in the README's findings.4-core rerun
The repo went public, so GitHub's standard runners now have 4 cores and 16 GB instead of 2 and 8. Every table was regenerated from run 39469262093, which covers all scenarios, the 100, 1,000 and 5,000-component apps, and all three real apps. Every gate passed, and no pipeline's median absolute deviation went above 2.7% of its median. The earlier 2-core runs (38888167669, 38899025048) stay linked.