Repository navigation
Comment the memory of every implementation on pull requests - #21
Merged
Merged
Conversation
The workflow of each implementation uploads the numbers of its measure scenario as the artifact memory-<implementation>. shared/tools/measure.sh writes them to the file in VFOX_PLUS_MEASURE_OUTPUT as well as to the log; Flutter's tool/measure.py already wrote memory.txt. memory.yml starts after each of those workflows ends, in the context of main, so that it also works for pull requests of forks, and .github/scripts/memory_report.py stops while another one is still running for the commit. The run after the last one writes one comment on the pull request: the table of docs/memory.md, sorted by the PSS of the Home page, with the difference from the latest run of the same workflow on main. An implementation that the pull request did not rebuild shows the numbers of main. Later commits replace the comment instead of adding another one. The same code does not use the same memory in every CI run. In the 107 runs of 2026-10-18 and 19 that shared their source with another one (30 groups of 2 to 8 runs), the PSS differed by up to 4.3 % for mygo, Uno, Capacitor, Tauri, Dioxus and Wails, by up to 7.5 % for Compose and Flet, and by 11 to 13 % for Gio and Flutter. A difference is a change, in bold, only when it is more than 5 MB and 5 % of the number on main (8 % for Compose and Flet, 12 % for Gio, 15 % for Flutter); none of the 940 comparisons of identical code was one by these limits. A different number of processes is a change too: it never varied. The script reads nothing but two lines of numbers from the artifacts of the pull request, which are not trusted. Its tests (83, run by the Repository workflow) use a fake of the API, and were also checked against the real one with read-only requests.
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
py/clear-text-logging-sensitive-data treats every name that contains "trusted" as secret data. It flagged the artifact name and ID that read_artifact() logs when it cannot read an artifact (memory_report.py, line 418), through the list that baseline_of() called "trusted". Nothing secret is logged: the list only held the artifacts that were made on the default branch.
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
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.
The workflow of each implementation uploads the numbers of its measure scenario as the artifact memory-. shared/tools/measure.sh writes them to the file in VFOX_PLUS_MEASURE_OUTPUT as well as to the log; Flutter's tool/measure.py already wrote memory.txt.
memory.yml starts after each of those workflows ends, in the context of main, so that it also works for pull requests of forks, and .github/scripts/memory_report.py stops while another one is still running for the commit. The run after the last one writes one comment on the pull request: the table of docs/memory.md, sorted by the PSS of the Home page, with the difference from the latest run of the same workflow on main. An implementation that the pull request did not rebuild shows the numbers of main. Later commits replace the comment instead of adding another one.
The same code does not use the same memory in every CI run. In the 107 runs of 2026-10-18 and 19 that shared their source with another one (30 groups of 2 to 8 runs), the PSS differed by up to 4.3 % for mygo, Uno, Capacitor, Tauri, Dioxus and Wails, by up to 7.5 % for Compose and Flet, and by 11 to 13 % for Gio and Flutter. A difference is a change, in bold, only when it is more than 5 MB and 5 % of the number on main (8 % for Compose and Flet, 12 % for Gio, 15 % for Flutter); none of the 940 comparisons of identical code was one by these limits. A different number of processes is a change too: it never varied.
The script reads nothing but two lines of numbers from the artifacts of the pull request, which are not trusted. Its tests (83, run by the Repository workflow) use a fake of the API, and were also checked against the real one with read-only requests.