You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 4345294
Browse filesBrowse the repository at this point in the historyBrowse files
Fix slow Go rescan from per-dep go.sum parse (#1424) (#1746)
* Start fix for #1424
Assisted-by: Claude Code:claude-opus-5-5
* Speed up Go rescan on large go.sum files
Hosted Go scans got about a third slower in #1403. For each patched
dependency the rewriter asks go.sum which versions the build loads,
and that tokenized every go.sum line every time. On a project with
1200 modules and 30 patches this was about 30% of the scan's work.
Lines that can't belong to the module are now skipped with one
prefix check before tokenizing, so the answer stays the same. Tests
pin the edge cases: shared path prefixes, /go.mod lines, odd
whitespace and short lines.
Fixes#1424
Assisted-by: Claude Code:claude-opus-5-5
* Parse each go.sum line once in zip_versions
Lines that pass the module prefilter are now parsed directly instead
of going back through the multi-line reader. The equivalence test now
also checks the editor after an edit, when it holds split lines.
Refs #1424
Assisted-by: Claude Code:claude-opus-5-5
---------
Co-authored-by: Claude <noreply@anthropic.com>
0 commit comments