Skip to content

Parse PmLog metadata for multi-digit PID and TID fields - #449

Merged
mariotaku merged 1 commit into
mainfrom
fix/pmlog-parser
Oct 24, 2026
Merged

mariotaku merged 1 commit into
mainfrom
fix/pmlog-parser

Conversation

@mariotaku

@mariotaku mariotaku commented Oct 24, 2026 •

Copy link
Copy Markdown
Member

Refs #354 as a related log-metadata prerequisite.

PmLog records with realistic IDs such as [1234:5678] fell back to an unknown process/context and kept the entire raw line as the message. The parser allowed only one character in each PID/TID field. Allow multi-character fields while preserving the existing single-character and empty [] forms.

Validation:

  • Nine fixtures exercise the public logread() stream: multi-digit/mixed-length IDs, single digits, empty IDs, timestamp/monotonic/severity/process/context/msgid metadata, nested JSON extras and message text, fallback records, and blank lines.
  • Original production parser: 3 behavioral failures / 6 passes. Fixed parser: all nine pass.
  • npm test: 131 tests passed across 15 files.
  • npm run ng -- build --configuration=development: passed (existing browser-support/Sass warnings).
  • git diff --check: passed.

No TV was used; fixtures reproduce the log-record parsing failure through a mocked command stream. Filtering and export remain separate work in #354.

Integration validation: #449, #450, and #451 apply together without conflicts; all 213 combined headless tests pass. A current public repository snapshot with 56 packages passes the combined page validation and same-version comparison checks.

@mariotaku
mariotaku merged commit 78c1416 into main Oct 24, 2026
8 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant