comphead commented on code in PR #7708: URL: https://github.com/apache/datafusion-comet/pull/7708#discussion_r4645585558
########## .ai/skills/bug-triage/SKILL.md: ########## @@ -168,7 +171,72 @@ If `gh issue edit` fails for any issue, leave that issue's `requires-triage` label intact and record the failure in the summary under a "Failed to label" section. -## Step 6: File the Summary Issue +## Step 6: Build the Open Critical Bugs Summary + +The report opens with the state of **all** open critical bugs, not just this +week's. Build it after Step 5 so that critical bugs labeled in this pass are +included. + +Fetch every open critical bug: + +```bash +gh issue list \ + --repo apache/datafusion-comet \ + --label bug --label priority:critical \ + --state open \ + --limit 300 \ + --json number,title,labels,url,createdAt +``` + +**Primary area.** Give each issue exactly one area so the counts add up to the +total. Use the first label from this list that the issue carries: +`area:Iceberg`, `area:Delta`, `area:udf`, `area:writer`, `area:shuffle`, +`area:joins`, `area:aggregation`, `area:memory`, `area:ffi`, `area:scan`, +`area:expressions`, `area:ci`. An issue with none of them goes under +`No area label`. Only read the labels that are already on the issue here; this +step never adds one. + +**New this week.** A critical bug is new if it was in this pass's +`requires-triage` queue and ended up `bug` + `priority:critical`. That includes +issues whose author already applied the labels. Review Comment: "New" is defined through the `requires-triage` queue, and Step 2 fetches every open issue with that label. The label is auto-applied to every new issue by `.github/workflows/label_new_issues.yml`, including the summary issues themselves. https://github.com/apache/datafusion-comet/issues/7707 currently carries `requires-triage` (the closed https://github.com/apache/datafusion-comet/issues/7640 and https://github.com/apache/datafusion-comet/issues/7574 still do too). If a report is still open at the next run, it is in the queue and would be classified as a bug or an enhancement. The body of #7707 says #7640 "was left alone", so the agent handled it ad hoc. This is not covered by the skill. Excluding titles that start with `Bug triage results: ` in Step 2 would make it deterministic. Also, a critical bug that a maintainer relabels to `priority:critical` after triage is counted in the area total and the `+N` but is never marked `NEW`. A `createdAt` within the last 7 days would be a si mpler and more robust definition if that is the intent. ########## .ai/skills/bug-triage/SKILL.md: ########## @@ -168,7 +171,72 @@ If `gh issue edit` fails for any issue, leave that issue's `requires-triage` label intact and record the failure in the summary under a "Failed to label" section. -## Step 6: File the Summary Issue +## Step 6: Build the Open Critical Bugs Summary + +The report opens with the state of **all** open critical bugs, not just this +week's. Build it after Step 5 so that critical bugs labeled in this pass are +included. + +Fetch every open critical bug: + +```bash +gh issue list \ + --repo apache/datafusion-comet \ + --label bug --label priority:critical \ + --state open \ + --limit 300 \ + --json number,title,labels,url,createdAt +``` + +**Primary area.** Give each issue exactly one area so the counts add up to the +total. Use the first label from this list that the issue carries: +`area:Iceberg`, `area:Delta`, `area:udf`, `area:writer`, `area:shuffle`, +`area:joins`, `area:aggregation`, `area:memory`, `area:ffi`, `area:scan`, +`area:expressions`, `area:ci`. An issue with none of them goes under +`No area label`. Only read the labels that are already on the issue here; this +step never adds one. Review Comment: This ordering lists 12 labels but Step 3 (unchanged, lines 99-103) and `docs/source/contributor-guide/bug_triage.md` only tell the classifier to apply 7 of them (`area:writer`, `area:shuffle`, `area:aggregation`, `area:scan`, `area:expressions`, `area:ffi`, `area:ci`). `area:Iceberg`, `area:Delta`, `area:udf`, `area:joins` and `area:memory` are never applied by this skill, so a critical bug triaged in this pass will not get them. https://github.com/apache/datafusion-comet/issues/6898 is an example. It is a join bug, `area:joins` exists, and it lands under `No area label` in https://github.com/apache/datafusion-comet/issues/7707. Please either add the missing labels to Step 3 and the guide table, or say here that only existing labels are read. Also, an issue whose only area label is not in this list (a new `area:*` label, or a rename) is reported as `No area label`, which is wrong. A fallback such as "any other `area:` label goes after the listed ones, alphabetically" avoids that a nd avoids having to edit the skill for every new label. ########## .ai/skills/bug-triage/SKILL.md: ########## @@ -178,44 +246,80 @@ TRIAGE_DATE=$(date -u +%Y-%m-%d) Title: `Bug triage results: ${TRIAGE_DATE}` -Body: a markdown report with these sections, in this order: - -1. **Header** - - Date, total issues processed, count of bugs vs. enhancements, and counts - per priority - - Link to `docs/source/contributor-guide/bug_triage.md` - - Note that labels have already been applied; the reviewer should spot-check - and close this issue when satisfied -2. **Bugs** — one subsection per priority, ordered highest priority first - (`priority:critical`, then `priority:high`, then `priority:medium`, then - `priority:low`). Omit any subsection whose count is zero. Do **not** use a - markdown table anywhere in this section; use nested bullet lists only. +The report is read in a 30 minute meeting, so keep it short. Detail goes where a +decision is needed; everything else is a title and a link. Sections, in this +order: - Within each subsection, one top-level bullet per issue: +1. **Header** — three lines at most: the date; issues processed with the split + into bugs and enhancements; bugs by priority for this pass. Link to + `docs/source/contributor-guide/bug_triage.md`. Say that labels are already + applied and the reviewer should close the issue when satisfied. +2. **Open critical bugs** — the heading must be exactly `## Open critical bugs`, + because next week's run looks for it. Use this layout: ``` - ### priority:critical + ## Open critical bugs + + 25 open, 2 new this week, 1 fewer than last week. + + ### area:Iceberg (8 open, 1 new, +1 vs last week) + + - **NEW** <title> (2 days) ([#1234](url)) - one sentence on why it is critical + - <title> (27 days) ([#1200](url)) + + ### area:expressions (7 open, -1 vs last week) - - <issue title> ([#1234](https://github.com/apache/datafusion-comet/issues/1234)) - - Area labels: `area:expressions`, `area:scan` - - Rationale: one sentence tying the call to the guide + - <title> (12 days) ([#1180](url)) ``` - The issue number (not the title) is the link target. The title is plain - text. If there are no area labels, write `Area labels: none`. - -3. **Enhancements** (omit section if empty) — one top-level bullet per issue in - the same `<title> ([#N](url))` form, with an `Area labels:` sub-bullet and a - one-sentence rationale for classifying it as an enhancement. Enhancements - have no priority subsections. -4. **Escalations to consider** (omit section if empty) — bullet per issue with - the same `<title> ([#N](url))` form, plus a sub-bullet explaining the - trigger from the guide. -5. **Skipped — needs more info** (omit if empty) — bullet per issue with the - same `<title> ([#N](url))` form, plus a sub-bullet explaining what is - missing. -6. **Failed to label** (omit if empty) — bullet per issue with the same - `<title> ([#N](url))` form, plus a sub-bullet quoting the `gh` error. + - Do not add a table. The counts are in the first line and the headings. + - Heading form: `### <area> (<N> open[, <N> new][, <+N|-N|0> vs last week])`. + Leave out `new` when it is zero. The area is the primary area from Step 6, + or `No area label`. Next week's run reads `<N> open` from these headings. + - Order the sections by open count, largest first, and leave out areas with no + open critical bugs unless the previous report had a heading for them. Review Comment: "leave out areas with no open critical bugs unless the previous report had a heading for them" never expires. A zero heading such as `### area:writer (0 open, -2 vs last week)` is itself a heading in this report, so next week's run keeps it again as `(0 open, 0 vs last week)`, and so on forever. Step 6 (line 234) says "dropped to zero", which is the intended meaning. Please state the rule as "the previous report listed it with a count above zero" in both places. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
