Hacktoberfest 2026:メンテナが10月に向けて印を付けた、オープンで初心者向けの issue。 Hacktoberfest の issue を見る

wip: improve search highlight experience

オープン
#1,000 コメント 1 件 リアクション 0 件 担当者 0 名 GitHub で見る

メンテナーはふだん 1 日以内に返信

まだ誰も着手していません。

評価

難易度
5/5
見積もり時間
1週間以上
初心者へのやさしさ
35/100
issue の種類
機能追加
明瞭さ
おおむね明確
活発さ
活発
技術スタック
git, typescript

調査の方向性

Start by reading the existing diff and context-plugin code, then inspect how whiteboard’s data model represents regions and evidence; the issue does not name specific files or tests. The proposal spans a new diffr search command, diff caching, and whiteboard integration, so clarify the interface and scope before taking it on. Done means highlighted regions can be surfaced in code peeks and diff lenses, with the surrounding diff folded appropriately.

索引モデルが issue の本文から書いたものです。

説明

We need support for actual highlights within regions in both code peeks + diff lenses. Example - you might imagine that the model may only want to show you lines 22-28 of this, but you're forced to see the whole thing:

Image

In the sense that a model is going to select a region of code to show you - it may not be the entire diff. But showing only those lines is sad because you lose context for the surrounding code (which is sad as you start to get a feeling for the system).

A tentative approach:

  • a 'diffr search' subcommand that:

    • takes in 'git grep' shaped input (commit:fileName:lineNo: text)
    • joins that with diffr-computed diffs
    • runs the context plugin to only show the region highlighted (within the context of the larger diff)
    • i'd imagine you'd only want to show the highlighted region at first, but should treat e.g. a highlighted region within a changed region as ok.
  • some wiring into whiteboard so that the data model accepts this

    • i think the mcp structure for "give me a region to show within a diff" is also a bit weird.
    • right now the model needs to give you precise line numbers etc.
    • instead, you could imagine the model could run:
      git grep <pattern> | diffr search and pipe that into whiteboard as evidence
      • diffr search returns a list of hits (so we can work in a decision model somehow)
  • other notes: computing diffs is expensive / slow. so i thought we could read cached diffs from .cache/diffr or something like that (the XDG standard or whatever)

  • internally, i broke up search into 'hydrate' -> 'postprocess' to resemble a normal search pipeline. this would let us insert e.g. jev/openai decisions at parts in the pipeline to filter results if we wanted to !

主要言語
TypeScript
スター
2.9k
フォーク
145
平均マージ
16時間 28分
マージ済み PR(30日)
553

環境構築

このプロジェクトには開発コンテナ、Dockerfile、コントリビューションガイドがありません。まず README を読み、一般的な手順ははじめてのコントリビューションガイドを参照してください。

はじめの一歩

  1. issue を最後まで読み、次にプロジェクトのコントリビューションガイドを読みます。
  2. 着手することを issue にコメントします — 二人が同じ作業をするのを防げます。
  3. リポジトリをフォークし、ブランチを切って変更します。
  4. issue 番号を参照したプルリクエストを送ります。

devdotfast/whiteboard のほかの issue

devdotfast/whiteboard の issue をすべて見る

似ている issue

TypeScript の issue をもっと見る

新しい issue をメールで受け取る

初心者向けの GitHub issue を短くまとめたダイジェスト。