Diamond operator ignores tied *ARGV after import-time traversal

オープン
#2,111 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
4/5
見積もり時間
3〜5日
初心者へのやさしさ
52/100
issue の種類
バグ
明瞭さ
明確に書かれている
活発さ
活発
技術スタック
java, perl
領域
backend, testing

調査の方向性

Start with OpcodeHandlerExtended.executeReadline, DiamondIO.readline and openNextFile, and TieHandle.tiedReadline to trace how reads are dispatched after *ARGV is tied. Reproduce the failures with the upstream t/01_argv.t commands in the issue, then compare tied-handle reads with ordinary @ARGV iteration and EOF behavior. Done when both backends honor the tied handle’s READLINE after import-time traversal and regression coverage passes.

索引モデルが issue の本文から書いたものです。

説明

area:io bug

Summary

<> stops using *ARGV after ARGV::JSON consumes the command-line file during module import and then ties *ARGV. PerlOnJava returns undef for reads that should invoke the tied handle's READLINE method.

This was reported by CPAN random run 20261118-104825-43075 for ARGV-JSON-0.01, target ARGV::JSON. The failure is in t/01_argv.t (tests 2 and 3, lines 32 and 37); t/00_compile.t passes.

Reproduction and results

The upstream test sets @ARGV to t/data/01.json, then loads ARGV::JSON. Its import method reads the file using <> and ties *ARGV. The test then expects scalar and list-context <> reads to return the decoded JSON values.

From the extracted ARGV-JSON-0.01 distribution directory:

PERL5LIB=blib/lib jperl t/01_argv.t
PERL5LIB=blib/lib jperl --interpreter t/01_argv.t

Both PerlOnJava backends reproduce the two failures. For the first read, the actual value is undef; the expected value is the first decoded hash. The list read is empty instead of returning the remaining array and hash values. The final EOF assertion passes.

The archived system-Perl oracle ran the same distribution and passed all 5 tests (t/00_compile.t and t/01_argv.t) under Perl 5.42.2. The archived PerlOnJava run failed 2 of 5 tests. A local follow-up reproduced the result on both backends; its executable reports build 543f9dfef, on Temurin Java 24.0.2 / macOS 15.6 ARM64.

Root cause

The import-time <> loop starts DiamondIO traversal while @ARGV contains the fixture filename. It consumes that filename and reaches EOF before ARGV::JSON ties *ARGV. The later <> operations are still routed by OpcodeHandlerExtended.executeReadline to DiamondIO.readline. That code continues the existing diamond traversal; when it needs another input, openNextFile shifts from the now-empty @ARGV and reports EOF. It does not dispatch those reads to the newly tied ARGV handle's READLINE method.

Relevant code paths are OpcodeHandlerExtended.executeReadline, DiamondIO.readline / openNextFile, and TieHandle.tiedReadline. The behavior to fix is the interaction between the diamond operator's traversal state and a tied *ARGV: reads after tying must honor the handle's READLINE, while ordinary @ARGV file iteration and EOF behavior remain intact.

Related coverage

The repository has adjacent diamond-state and tied-handle coverage, including the changes in PR #1463, but that coverage does not exercise tying *ARGV after an import-time diamond traversal has consumed its input. The upstream ARGV-JSON-0.01 test is a focused reproducer suitable for project regression coverage.

No XS or native dependency is involved. Searches of existing issues for ARGV::JSON, tied ARGV, and DiamondIO found no issue covering this behavior.

主要言語
Perl
スター
67
フォーク
8
平均マージ
5時間 12分
マージ済み PR(30日)
131

環境構築

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。

fglock/PerlOnJava のほかの issue

fglock/PerlOnJava の issue をすべて見る

似ている issue

Perl 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。