Text::CSV::csv fails with bundled Text::CSV_XS shim

Ouverte
#2,116 0 commentaires 0 réactions 0 personnes assignées Voir sur GitHub

Les mainteneurs répondent en général sous 1 jour

Personne n'a encore pris cette issue.

Évaluation

Difficulté
3/5
Temps estimé
1-2 jours
Accessibilité débutants
68/100
Type d'issue
Bug
Clarté
Clairement spécifiée
Activité
Active
Stack technique
perl
Domaine
backend, testing

Piste de recherche

Start by reading the bundled Text::CSV_XS shim and the Text::CSV wrapper’s direct backend call. Add a focused regression test for Text::CSV::csv; confirm it fails on the unfixed parent and passes under standard Perl and both PerlOnJava backends. Then run WHO-GrowthReference-Table 0.013’s t/00-compile.t on both backends and confirm it passes.

Rédigé par le modèle d'indexation à partir du texte de l'issue.

Description

area:cpan-port bug

Summary

WHO::GrowthReference::Table 0.013 fails its compile test under PerlOnJava because the bundled Text::CSV_XS compatibility shim does not define a package-level csv function. The failure is reproducible on both the JVM and interpreter backends. The same upstream distribution suite passes under system Perl.

Root cause

The bundled Text::CSV wrapper implements Text::CSV::csv by directly calling the selected backend's package symbol (&{"$backend\::csv"}). When it selects Text::CSV_XS, that requires Text::CSV_XS::csv to exist directly in the package stash.

The PerlOnJava Text::CSV_XS shim inherits from Text::CSV_PP and explicitly installs inherited methods into its own stash so direct symbol lookups work. Its export/alias block includes methods such as version, error_diag, and the CSV constants, but omits csv. Inheritance alone does not satisfy the wrapper's direct symbolic call, so the invocation dies with:

Undefined subroutine &Text::CSV_XS::csv called at jar:PERL5LIB/Text/CSV.pm line 97.

WHO::GrowthReference::Table calls Text::CSV::csv in compile-time data initialization. Therefore its t/00-compile.t fails while loading the module.

Reproduction

A minimal reproducer is:

use Text::CSV;
my $rows = Text::CSV::csv(in => "a,b\n");

On PerlOnJava source commit 36d061932880b125578ca915c78613ce469b1675, this fails with the undefined Text::CSV_XS::csv error on both backends:

timeout 60 ./jperl -e 'use Text::CSV; my $x = Text::CSV::csv(in => "a,b\n"); print scalar(@$x), qq(\\n)'
timeout 60 ./jperl --interpreter -e 'use Text::CSV; my $x = Text::CSV::csv(in => "a,b\n"); print scalar(@$x), qq(\\n)'

Both exit 255 with the same error. Under system Perl, the archived CPAN oracle passed the complete WHO::GrowthReference::Table 0.013 test suite: 5 test programs, 1 assertion, all successful.

CPAN evidence

  • Distribution: WHO-GrowthReference-Table 0.013 (WHO::GrowthReference::Table)
  • CPAN random tester run: 20261118-104825-43075
  • Failing test: t/00-compile.t (WHO/GrowthReference/Table.pm loaded ok)
  • PerlOnJava result: 1 of 5 test programs failed; the remaining four were author/release tests skipped by the distribution.
  • System Perl oracle: all 5 test programs passed (1 assertion).
  • Runtime prerequisites used by the failing run: Data::Section::Seekable::Reader 0.092 and Data::TableData::Lookup 0.003; both dependency test suites passed under the system Perl oracle.
  • Reproduction environment: macOS 26 / Darwin 25.6, ARM64, OpenJDK 24.0.2, PerlOnJava Perl 5.44.1.

The closedir() warning in the archived log is unrelated; the test's fatal error is the missing Text::CSV_XS::csv symbol.

Acceptance criteria

  • The bundled Text::CSV_XS backend exposes csv as a package-level function compatible with Text::CSV::csv's direct backend call.
  • Add a focused project regression test for the Text::CSV::csv functional interface.
  • Validate the new test with standard Perl, retain evidence that it fails on the unfixed PerlOnJava parent, and verify it passes on both JVM and interpreter backends.
  • Re-run WHO-GrowthReference-Table 0.013 t/00-compile.t on both PerlOnJava backends.
Langage dominant
Perl
Étoiles
67
Forks
8
Merge moyen
5 h 12 min
PR mergées (30 j)
131

Préparer son environnement

Par où commencer

  1. Lisez l'issue en entier, puis le guide de contribution du projet.
  2. Signalez en commentaire que vous la prenez — cela évite que deux personnes fassent le même travail.
  3. Forkez le dépôt et travaillez sur une branche.
  4. Ouvrez une pull request qui référence le numéro de l'issue.

Autres issues de fglock/PerlOnJava

Toutes les issues de fglock/PerlOnJava

Issues similaires

Plus d'issues Perl

Recevez les nouvelles issues par e-mail

Un résumé court des issues GitHub adaptées aux débutants.