Shebang callback support in language bindings

Ouverte
  • Quelque chose ne va pas ?
#4,344 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
5/5
Temps estimé
Plus d'une semaine
Accessibilité débutants
30/100
Type d'issue
Fonctionnalité
Clarté
À clarifier
Activité
Active
Stack technique
c, ruby, rust
Domaine
backend-api-design

Piste de recherche

Start with the Rust binding callback branch and the referenced main_script option in ext/prism/extension.c, then compare how Prism’s language bindings handle parser options. The issue asks whether shebang processing belongs in Prism or in each binding, including the effects of flags such as -n on the AST. Done means that design question is resolved and the chosen behavior is supported consistently.

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

Description

related to https://github.com/fables-tales/rubyfmt/issues/916

Some files are only syntactically valid after they've been processed with shebang comments, e.g.

#!/usr/bin/env ruby -np
next

Currently, the way these are handled is by passing a shebang_callback which reads the flags and sets the appropriate options (like here in CRuby).

I'd like for rubyfmt to be able to support files like this, and as for the immediate solution it wouldn't be too hard to add support for shebang callbacks in the Rust bindings (I've got a branch with those changes), but I suppose my broader question is whether or not language bindings all having to reinvent shebang handling is the desired end result.

Right now, this example also fails in the prism gem (Prism::parse('#!/usr/bin/env ruby -np\nnext') returns errors), and presumably does the same for all the language bindings, and yet echo '#!/usr/bin/ruby -n\nnext' | ruby -c says the syntax is OK, which is somewhat surprising and leads to things like ruby-lsp also reporting errors on similar files. That said, a reasonable argument for Prism not doing this handling by default is that some shebang flags (like -n) result in nodes being inserted into the AST, and it might not be ideal to do that without some kind of explicit choice -- though as it currently stands it does require the main_script option to be set, which may be sufficient.

So my question is, is the preferred approach for users of Prism to re-implement shebang handling (and it can be exposed as an option by the various language bindings), or is there some world where Prism itself handles shebangs?

Langage dominant
C
Étoiles
1k
Forks
202
Merge moyen
6 h 18 min
PR mergées (30 j)
77

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 ruby/prism

Toutes les issues de ruby/prism

Issues similaires

Plus d'issues C

Recevez les nouvelles issues par e-mail

Un résumé court des issues GitHub adaptées aux débutants.