0004675: Compiler fails to warn about pattern Top _ when Top is a 0-ary constructor - MantisBT

Mantis Bug Tracker

View Issue Details [ Jump to Notes ]

[ Issue History ] [ Print ]

ID

Project

Category

View Status

Date Submitted

Last Update

0004675

OCaml

OCaml general

public

2008-12-17 16:07

2009-12-08 17:49

Reporter

Pascal Cuoq

Assigned To

Priority

normal

Severity

minor

Reproducibility

always

Status

closed

Resolution

fixed

Platform

OS

OS Version

Product Version

3.11.0+beta

Target Version

Fixed in Version

3.12.0+dev

Summary

0004675: Compiler fails to warn about pattern Top _ when Top is a 0-ary constructor

Description

# type t = Top | Base of int ;;
type t = Top | Base of int
# let f x = match x with Base _ -> 0 | Top _ -> 1 ;;

Résultat obtenu :
val f : t -> int = <fun>

Résultat attendu :
Error: The constructor Top expects 0 argument(s),
but is here applied to 1 argument(s)

Tags

No tags attached.

Relationships

related to	0004052	closed		Constructeur constant et _
has duplicate	0004717	closed		Ocaml does not warn about invalid wildcard uses
has duplicate	0004835	closed		None _ accepted as pattern match
has duplicate	0004911	closed		weird behaviour of underscore in pattern matching

Notes
(0004804) doligez (manager) 2008-12-17 18:52	This is a feature that was implemented to facilitate automatic generation of OCaml code...

(0004805) Pascal Cuoq (reporter) 2008-12-18 00:53	I can see how some camlp4 extensions may get lighter this way, but the principle behind any single OCaml warning is that it's better to alert the programmer at the first sign that she might be confused about her program (the earlier the better), and this feature is certainly not homogeneous with this (desirable) general principle. Shouldn't the pattern generated by camlp4 be something that does not coincide with something that a human programmer might write by accident and want to hear about? "Top _" is already as ugly as it gets, therefore if a camlp4 extension is generating this it does not look like a very big stretch to force it to generate "Top __" or "Top ((_))" or whatever other construct that 1) would also work if the constructor had arguments, so as not to return to the initial problem, and 2) no-one in her right mind would write.

(0005023) Etienne Millon (reporter) 2009-07-15 15:51	I agree with Pascal Cuoq. It may be nice to factor some camlp4 code, but it creates a nasty compatibility issue.

(0005150) doligez (manager) 2009-11-06 19:15 edited on: 2009-11-09 15:13	In 3.12.0, there will be a warning (off by default) for when you give an _ as argument to a constant constructor.

Issue History
Date Modified	Username	Field	Change
2008-12-17 16:07	Pascal Cuoq	New Issue
2008-12-17 18:51	doligez	Relationship added	related to 0004052
2008-12-17 18:52	doligez	Note Added: 0004804
2008-12-17 18:52	doligez	Status	new => feedback
2008-12-18 00:53	Pascal Cuoq	Note Added: 0004805
2009-02-16 10:05	ertai	Relationship added	has duplicate 0004717
2009-07-15 15:06	xclerc	Relationship added	has duplicate 0004835
2009-07-15 15:51	Etienne Millon	Note Added: 0005023
2009-11-06 19:02	doligez	Relationship added	has duplicate 0004911
2009-11-06 19:15	doligez	Note Added: 0005150
2009-11-09 15:13	doligez	Note Edited: 0005150
2009-11-09 15:13	doligez	Note Edited: 0005150
2009-12-08 17:49	doligez	Status	feedback => closed
2009-12-08 17:49	doligez	Resolution	open => fixed
2009-12-08 17:49	doligez	Fixed in Version	=> 3.12.0+dev

Copyright © 2000 - 2011 MantisBT Group