ClojureScript

RegExp objects print incorrectly

Details

  • Type: Defect Defect
  • Status: Resolved Resolved
  • Priority: Minor Minor
  • Resolution: Completed
  • Affects Version/s: None
  • Fix Version/s: None
  • Component/s: None
  • Labels:
  • Patch:
    Code and Test

Description

#"x" is currently printing as #</x/>

js/RegExp needs an IPrintable implementation: (-pr-seq [re _] (list "#\"" (.-source re) "\""))

Unfortunately, this is blocked by the same GClosure issue as http://dev.clojure.org/jira/browse/CLJS-68

Attachments

  1. Text File
    print-regex.patch
    04/Jun/12 6:10 PM
    4 kB
    Brandon Bloom
  2. Text File
    print-regexp2.patch
    04/Jun/12 8:42 PM
    2 kB
    Brandon Bloom

Activity

Ascending order - Click to sort in descending order
Hide
Permalink
David Nolen added a comment -

The simplest solution is to add a RegExp case to pr-seq.

Show
David Nolen added a comment - The simplest solution is to add a RegExp case to pr-seq.
Hide
Permalink
Brandon Bloom added a comment -

That was surprisingly not that simple...

(instance? js/RegExp obj) compiles to `cljs.core.instance_QMARK_.call(null, RegExp, obj)` which also triggers the GClosure warning. So I thought I could inline `instance?` with a macro, but the instanceof operator's operands are in reverse order of instance?'s arguments. To preserve evaluation order, the expansion would be `(let t# ~t ('js* "({} instanceof ~{})" ~o t#))) however even that triggers the warning because the output leaves RegExp as an expression by itself again. I wound up just special-casing simple instanceof checks against symbols in the macro.

Attached patch fixes printing of RegExp objects and also enables calling (instance? js/RegExp X) without triggering the warning. Sadly, any more complex expression which evaluates to js/RegExp, or any higher order usage of instance? against that type will still trigger the warning.

Show
Brandon Bloom added a comment - That was surprisingly not that simple... (instance? js/RegExp obj) compiles to `cljs.core.instance_QMARK_.call(null, RegExp, obj)` which also triggers the GClosure warning. So I thought I could inline `instance?` with a macro, but the instanceof operator's operands are in reverse order of instance?'s arguments. To preserve evaluation order, the expansion would be `(let t# ~t ('js* "({} instanceof ~{})" ~o t#))) however even that triggers the warning because the output leaves RegExp as an expression by itself again. I wound up just special-casing simple instanceof checks against symbols in the macro. Attached patch fixes printing of RegExp objects and also enables calling (instance? js/RegExp X) without triggering the warning. Sadly, any more complex expression which evaluates to js/RegExp, or any higher order usage of instance? against that type will still trigger the warning.
Brandon Bloom made changes -
Field Original Value New Value
Patch Code and Test [ 10002 ]
Attachment print-regex.patch [ 11291 ]
Labels patch patch,
Hide
Permalink
Brandon Bloom added a comment -

As discussed, will make a separate ticket for inlining 'instance? and will (defn regexp? [o] (js* "~{o} instanceof RegExp"))

Show
Brandon Bloom added a comment - As discussed, will make a separate ticket for inlining 'instance? and will (defn regexp? [o] (js* "~{o} instanceof RegExp"))
Brandon Bloom made changes -
Attachment print-regexp2.patch [ 11292 ]
Hide
Permalink
Brandon Bloom added a comment -

Updated with simplified patch

Show
Brandon Bloom added a comment - Updated with simplified patch
David Nolen made changes -
Resolution Completed [ 1 ]
Status Open [ 1 ] Resolved [ 5 ]

People

Vote (0)
Watch (1)

Dates

  • Created:
    Updated:
    Resolved:
Atlassian JIRA (v4.4#649-r158309) | Report a problem
Powered by a free Atlassian JIRA open source license for Clojure. Try JIRA - bug tracking software for your team.