simple bug: Inconsistencies around null and undefined

[amaramth]

In code, plist.build({foo:undefined}) === plist.build({}) and plist.build({foo:null}) === plist.build(plist.parse(plist.build({foo:null}))) are not true and I think they should be.

exposition:

Javascript has null and undefined and [key missing]; JSON has null and [key missing]; plist has [key has no value] and [key missing]. I derive this from plist.build({foo:null}) !== plist.build({}), which is true and seems like it should be - this assumption might be false, though.

JSON.stringify,parse converts

• javascript null ↔ json null
• javascript [key missing] ↔ json [key missing]
• javascript undefined → json [key missing]

plist.build,parse should do likewise, converting

• javascript null ↔ plist [key has no value]
• javascript [key missing] ↔ plist [key missing]
• javascript undefined → plist [key missing].

Currently, it actually does

• javascript null,undefined → plist [key has no value]
• plist [key has no value] → javascript [key missing]
• javascript [key missing] ↔ plist [key missing]

Can you reply with clarification? I don't know if that's the most consistent way to change this, or if it should be changed.

[TooTallNate]

👍 PR welcome

[mreinstein]

technically this would be an API change, so if you send a PR this would require bumping the major version on publish. Just putting this here as a reminder if/when that happens.

[lordgreg]

This i causing exactly the issue a-lot of Cordova users currently have:

Here is an example:

Cordova adds two keys with nullified values to plist, which plist reads them out:

  NSMainNibFile: null,
  'NSMainNibFile~ipad': null,

What plist (latest) is then doing is, it writes them back to plist (including all others possible options we write to config.xml). The objects above become:

    <key>NSMainNibFile</key>
    <value>NSMainNibFile~ipad</value>

Expected output:

--no output--

One option currently is to use the version 1.2.0 (yea, 1.2.0!) or, the hacky way to just remove both keys from plist "manually" from the string that plist module generates.

Both of them are totally wrong.

[mreinstein]

@amaramth I know this is a crazy-late reply, but I'm getting ready to start tackling a v4 of plist.js. If you're still interested in having this change go into a new major version, it's upcoming.