jsonsublimetext3sublimetexthtmltidysublimelinter

SublimeLinter ignore_match not working when extension is specified

I'm trying to get SublimeLinter and HTMLTidy to ignore certain php warnings. Specifically the missing <!DOCTYPE> declaration, inserting implicit <body>, and inserting missing 'title' element. It tells you right in the docs how to do this:

{
"ignore_match": [
    "missing <!DOCTYPE> declaration",
    "inserting (?:implicit <body>|missing 'title' element)"
    ]
}

Source: Linter Settings

Adding that to my SublimeLinter user settings file worked fine. The problem is that I want it only to apply to PHP files, because using just the above will also ignore it on HTML files.

The docs have an example for applying ignore_match only to certain extensions:

{
"ignore_match": {
    "inc": [
        "missing <!DOCTYPE> declaration",
        "inserting (?:implicit <body>|missing 'title' element)"
        ]
}

So I copied that and changed the inc to php and I still get those warnings on php files.

Here is what the htmltidy section of my SublimeLinter settings file looks like:

"htmltidy": {
            "@disable": false,
            "args": [],
            "excludes": [],
            "ignore_match": {
                "php": [
                    "missing <!DOCTYPE> declaration",
                    "inserting (?:implicit <body>|missing 'title' element)"
                ]
            }
        }

Any idea why ignore_match works as it is supposed to until I specify an extension that the ignore_match rule is supposed to apply to?

Solution

  • This is SublimeLinter bug. The settings tokenizer allows to replace tokens like ${project} or ${home} in settings. It tries to do it recursively but mess up when you use nested list inside ignore_match dictionary.

    I've tested it and it changes all strings, wraps it to list and assigns the list name as key to dictionary, so:

    {"ignore_match": { "php": ["Error1", "Error2"]}}

    is converted to: {"ignore_match": None, "php": ["Error1","Error2"]}}

    It's quite hard to debug what is going on so I have to recommend saving ignore_match settings before tokenizing and reset after it tokenize. Something like that in lint/linter.py get_merged_settings:

    ims = None
if 'ignore_match' in settings.keys():
    ims = settings['ignore_match']
    del settings['ignore_match']
self.replace_settings_tokens(settings)
if ims:
    settings['ignore_match'] = ims