Inconsistent precedence in Markdown parser #42

New issue

Open

opened 2024-06-14 14:25:26 +02:00 by danieldongit · 1 comment

danieldongit commented

2024-06-14 14:25:26 +02:00

(Migrated from github.com)

There seems to be a conflict between my custom inline parser and image tags.
This is my code:

const copyOpen = '[copy]';
const copyClose = '[/copy]';
export const CopyExtension: MarkdownConfig = {
    defineNodes: [NodeTypes.copy, NodeTypes.copyTag],
    parseInline: [{
        name: NodeTypes.copy,
        parse(cx: InlineContext, next: number, pos: number) {
            const isOpenTag = cx.slice(pos, pos + copyOpen.length) === copyOpen;
            const isCloseTag = cx.slice(pos, pos + copyClose.length) === copyClose;

            if (isOpenTag) {
                return cx.addDelimiter(CopyDelim, pos, pos + copyOpen.length, true, false);
            } else if (isCloseTag) {
                return cx.addDelimiter(CopyDelim, pos, pos + copyClose.length, false, true);
            }
            return -1;
        },
        after: NodeTypes.italic,
    }],
    props: [
        styleTags({
            [NodeTypes.copyTag]: tags.special(CustomTags.CopyTag),
            [NodeTypes.copy]: tags.special(CustomTags.Copy),
        }),
    ],
};

This issue is that my copy tag gets mistaken for on image when placing a ! character before the closing tag.

Since it is parsed after Italic it should be parsed before images. If I set it to be before image it gets interpreted as a link.
I don't see what else can be done on my end to fix this.

Here are some screenshots of the behaviour:

There seems to be a conflict between my custom inline parser and image tags. This is my code: ``` const copyOpen = '[copy]'; const copyClose = '[/copy]'; export const CopyExtension: MarkdownConfig = { defineNodes: [NodeTypes.copy, NodeTypes.copyTag], parseInline: [{ name: NodeTypes.copy, parse(cx: InlineContext, next: number, pos: number) { const isOpenTag = cx.slice(pos, pos + copyOpen.length) === copyOpen; const isCloseTag = cx.slice(pos, pos + copyClose.length) === copyClose; if (isOpenTag) { return cx.addDelimiter(CopyDelim, pos, pos + copyOpen.length, true, false); } else if (isCloseTag) { return cx.addDelimiter(CopyDelim, pos, pos + copyClose.length, false, true); } return -1; }, after: NodeTypes.italic, }], props: [ styleTags({ [NodeTypes.copyTag]: tags.special(CustomTags.CopyTag), [NodeTypes.copy]: tags.special(CustomTags.Copy), }), ], }; ``` This issue is that my copy tag gets mistaken for on image when placing a `!` character before the closing tag. Since it is parsed after Italic it should be parsed before images. If I set it to be before image it gets interpreted as a link. I don't see what else can be done on my end to fix this. Here are some screenshots of the behaviour: ![image](https://github.com/lezer-parser/markdown/assets/150142406/cb7a886c-72ff-407e-93bb-663e59b12884) ![image](https://github.com/lezer-parser/markdown/assets/150142406/3524bd11-3899-4bd8-aaf7-170dd081ea01)

marijnh commented

2024-06-14 17:02:49 +02:00

(Migrated from github.com)

The order in which inline parsers get applied applies to the first character they match. So at the !, all inline parsers would be tried, in order, and since your custom syntax doesn't match there yet, even if it has a higher precedence than images, it won't match yet, but then the image parser will match, and consume the bracket syntax. As such, you may not be able to parse this notation in combination with images using @lezer/markdown, at the moment.

The order in which inline parsers get applied applies to the first character they match. So at the `!`, all inline parsers would be tried, in order, and since your custom syntax doesn't match there yet, even if it has a higher precedence than images, it won't match yet, but then the image parser will match, and consume the bracket syntax. As such, you may not be able to parse this notation in combination with images using @lezer/markdown, at the moment.

No labels

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

lezer/markdown#42

No description provided.

Rows
Columns