WordPress importer fails on post types with dashes in the slug

[ilicfilip]

Bug

The WordPress importer passes custom post type slugs through to EmDash collection slugs without sanitization. WordPress allows dashes in post type keys (e.g. my-custom-type), but EmDash's SLUG_VALIDATION_PATTERN only allows lowercase letters, numbers, and underscores (/^[a-z][a-z0-9_]*$/). This causes the import to fail for any custom post type containing a dash.

Reproduction

1. Export a WordPress site that has a custom post type with a dash in the key (e.g. book-review)
2. Run emdash import wordpress export.xml --execute
3. Import fails with: collection slug must start with a letter and contain only lowercase letters, numbers, and underscores

Cause

In packages/emdash/src/cli/commands/import/wordpress.ts, the mapPostTypeToCollection function (line ~445) returns unmapped post types as-is:

function mapPostTypeToCollection(postType: string): string {
    const mapping: Record<string, string> = {
        post: "posts",
        page: "pages",
        // ...known mappings
    };
    return mapping[postType] || postType; // no sanitization
}

The slug validation in packages/emdash/src/schema/registry.ts (line 27) then rejects it:

const SLUG_VALIDATION_PATTERN = /^[a-z][a-z0-9_]*$/;

Proposed fix

Sanitize the post type slug in mapPostTypeToCollection before returning it:

function mapPostTypeToCollection(postType: string): string {
    const mapping: Record<string, string> = {
        post: "posts",
        page: "pages",
        // ...
    };
    if (mapping[postType]) return mapping[postType];
    // Convert dashes to underscores to match EmDash slug requirements
    return postType.replace(/-/g, "_");
}

Context

WordPress post type keys: "Must not exceed 20 characters and may only contain lowercase alphanumeric characters, dashes, and underscores." The only incompatibility is the dash, so a simple replace(/-/g, "_") covers all cases.

[ilicfilip]

Closing as there is already PR which fixes this: #83

[ascorbic]

We keep the issue open til the PR is merged