Decode HTML entities in excerpts - #2257

I think the other getExcerpt() functions need some attention too; I've marked this PR as a draft until I can review the other classes.

Thanks for starting a pull request. I think the HTML entities should probably be decoded before the text is stored in the DB, that way you won't have to alter every point that the text is used.

I think the HTML entities should probably be decoded before the text is stored in the DB

I wasn't sure if anything else was relying on the stored content being encoded, but if you think that's safe then I'm all for it. I'll replace this commit with a change that stores decoded text in the database.

The new commit saves decoded text in the database for pages. This does affect search functionality, but I think the existing behavior is a bug. Currently you have to use the HTML entity reference in the search string to get the expected results:

This commit fixes the search behavior along with the excerpt rendering. However, the PR will need a database migration to decode existing text content that's already been saved—I'll work on that next.

Hi @mr-vinn,
I agree, That current search behaviour is just due to the existing issue and therefore not correct in itself.

I'd say don't worry about decoding existing content in a migration. This issue is mainly cosmetic and I'd prefer to not have a fairly heavy operation in the migrations to just solve a cosmetic issue. I always try to minimise risk in the migrations where possible.

If a way to update all existing plain text content is really warranted I'd prefer a command to be added which a user could optionally run but I don't think it is.

I'd say don't worry about decoding existing content in a migration.

That works for me; in that case this PR is ready to merge.

Thanks again @mr-vinn, Testing added to prevent regression & merged to be in the next release.