[RFC]Quote link to quoted message

[brunoais]

• Backward compatibility proved to be too complicated so instead of modifying the original quote_username_open template fragment, it uses a new, separate template fragment named quote_extended which contains XSL for the logic.

That seems fine to me. I think there shouldn't be any issues with that.

• The quote's date is unstyled, it just sits there. The link to the original post is represented by a Unicode arrow. I didn't realize is at the time but that's the same thing XenForo uses. Here's an updated album showing quotes in different forum software.

I think the date should be on the right of the rest of the quote information. IMO, it should be (I do not include the text such as the "L_WROTE"):
{urlToSource {author}} {date}
OR
{sourcePostLink(icon)} {urlToAuthor {author}} {date}
OR
{sourceLink(externalLinkIcon)} {urlToAuthor {author}} {date}

If it can be confusing if the author link can have two different things, then use these instead:
{sourcePostLink(icon)} {urlToAuthor {author}} {date}
OR
{sourceLink(externalLinkIcon)} {author} {date}
OR
{sourceLink(externalLinkIcon)} {urlToAuthor {author}} {date}

Opinions?

• The post link contains a small bit of inline JavaScript that replaces its full URL with an anchor link if the post exists on the current page. If someone has a better solution, I'll gladly accept PRs against my repository.

I like the concept. I don't like that it is inline. I think it should be external js and you should use a data-* attribute on it.
Then, develop javascript that searches for all tags with that data-* attribute and execute the check and edit on it. If you'd rather have me doing it, just tell me and I'll make a proposal.

[JoshyPHP]

I think it should be external js and you should use a data-* attribute on it.

What would be the upside though? Moving that code to an external file means either adding a bunch of listeners (one on every link) or one listener that will trigger on every click made anywhere in the topic. It's going to execute some code on every page load regardless of whether the user clicks any links, and it will be active only after the page is loaded. It will require more code and be less efficient with it.

[JoshyPHP]

Edit: one test fails on Travis's PHP 5.3 with the following error message. PHP 5.3 has been removed from my Linux distribution and the test passes on my local version (PHP 5.3.29-pl0-gentoo) so debugging it is especially challenging.

DateTime::setTimezone(): Can only do this for zones with ID for now

Looked into this bug. I can't reproduce locally and the test runs fine on Travis when it's run apart from the rest of the test suite. It's probably an interaction with some other test in the suite, I have no idea. If I try mocking phpbb\user::format_date() the test suite commits sudoku. I may very well just yank the test and pretend it never existed. Disregard that.

[brunoais]

I think it should be external js and you should use a data-* attribute on it.

What would be the upside though? Moving that code to an external file means either adding a bunch of listeners (one on every link) or one listener that will trigger on every click made anywhere in the topic. It's going to execute some code on every page load regardless of whether the user clicks any links, and it will be active only after the page is loaded. It will require more code and be less efficient with it.

The loss of efficiency goes only to the DOM search. The number of listeners is the same.
Besides, this is an useful piece of code that can be use in many different places, not just the link in the quote tag.

[JoshyPHP]

Then I guess you should write it and see with Nicofuma whether he wants it in.

[Nicofuma]

I like it

[brunoais]

Which thing, Nicofuma?

[Nicofuma]

the data-* attribute. That's something we could use in other places.

Implemented by https://area51.phpbb.com/phpBB/viewtopi ... 23&t=48026 and deployed on area51