mtekk / wp-lynx Goto Github PK

Part of #15, we need to do a few things to protect us from evil people as not just anyone should be able to hammer llynx_scrape. It's likely we'll limit to logged in users, and set a rate limit of 3 requests a minute, along with a quickly rotating API key (nonce, but better than the built in WP nonce which is valid for too long). May even be wise to only allow opengraph scraping from the frontend.

Seems to be a possible JavaScript bug.

The media_buttons_context filter has been deprecated in favor of the media_buttons action.

Looks like WP4.4 will have a new jQuery UI version that changes the alignment of content below UI tabs. Need to fix this before 4.4 is released.

Looks like the show/hide link broke some time in the not so distant past. Need to fix this.

Right now we don't exactly do this, should not pollute the admin.

In WordPress 3.5 the media adding lightbox was updated with a new look and feel, we should update to match that. Will cause our required version of WordPress to bump up to 3.5.

For whatever reason, in 1.0.0 the Lynx Print Template is not longer followed.

General design goals for latest adminKit:

Change to a module loader architecture
Admin stuff move to a module that is only loaded when necessary
Module loader takes care of initializing/loading the options from the DB and loading modules

Need to move from wp_lynx to wp-lynx for the textdomain to be ready for Language Packs.

Remind the user that thumbnails are disabled when the uploads directory can't be written to within the lynx print adding interface.

This is the future of where WP Lynx is going anyways, so might as well get on board.

Ref: https://github.com/WordPress/gutenberg-examples
Ref: https://wordpress.org/gutenberg/handbook/extensibility/

In some cases, we end up grabbing/reporting the same URL twice. Need to determine why this is happening and fix it.

Might as well port over the Breadcrumb NavXT unit tests for the uninstaller since it isn't something that is fun to test manually. Not to mention, it keeps burning me.

Should include these, see if WordPress has a preferred way of switching between dev and .min.x

As part of #30 we introduced PDF handling, let's step it up one notch and switch to the PDF previewer introduced in WordPress 4.7.

Introduce an advanced mode that allows more aggressive image fetching (in particular grabbing background images from CSS).

No longer need/no longer should check for safe_mode in llynxScrape.

This may not be the easiest, but should be possible for users with ImageMagick.

For the image, can use something like
<?php $imagick = new Imagick(); $imagick->readImage('myfile.pdf[0]'); $imagick = $imagick->flattenImages(); $imagick->writeFile('pageone.jpg'); ?>
http://php.net/manual/en/imagick.readimage.php

Since WordPress has the HTTP class, and the make_absolute_url, we can move away from the external library. See https://developer.wordpress.org/reference/classes/wp_http/make_absolute_url/

If we never see a response from the server (e.g. CURL timeout > PHP timeout) we need to handle it in JS rather than just throw JS errors.

Right now we hard set CURLOPT_MAXREDIRS to 3, should let users configure this. See http://wordpress.org/support/topic/maximum-redirects-exceeded?replies=1

Use og:title og:image and og:description as defaults for the scrape of requested URL

See http://ogp.me/

Now that the lynx print adding interface is primarily backbone.js based with JSON communication with the server (as of 1.0.0), we can feasably place this in the frontend.

Right now we have a awful mix of HTML and PHP in the main plugin file. This should be split out akin to what is happening in https://github.com/mtekk/Breadcrumb-NavXT/tree/plugkitdev this is part of #2

Part of #9

Need to update screenshots in the documentation and plugin page along with the documentation for 1.0.0.

Re-add the feature from the pre-1.0 rewrite versions where multiple URLs can be fetched at once.

See http://make.wordpress.org/core/2014/08/21/introducing-plugin-icons-in-the-plugin-installer/

We're going to phase in unit tests and automated testing. llynx_scrape will be the first victim.

Per request, add support for website thumbnails (using external thumbnail provider).

In some cases the user will not have TinyMCE, or the editor instance isn't available. We need to check for this.

Need to instantiate llynx_uninstaller() rather than wp_lynx_uninstaller()

Rather than depend on the server for the entire lynx print assembly, let's do it in JavaScript. The less we have the server do, the faster this will be. Just need to be careful with image saving (need to wait on the image URL from the server).

Should integrate with the revamped post formats in Wordpress 3.6, see if we can hook into the link post format.

We currently try to save bandwidth by making intelligent requests for the headers of images. However, some poorly configured servers may try to gzip or deflate compress their response to this ranged request, which comes back to us a garbage. We should instead tell the server we do not accept gzip or deflate for partial requests.