Embedded images cause problems

On my site it seems that sometimes people add images to pages by copying and pasting them directly into the TinyMCE editor. I guess this is easier than adding an attachment. This results in base64 encoded image data directly in the topic .txt file instead of in an attachment. This causes a couple of problems:

Pages are quite slow to edit - especially in natedit wiki text mode where you can actually see all that binary data encoded as text.
Pages are VERY slow to index (I'm using Solr). I've see the CPU pegged and indexing stalled for a very long time (half an hour or more) on some of these topics.

I'm wondering if there's something we can do about this. For example, could the Wysiwyg plugin automatically recognize these images and - during the html2tml conversion or on save - convert them to attachments and update the image elements accordingly? This would be slow but would leave the topicin good shape for indexing and future edits. Generic attachment names would be fine.

I've manually fixed the topics on my site that had this problem, but the problem is sure to come up again since there is nothing to prevent users from inserting images this way.

Thanks in advance for whwatever help you can offer.

-- LeilaPearson - 16 Apr 2014

Interesting question. Not sure if there's any simple solution that doesn't require some heavy recoding in the TinyMCEPlugin. It's possible, I think, as TopicInteractionPlugin does it.

-- Main.CrawfordCurrie - 07 May 2017 - 09:20