Easyling Release Notes January, 2023

January has been a month of behind-the-scenes preparation and bug-fixing. I’m not supposed to tell you yet, but rest assured we’re working on some very cool stuff. The two main highlights for this month are additional information on the state of crawls and support for the srcset attribute in the JavaScript translation engine. Read the full article for the details!

srcset in JavaScript

The JavaScript translation engine provides better compatibility than the Translation proxy most of the time. However, this hasn’t been the case for the srcset HTML attribute. This attribute allows developers to specify different images depending on various factors such as screen size. This allows optimisation of the resources to be loaded, ensuring that the correct one is displayed.

We fixed this lack of support in January. If you need to localise images specified by srcset in a project that uses Basic publishing, you can do so in the same way as with the Translation proxy.

Crawl state information

We have added a new section to the Request summary card in the Crawl list. It shows statistics for pages that couldn’t be processed during the crawl. Don’t treat these as warnings. Depending on your settings, unprocessed pages can be perfectly normal. Take the screenshot below, for example. I took it after recrawling our own site, easyling.com. I set the crawl to check the current page list. There are 35 pages where the response code is an error code, but error pages are not processed. These are pages that used to be on the site but were later removed. As they are still in the Pages list, the crawler tried to visit them. This is perfectly normal, but I should probably clean up my Pages list. There are also 8 pages that are redirects. I’ve set the crawler not to follow them, so it’s normal for them to be skipped.

Longer APT

Auto pre-translate, or APT for short, allows you to collect new content for a certain amount of time before it’s sent to third-party systems for processing. Due to technological limitations, this time has to be limited. With a release this month, we have increased the limit to 1440 minutes, one day.

Miscellaneous

It’s a bit of a tradition at this point to have a list of minor improvements, isn’t it? We’ll do it again this month, just in case.

• We have prepared the proxy to handle an edge case where JavaScript code has HTML code embedded in it and the code relies on attribute names being enclosed in single quotes (e.g. foo='bar' instead of foo="bar"). The previous implementation would change the ' to ".
• We have fixed an issue with mass page exclusion that popped up when some of the items to be excluded were already excluded.
• We have improved the new content ingestion algorithm of the JavaScript translation engine. It can now detect if the page has already been translated by Crebbl or Google Chrome’s built-in Google Translate and avoid ingesting translated content.
• A workaround was implemented to enable support for code containing optional chaining (the ?. operator) in JSON/JS/XML processing.
• We’ve fixed a bug where the translator script would insert garbled translations if the original site had a different character set.
• We fixed a bug where changing one crest setting would delete the others.
• Finally, we fixed the “Find URLs in JavaScripts” feature for non-embedded JavaScript content (e.g. JavaScript code directly in script tags.)