Small things and bugfixes, nothing fancy this time.
- Changed order of the charset detection, according to w3.org rules
- Better support of crazy pages with multiple charset instructions.
- In absence of the body (or links in body) uKeeper will use the subject for link extraction
- Fixed an issue with empty subject caused errors in some cases
- Changed an attachment flag for “binary” files in order to make it recognizable by gmail web client
Added personalized “report error” link to the header of every email. Obviously, we expect users to click on this link only in case if the received page was not processed right, or seems to be incomplete or inaccurate. Don’t click it just for fun, but to report the real issue.
In order to keep effective processing and smooth capturing, the uKeeper relies on the voluntary reporting of incorrect or ugly looking emails. uKeeper uses these data to maintain high quality and provide quick fixes. Your report may be the critical action that prompts a modification of the service, improves its accuracy and leads to increased overall product quality.
- Supports for “binary” links. You can send a direct link to image or pdf and uKeeper will send it back as an
attachment.
- Preparation for Chrome’s plugin (JSON request/response)
- Fixed embedding of images with relative address
- Fixed encoding detection for some cases with bad HTTP response and partial HTML charset info.
- Multi-line title normalized for the email’s subject
- Supports for text-only pages (content-type text/plain). Because such pages don’t have titles, ukeeper will always use a subject, regardless ! modifier.
There is a known issue with some binary links sent to evernote. This is true for images only, not pdf. Evernote seems to duplicate attached pictures. I’m trying to understand the root cause and fix it somehow
- Bookmarklet’s code for manual creation generated automatically as a part of confirmation email
- Fixed charset / encoding issue for Kindle, in both article and document name
- Added …@free.kindle.com as a valid kindle’s destination
- Preserving original title in case if it is starting with “!”
- Some anti-flood protection in WEB (bookmarklet) mode
- Added uKeeper-web (bookmarklet). See extras for more details. Available for registered users only. Currently supports NICE capturing only. Thx to Aleksandr Ch. for invaluable help with all JS magic involved on browser-side.
- Improved error handling and better error messages
- SSL certificates installed and approved. Starting to protect all communications by pretty strong (2048 key) encryption.
- Kindle support was activated. Any link sent to [email protected] from an account with kindle’s forwarding email will be processed in kindle-compatible mode. Technically it means creating email with attachment, get rid of pictures and format body in some special way.
- NICE-mode embeds pictures into the email body to makes all pictures preservable.
- Fixed problem with some servers stupidly ignoring Accept-Encoding header. In those cases users received compressed garbage instead of good-looking email. Not anymore.
- Added preliminary support for attachment mode. Will be useful for Kindle, not activated yet
- Some preparation for full-pdf capturing on multiple nodes
For last two days I have been working on several interesting additions. Technically, these new functions almost ready for prime-time, but I don’t have a clear view how to release it and if I even want to release it to all users. This is the story:
I have started implementation of kindle-friendly processing and immediately run into several strange issues. In addition, I don’t use Kindle regularly and don’t have a clear view on Kindle related use-cases. So, asking for your feedback / advise / bright ideas.
We have moved web server from Google’s blogger, to AWS S3 cloud. At the same time we have switched to the Octopress engine. Registration links adjusted as well, but untill all DNS catch up with ukeeper.com, new registrations could be a little problematic. Core functionality should not be affected by this, just new registrations for some users. Hopefully, in a couple of hours everything will get back to normal.