Xarantolus Collect Versions Save

You can now filter by multiple domains, e.g. if hosted on http://yourserver:100, you can now visit

http://yourserver:100/site/example.com+example2.com 

to see all sites from example.com and example2.com.

See here on how to update your installation.

From #55:

• Update the website-scraper dependency to version 4
• Small consistency fixes when running on Windows
• Add a /details/:id/index route that redirects you to the main index page of the entry

See here on how to update your installation.

This release fixes a few bugs I noticed:

• Do not simultaneously write into the data file (that way it should no longer get corrupted) #52

Other improvements:

• Videos are now downloaded in a higher quality (youtube-dl's "best" video format is not the highest quality possible) #54

See here on how to update your installation.

Add options for public access #41 (thanks TeoGoddet).

The new options allow_public_view and allow_public_all allow configuring if sites can be viewed without logging in and if logging in is required at all.

If you added an url like https://example.com/some%20file.pdf, the server would download it an save it as some%20file.pdf. This would break the link to the file so you couldn't access it directly. This release fixes this issue by saving that file as some file.pdf, which means that the link will work correctly.

The server should no longer throw an error if the content file doesn't exist, but treat it as if it contained an empty array. This file is now also created at startup so this only happens when you delete it while the server is running.

It now also checks the more generic error code ENOENT instead of a specific one that only worked on windows.

See #39 for more details on the problem.

Page generation should now be faster as it generates HTML elements instead of a html string that needed to be parsed again. Depending on the number of elements on the page this was taking long.

This release fixes the following bugs:

• You couldn't type a 'V' in the title input on the new page because it was used as keyboard shortcut for the url field. This shortcut will now only work in the url input.
• If youtube-dl doesn't provide a title, the server will no longer crash (youtube-dl usually returns a title, so this wasn't really an issue).

The documentation has been expanded:

• User Guide: a general guide on how to use the server. Includes keyboard shortcuts and all kinds of information, e.g. how video downloads work and their limitations regarding cookies.
• An error in the API documentation has been fixed: Some zip implementations will not be able to open files bigger than 4GB, e.g. the windows explorer. Files are not corrupted and can be opened using other programs.

This release adds the following keyboard shortcuts:

• General
  • ESC to return to the main page
  • N to go to the 'new' page to add an url
  • Space to scroll up/down depending on the current scroll height
• Specific
  • Shift + V to fill in the 'video:' text in the url field when adding an entry on the 'new' page

Instead of comparing strings when setting event listeners, the function now checks the current state in order to avoid unnecessary loops.

In short: This speeds up the frontend when a page is generated