These improvements rely on the new JSON operators to be useful. Those improvements were originally in #6522 but now there's an updated version in #6666. Managing things is simpler if I merge these changes now
* feat: expose isMobile to info
* feat: allow access browser info from $tw.browser
* fix: adapt typo
* refactor: only export selected properties
Jermolene 5 hours ago
The trouble is that the properties of bowser.browser are not under our direct control, and so subsequent updates to Bowser might overwrite important properties of our own. I'd rather explicitly import the properties that we choose to support.
* refactor: put things into `is`
* Fix for Bug #6618
This Commit fixes Bug #6618. It is a little bit more complicated than
using one tiddler to store the new value for a field. Because the
following can happen:
* The user types "not-a-date" into the field value of a simple text field.
* The user now selects a field name that uses a HTML5 date editor. The
Editor will show no date because the value cannot be parsed.
* The user saves the tiddler by clicking the checkmark.
Now the date-field contains the value "not-a-date" but the user was not
aware that this will be added. The edit control showed no date (because
the value was invalid) and the user assumed the field was empty and
won't be added to the tiddler.
To prevent this, every kind of field editor gets its own storage tiddler.
Its name is derived from the SHA256-hash of the name of the tiddler that
is returned by the Field Editor Cascade. That way every editor in the
cascade is only seeing its input. As long as the default setup (with one
default editor) is used, everything works like in 5.2.1.
This commit also fixes the bug that the after adding a field the
field-type input box was not focused again.
* Update Documentation for Field Editor Cascade
The fix for bug #6618 makes the handling of the tiddler backing the edit
operation much more complicated. See previous commit "Fix for Bug #6618"
for more details.
* Fix: eventcatcher widget - variables can be undefined
* Fix: selectedNode can be an svg where offsetLeft ... are undefined
* Make check for offsetLeft short
* remove second collectDOMNodeVariables
* Fix lazy all template with user defined macro cause error
Fixes https://github.com/Jermolene/TiddlyWiki5/issues/6637
* fix: exclude the SJCL library when saving
@Jermolene said:
The construction -[type[application/javascript]library[yes]] is used in the core as a rather clumsy way to exclude the SJCL library when saving. The same construction is in the usual $:/core/save/all filter too.
It's possible that we should review unintended side effects of that behaviour, but here we should leave it alone.
A user was wondering how to add a background image to the password screen of an encrypted wiki (https://talk.tiddlywiki.org/t/background-image-on-login-page/3145). It was not obvious that an encrypted wiki is considered as "loading", so I updated this tiddler to hint at this.
I also added a hint to the other system tags $:/tags/RawMarkup.. since they can also be used for the same purpose.
Adding tiddler 'How to update TiddlyWiki to the latest version' which transcludes the existing tiddler 'Upgrading'.
Some people search for term 'update' instead of 'upgrade' .
This batch includes:
- Minor corrections of metadata from my previous translations
- Translation updates for tiddlers tagged WikiText
- New translations for Parser Modes documentation
- Update of outdated core tiddlers translations (ie Open sidebar tab)
* Documentation for indeterminate checkboxes
* Unit tests for indeterminate checkboxes
* Implement indeterminate checkboxes
* Simplify indeterminate checkbox example
* Slightly simplify refresh logic for indeterminate
That five-line if statement can be turned into a simple assignment.
* Use "yes" and "no" for checkbox indeterminate attr
This makes the "indeterminate" attribute of the checkbox widget work the
same way as other boolean attributes of other widgets.
* Fix bug with invertTag attribute
One place in the checkbox widget code was checking invertTag for
Javascript truthiness rather than the value "yes", which could have
produced incorrect results if anyone wrote invertTag="no". Fixed.
jeremy@jermolene.com