Are you able to capture the slow rendering in a short video so I can get a feel for it? Does fiddling with the graphics settings make any difference? Anki opens but all windows are blank - #17 by dae
I’m guessing we’d need to exclude it at state saving time? Without other options available, this may be our best bet.
I am assuming some menu should pop up when you click on the tag icon, because it changes to have a + sign and shows a hand mouse cursor. It doesn’t pop up, however.
If a menu is not supposed to pop up, I suggest implementing a menu with the most often used tags.
the tag selector popup does not disappear when you type the full tag. to reproduce,
focus tag editor
type the full tag name
press tab to select tag
press enter to accept tag
tag selector popup displays the scrollbar even when they are not needed
tag selector text is tiny compared to other fields, and also to the tag selector popup
there’s still no way to make tags not sticky, that is, to clear the tag field after adding a new card. I was hoping for this one. I tag only some of the cards with one specific tag, and of course I end up forgetting to remove the tag for the next card. I have to revisit added cards to fix up the tags afterwards. Please please add a sticky toggle to the tags as well.
I also think that having borders around input field titles, the dotted lines, and a line that separates tags from the rest of the input fields, only create a messy look.
Running the most recent beta on my Mac, the card browser seems not to be able to pick a proper fallback font for the the card overview (see attached screenshot), where it fails to render most Coptic characters (specifically those characters that are in the specific “Coptic” unicode block (and not those few Coptic characters who were encoded early and are in the “Greek and Coptic” unicode block). In the first line of the screenshot, for instance, it fails to render ‘ⲉϥⲧⲙ̅ⲥⲱⲧⲙ̅’ (apart from the ϥ). As can be seen in the rich text fields below, the computer has no problem rendering the text otherwise.
The same problem occurs with a number of other less used scripts, e.g. Syriac, Phoenician, Cuneiform, Egyptian Hieroglyphs etc. As above, there are no problems with displaying these in the text fields at the bottom nor during normal review.
Clicking the button puts the focus in the tag editor. It also exposes the shortcut for doing so, when you hover over it (Ctrl+Shift+T). Additionally if you already typed in the tag, it will accept it.
there’s still no way to make tags not sticky
I’ve had the PR draft somewhere, I’ll see whether I can make it ready before the full release.
I also think that having borders around input field titles, the dotted lines, and a line that separates tags from the rest of the input fields, only create a messy look.
I’ve also had that thought before. What would you suggest to fix this? I’ve personally thought whether it would make sense to put the tags, and the bottom buttons above the fields buttons (bold, italic, etc.)
I think it looks much better than the previous design. It makes sense to have a border around the whole field, because the UI elements of that header (LabelContainer) exclusively affect that field.
Regarding the dashed bottom-border of the header, I agree - it could be omitted. This results in a cleaner look:
Maybe a darker border (between --border and --medium-border) and replacing the bottom sticky border with a background for the tag input would help reduce the clutter?
There are some power-users that would probably cringe at that thought With a lot of space occupied by add-on buttons, moving the tags and add-button etc. on top would create a very imbalanced look. I think the current bottom position of these elements looks good. Is there any specific reason why you’d want to move them above? Or did I misunderstand the suggestion?
Sure, here goes a typical example, with Qt5 the .png image (1.3 mb) is showed nearly instantly:
The same card with Qt6:
Not the end of the world and still perfectly usable, but notably slower. This does not happen with all large images, I have tried different sizes and formats and so far I have not recognised any pattern.
I’m tecnhically very limited and get a little bit lost with the info in that link, sorry. Not really sure about how can I tweak the graphic settings when using the Qt6 version, but if you let me know how can I do that I will be happy to try.
Bug: esc key while previewing a card in Browser throws this error:
Anki 2.1.50 (43c41d76) Python 3.9.7 Qt 6.2.0 PyQt 6.2.0
Platform: Mac 12.0.1
Flags: frz=True ao=False sv=3
Add-ons, last update check: 2021-12-07 20:04:33
Caught exception:
Traceback (most recent call last):
File "aqt.progress", line 53, in handler
File "aqt.browser.previewer", line 319, in _on_close
File "aqt.browser.previewer", line 124, in _on_close
AttributeError: 'NoneType' object has no attribute 'cleanup'
Workaround: Click the exit button in top-left of preview window
Shift-started Anki, and turned off all keyboard/trackpad programs (BetterTouchTools, Karabiner Elements, Keyboard Maestro).
Does selecting an appropriate font in Cards>Browser Appearance help? If not, it may be that Qt’s font rendering can’t handle this case, but Chromium’s can.
Yep, that doesn’t look very good! Could you provide me with that image (privately if necessary)?
This doesn’t seem to happen on a Mac - I’m guessing it’s an issue with the HTML editor component we’re using.
Refreshing that cards in the browser is also a little bit slower in 2.1.50 than in 2.1.49, but (at least in my system) it gets much worst when Qt6 is used. An example using the same deck:
Thanks, I can reproduce the problem here too. 2.1.50 is using Qt 5.15, as the 5.14 version 2.1.49 uses isn’t available for our current tools. Slowdowns when loading large images were observed in the past (second point on Qt 5.15 issues · Issue #982 · ankitects/anki · GitHub). Qt 6 seems to perform even worse.
If you shrink the images down to less than 4000 pixels in either height or width as an experiment, does that make them load quickly?
Is your _Test3.png file (the one with the squares) freely distributable? If so, we can use it to report this to Qt.
Not sure it’s the right setting, but if you mean the “Cards” button below the card list in the browser and then the “Options” button and then “Browser Appearance” and “Override font”, I’ve tried changing it to an appropriate font (Antinoou), and it doesn’t change anything. Still display replacement characters instead of proper glyphs when I return to the card browser. Anyways, that option doesn’t seem like it would be the right place for a setting that would change the font used in the general card browser a few windows back.
Yes, I’ve made some experiments with different image sizes, and it seems that the problem affects only to images larger than 5000 pixels height or width. Here is the test deck, if you want to check yourself.
And yes, the squares image is just a quick doodle made with paint (but it’s nice, isn’t it ), you can use it to report the problem.
Personally, I’m adding only a few tags to my cards so I would enjoy a menu with the most used tags.
Also, most other icons are hollow (not solid) and the tag icon is filled in. This looks a bit inconsitently.
I would also remove the tags icon altogether, and, instead I would put a hint into the tag input filed (as in, text that disappears when you start typing) that reads “Tags” or “Add a tag”, or perhaps combined with the current “Tags” label, “Click to add”, or maybe a “+” button at the end of current tags that would list the most used tags.
I think that the design of the current stable is pretty good? I would remove the line that separates the buttons from the input fields as well; otherwise I like the way it looks. Oh, and I would also remove the sliiightly different background color behind the input elements. Why is it different to the background color of the window itself?
I see what you mean, but I doubt that people are actually confused by the labels. While the labels are not closer to their field than to the previous one, I just don’t see how someone could think that the labels are pointing to the field above? Is this a problem in practice?
I should also say, not putting any borders around individual field labels has been a mainstream design for, like, forever. In all apps the labels just float on top or on the left. Lately even category labels (labels that denote a group of input fields/controls) are not using borders, instead, different font sizes, typefaces and colors are used. If anything, this makes Anki look weird
The “Create copy” feature is really useful, and I’m sure a lot of people are going to love it, thanks!. Maybe, some users might be confused if they select several notes and try to create a copy of all them, so I would suggest either not allowing using that command when several notes are selected, or inform the user that only the first note will be copied in those cases.
On the other hand, the window position when in full screen is still being saved in my system, so I’m still experiencieng this issue:
This action uses the current row, which is not necessarily the first selected, or even part of the selection, if keyboard modifiers were used to create it. It’s the same behvaviour as with the card info dialog, so I wouldn’t special case it. In both cases, the opening dialog should make clear what’s happening.
That’s strange. Are you sure it’s not the invalid state saved with beta 1? If you switch off fullscreen before exiting a window (multple toggles may be required to reach a consistent state), it should be back to normal when reopening it.
I’m not sure if I’ve understood correctly then. When exiting Anki with fullscreen mode off, everyhting works OK, but if I start Anki > F11 > Exit > Start Anki again, the program start in full screen mode and the tool bar/ gear icons are not respondig, as described above. Shouldn’t this prevent Anki to start in fullscreen mode?