3.2.x Bug Report / Feedback Thread

Shun

状元
Shouldn't have anything to do with database size, no - SQLite is very lean in its RAM usage (only brings a little piece of the database into memory at a time) and in any event nothing in Settings / Flashcards ought to cause any extra database data to be reloaded since it's actually just adjusting in-app settings relating to flashcards.
Interesting, good to know.

It could be something iCloud related, since even if you're not using iCloud sync the app checks to see whether you have a dormant iCloud database when you enter that screen - does Pleco have permission to use iCloud storage? (if so, does it get rid of the crashes if you turn that permission off?)
I turned that permission setting off, and no crashes so far. I will have to give it some time and will get back to you in case I still get the error.
 

rizen suha

状元
thx, two more:
1) in reader (on ipad, havent tried iphone) new bookmarks are no longer saved after app close.
2) when app opens, reader resumes / points to last read page "before update", ignoring where you actually left off the last time app was open.
 

mikelove

皇帝
Staff member
1) in reader (on ipad, havent tried iphone) new bookmarks are no longer saved after app close.
2) when app opens, reader resumes / points to last read page "before update", ignoring where you actually left off the last time app was open.
Hmm... what type of document? And are you on iOS 14 or 15?
 

mikelove

皇帝
Staff member
@lostinasia - went back and I think we were able to reproduce your issues on an iPad now; had to do with a weird view initialization glitch in the new iOS SDK. Anyway should hopefully be fixed in a day or so whenever Apple approves our next update.
 

rizen suha

状元
fyi, new hanzi neural hand writing recognition:
~100 hanzi tried => 1 infrequent hanzi failed = 隺 (modern use almost exclusively as component). in non-neural hwr 隺 hanzi is recognized. thats all i know.
 

mikelove

皇帝
Staff member
fyi, new hanzi neural hand writing recognition:
~100 hanzi tried => 1 infrequent hanzi failed = 隺 (modern use almost exclusively as component). in non-neural hwr 隺 hanzi is recognized. thats all i know.
The data set is smaller, yes - my tentative thinking is that we might eventually pull in results from our old handwriting recognizer as well and merge them (at least for people who own that or buy it).
 
I think I've mentioned this before but it is still triggering in the latest update: I have created a dictionary group for Outlier (T+S) - but every time there is a dictionary update the group gets deleted.
 
Top