3.2.x Bug Report / Feedback Thread

[rizen suha]

keep old ios 18 keyboard, ios 26 eats 3% of the width. wtaf.

 

[mikelove]

Not something we have much control over, sadly.

 

[lookslikerain]

Bug report for newest Legacy beta version: When you tap on a word from search, the screen can be dragged back to search from anywhere on the screen, not just the far left. This means that when you try to select multiple characters, a drag starts and you can’t select text. After touch release, when you select the next time the handles are gone (they return on subsequent selections, but it doesn’t matter because they can’t be used while this bug is present anyway).

 

[mikelove]

Thanks! Will fix shortly. (We dealt with this in 4.0 months ago but forgot to deal with that particular iOS 26 issue when we were updating 3.0)

 

[mikelove]

Fix should be up now, let me know if it works OK - thanks again for flagging this.

 

[linhai]

Version: 3.2.80

If no phonetic system is selected under Settings -> Languages + Text -> Phonetic system, Tone Practice tests won't display any text in the input boxes, and selecting any tone option will result in a failed question no matter if the tone is right.

 

[mikelove]

Thanks.

 

[kuduman]

Bug report - I experience an issue with the still OCR on iOS (iOS26, Pleco 3.2.81, with new OCR interface and algorithm enabled): when I try to correct the character that was wrongly OCR‘ed, I hold my finger on it to open the handwriting interface, I write the correct character and select the correct one from the options is displayed, but the character is not changed to the correct one I selected.
Thanks in advance.

 

[mikelove]

Sorry, you have the new interface and algorithm enabled? I didn't know the correction interface was even hooked up there - it doesn't output alternate candidates (not the way the algorithm works). Will investigate whether we can get just the handwriting working now that we know it's still there.

 

[kuduman]

Sorry, you have the new interface and algorithm enabled? I didn't know the correction interface was even hooked up there - it doesn't output alternate candidates (not the way the algorithm works). Will investigate whether we can get just the handwriting working now that we know it's still there.

Thanks. Yes that would be great, as the OCR is never 100% correct (especially in documents of low quality printing). The new algorithm does produce better results.