2.2.13 Bug Report / Feedback Thread
- Thread starter mikelove
- Start date
Maximilian
Member
Hello Mike,
I have IOS6 and the latest release of pleco running. I encountered a problem with adding a new flashcard. All works fine, except that I can not insert a new line via the return key. Even if a new line is displayed at the creation, later it will be all in a single line :-( Any idea what could be the reason?
Regards
Maximilian
I have IOS6 and the latest release of pleco running. I encountered a problem with adding a new flashcard. All works fine, except that I can not insert a new line via the return key. Even if a new line is displayed at the creation, later it will be all in a single line :-( Any idea what could be the reason?
Regards
Maximilian
I noticed this bug a few months ago. I have the popular Tencent 微信／WeChat app and often jump between Pleco and conversations in that app, while using the Pasteboard Reader. This app, like many messenger apps, allows you to record little pieces of audio as part of a chat conversation and to play these back. However, this does something to Pleco's sound in that the Pleco speech merely becomes a whisper, whatever the audio level. So when I try to study Flashcards that use audio, I can barely hear the audio even though the volume is maxed. The workaround I have is I double click the hardware home button to bring up current processes, swipe to the right to bring up IOS's 'Music' icon and quick menu, press play, wait for the music to start playing, pause, click out and reopen Pleco again and Pleco's sound levels are back to what they should be. I notice the problem crops up a few times a week and will try to trace back exactly what I did shortly before but if you already fixed this in the last update, please let me know.
As for the handwriting recognition, it 'seems' to work better. Before I had endless trouble writing the character 断 but now the engine has no problem. Also before I sometimes had trouble with characters that repeat elements such as 品，晶 and 鑫 but now there is no such issue. However, now every time I write 成 it usually comes up with 戌, which means I'm often correcting the score in flashcard sessions because I don't notice it cropping up. Also, my 己 is typically always interpreted as 已 and I think the algorithm could be tweaked a little better for that one as I think it's a bit too exacting on the shape. Overall though, seems better. Does this also extend to the OCR too as that appears to have been getting better and better, understanding more and more fonts?
Someone else mentioned in another thread that sound levels can't be changed until audio sounds in the app. Yep, I've always had that issue too. Initially, the volume controls just changed the master phone audio. No idea if this has already been fixed though.
Last thing and not really a bug but more of how Pleco manages crash recovery. I've always been very impressed by how Pleco seems so robust, even after crashes and flashcard sessions can usually be restored. However, I was studying flashcards the other day (character writing practice, filling in the boxes), had already studied a couple of hundred cards and suddenly Pleco crashed. I reopened Pleco and it never got past the loading screen. Reopened it again and the Flashcard session no longer existed, although everything else was stable. I was a little nonplussed as I rely on the 'repeat incorrect/loop/shuffle' settings in my learning. Is this because there was a double crash and Pleco perhaps sacrificed the session in case this was causing the crash? I guess I could have used the Search Flashcards to find the last incorrectly answered ones and put them in a temporary category for a one-off review, but it didn't occur to me at the time.
As for the handwriting recognition, it 'seems' to work better. Before I had endless trouble writing the character 断 but now the engine has no problem. Also before I sometimes had trouble with characters that repeat elements such as 品，晶 and 鑫 but now there is no such issue. However, now every time I write 成 it usually comes up with 戌, which means I'm often correcting the score in flashcard sessions because I don't notice it cropping up. Also, my 己 is typically always interpreted as 已 and I think the algorithm could be tweaked a little better for that one as I think it's a bit too exacting on the shape. Overall though, seems better. Does this also extend to the OCR too as that appears to have been getting better and better, understanding more and more fonts?
Someone else mentioned in another thread that sound levels can't be changed until audio sounds in the app. Yep, I've always had that issue too. Initially, the volume controls just changed the master phone audio. No idea if this has already been fixed though.
Last thing and not really a bug but more of how Pleco manages crash recovery. I've always been very impressed by how Pleco seems so robust, even after crashes and flashcard sessions can usually be restored. However, I was studying flashcards the other day (character writing practice, filling in the boxes), had already studied a couple of hundred cards and suddenly Pleco crashed. I reopened Pleco and it never got past the loading screen. Reopened it again and the Flashcard session no longer existed, although everything else was stable. I was a little nonplussed as I rely on the 'repeat incorrect/loop/shuffle' settings in my learning. Is this because there was a double crash and Pleco perhaps sacrificed the session in case this was causing the crash? I guess I could have used the Search Flashcards to find the last incorrectly answered ones and put them in a temporary category for a one-off review, but it didn't occur to me at the time.
caesartg said:
We haven't fixed it yet, no, though we've had a few other reports of it. As far as I know there's no way for us to actually adjust the system volume in-app, though, so there may not be a whole lot we can actually do about this unfortunately.
caesartg said:
Interesting, I'll pass those both along - 成/戌 seems like it might be a direction problem, is the "tail" of the third stroke in your 成 curving back into itself like it's supposed to? 己/已/巳 is the handwriting recognizer designer's worst nightmare - honestly I'd probably just separate the third stroke from the second completely in 己.
OCR hasn't had any recognition algorithm improvements lately, but we have been tweaking the image processing code a bit (and are continuing to do so) so you might find some minor benefits from that.
caesartg said:
Actually that probably just meant that it hadn't completely exited when it showed you that second loading screen - I see that all the time on development builds, crashes and then you immediately reopen it and it hangs because the system hasn't actually finished cleaning up after the last crash yet. We don't store in-progress flash sessions to memory until you exit the app or the session - would slow things down and drain a lot of battery life if we did - so it went away simply because it wasn't saved anywhere after the crash.
Thanks Mike - that all makes sense. The 成 problem seems to come from my occasionally ugly handwriting when the 5th stroke is a little long and just touches the edge of the third stroke - the writing still looks much more like 成 than 戌 though and if writing in a hurry or cursively it would be best to pick it out as 成. If I take more care in writing the character without overreaching the 5th stroke it seems to identify it fine though, so no worries.
caesartg said:
That's a new one - did the system put you on our HK download server correctly or did it use another one? I just ran a check and the HK server seems to be accessible from test servers beyond the Great Firewall, but it's possible there may have been something wrong with our geolocation system that had you pointed to another server instead.
Ah okay, but actually, the system couldn't connect to the server at all. Then I tried with VPN and it could connect (and there were no files to download). Then I switched off VPN and again it couldn't connect. Anyway, if I notice it again I'll mention it. By the way, the reason why I've been checking the Add-ons section is to see if new dictionaries are appearing. Will the new dictionaries be appearing as soon as Apple accept them, or will you only make them available with the new major version release? Cheers
Ben
Ben
Maximilian
Member
Hello Mike,
I always get this behaviour when I try to edit or change a existing flashcard or create a new flashcard. When I perform the following steps, up to now, I always get this behaviour:
Flashcards - Organize - Organize Cards - Select Category - Select Card - Edit card text - (Edit) Definition - New line displayed - Back to Reader - Text in Definition still correctly displayed - Save - Text without new line/return!
Regards
Maximilian
I always get this behaviour when I try to edit or change a existing flashcard or create a new flashcard. When I perform the following steps, up to now, I always get this behaviour:
Flashcards - Organize - Organize Cards - Select Category - Select Card - Edit card text - (Edit) Definition - New line displayed - Back to Reader - Text in Definition still correctly displayed - Save - Text without new line/return!
Regards
Maximilian
mikelove said:
Maximilian
Member
Hello Mike,
anything new about this issue?
Regards
Max
anything new about this issue?
Regards
Max
I just noticed some odd behavior using the most recent version of Pleco on the iPad (a 3rd generation iPad). When I have a multiple character word open in the dictionary window, showing the definition, and I try to advance through the individual characters using the arrow keys (both type of arrow key: the "single" and "compound" advance arrow keys), the arrow sticks glowing white and the selection in the window doesn't advance. Everything just freezes. The program doesn't freeze however. If I tap the screen it clears the problem. I tried the same thing on my iPhone and didn't notice a problem, so perhaps this is specific to the iPad. Has anyone else noticed this behavior? (by the way, the word was was looking at when I noticed this is jue2wu2jin3you3). But I've reproduced the problem with other words as well.
dcarpent said:
It sounds to me like there must be something that's causing the document reader to think it can't advance farther, a character it doesn't know how to deal with... are you using traditional or simplified characters? Have you changed any of the color scheme settings, or configured headwords to display with characters in a smaller font in the alternate character set? Which dictionary is this in?
I'm pretty sure it was the Oxford. I haven't made any of the changes that you mention. Playing around with it this morning I am unable to reproduce what I saw last night. I opened the dictionary page for the same word and moved back and forth with the arrow keys within the word and it worked fine. If I encounter similar behavior again I'll post again. I know you can't do much about a problem if it is not reproducible.