Beta 6 Bug Report / Feedback Thread

goog1e

举人
Quick flashcard question:

If you switch from Spaced-repetition to Frequency-adjusted or Random, what happens to the score data? Is this not recommended? I want to review some cards more often than recommended in the first day when I learn things.

Thanks
 

marsch

举人
Hi,
thanks for beta 6 – many things work better for me now! However, some still don’t (all Flashcard / Flash card related – so you can google for both, if you like :))

  • “Only change once” still isn’t persistent – whenever I come back to the screen, it’s ticked – haven’t checked for function

    Audio: maybe a warning when system volume is set to 0? I was about to complain that it doesn’t work at all… :? Works great now even for more than 200 cards.

    saving/resuming a session doesn’t work: nothing is saved (in b5, it at least created a 0-byte file). Did I do something wrong when I installed it, as it seems to work for others…?

    quitting a session is not quite what I expected it to be: It asks me twice whether I want to save the session (i.e. nothing happens after the first time), then it says “no flashcards found”, doesn’t review incorrect, let alone loop, nor show statistics. Shouldn’t that have been implemented in b6?

    if I finish a batch of flashcards, I have to confirm the last card twice before I can review incorrect cards.

    Manage FC: I tried to add a cat to all flashcards that had another cat (i.e. they had the cat for “lesson 3”, I wanted it to have cat “read” in addition). Didn’t work. It showed that it was processing, but then nothing changed.

    “keep cat” is a great feature. Just in time for the new semester, so I can add my lessons’ vocab. One “problem”, however: I was used to having an option of displaying the words in the order they were added (good for checking whether you've done them correctly, or for exchanging flashcard files with others). Of course, sorting by creation date doesn’t work if they were already in another cat. So sorting by pleco_flash_categoryassigns.id would be a nice feature (as that’s the same order than my adding flashcards to a cat).

    from the new counter (thanks!) in Manage FC I conclude that you do the batch processing one by one? I don’t know how others use that screen, but I do quite a lot of batch processing (s.a.), so if instead of doing it one by one you could just execute an SQL-statement, that would give me a lot more time to memorize instead of to manage :roll: .

You wrote:
If you do a Google search you can find a lot of information on how to prevent the system from disabling the card slot when it goes to sleep.
Maybe I’m just too stupid to do a proper Google search, but I haven’t found anything relevant yet. Do you happen to have a link?

To Fleminator: One suggestion to emulate a “not in cat” – I have to do that often, too: I create a dummy cat, add that to all cards (or add all cards to that – depending on your perspective :) ), then filter for the cards in “cat” and remove those from “dummy cat”. Then “dummy cat” contains all cards that are not in “cat”. But I’d also vote for a “not in cat”-option!
 

mikelove

皇帝
Staff member
goog1e - the scores carry over between modes, and should work similarly in both; the score in both rep-spaced and freq-adjusted is essentially just specifying how often you see a card, frequency-adjusted it's relative to other cards and rep-spaced it's an absolute number of days but either way a card with a score of 400 is going to come up half as often as a card with a score of 200. (which is why we make frequency-adjusted the default option, since it gives you the optimization of rep-spaced but with the flexibility to review as many or as few cards as you feel like reviewing on a particular day, though you do lose some of the memorization benefits without those strict intervals between showings)

marsch - we've heard of a few preferences still not saving on PPC, actually - not sure why, we haven't managed to reproduce it here yet. Sorry about the volume 0, you're right that we ought to put up an error message for that. We've had one or two other reports of sessions not saving even in B6 - are you sure that that empty file still isn't sitting around / left behind (either in the root directory or in My Documents)? What type of session were you running? The double-prompting is a symptom of a larger bug with review-incorrect which we think we've fixed for B7. Manage Cats' batch category-add function is broken in B6 on PPC (we moved all of those batch processing functions to their own thread so the window wouldn't freeze) but should be fixed in B7. If you set the sort to "none" then the results of any category-only search should show up in the order they were added to the category (simply because that's how they get pulled out of the database) - does that not work? The batch processing actually has to be done one-by-one because there are some extra checks involved in each operation - it shouldn't be significantly slower than a single operation, though, since all of the tables are indexed by card ID and hence it doesn't need to scan through the entire table each time it wants to find a card to change.

I don't have a specific link for card auto-off delay timing, but I think the relevant registry key is named "PnPUnloadDelay" so if you do a search for that you should find a bunch of info.
 

sych

榜眼
I suspect I might have a corrupted FlashCards DB.

I have a Treo 680, I upgraded from Pleco 1.x to Pleco 2 B5, and then recently upgraded to B6 by overwriting the previous PDBs/PRCs using the new ones from the B6 zipfile. I left my user dictionary and flashcards DBs in place. I am using PDB files for my flashcards and custom db.

In B5, I created 2 flashcards profiles - "Reading" and "Writing". Each has a separate score file. Last night I tried to use the "Writing" one for the first time in quite a while, and after switching to the "Writing" profile and doing a session, I decided to have a look at the options in "Advanced" on the flashcards screen. I can't remember exactly what I did (I don't think I actually made any significant changes) but when I was finished, I clicked "start new session" and no cards matched the selection criteria - which was odd, because I'm using repetition spacing and I should have had about 600+ cards to do that day (!!). So I switched back to the "Reading" profile to see if there were any cards there, and I got a message saying "your score file is invalid".

I couldn't work out how to fix it, so I decided to try restoring from my previous HotSync backup. I restored Pleco Flashcards.pdb, PlecoCUser48904B61.pdb, and tmp_XSXSp6NsWLIkb73.pdb back onto my Treo using HotSync. After doing this, I found that when using "Review incorrect cards at end of session", I was getting a crash just before the review started (hadn't got this before), but otherwise everything seemed OK.

Until I found I can't add cards to categories. I can add a card to "uncategorised", but when I go to Manage Cards, select Category: Uncategorised, then use the "plus" icon at the bottom to put the card into a category, it first disappears from the list, but if I go back to the same screen (or hit search), it's still in Uncategorised. If I edit the card, and try to assign one or more categories there, it has no effect. If use the add flashcard toolbar button to add a card to a category it also has no effect - the card ends up in "uncategorised".

I decided I would try to export and re-import my flashcards. I tried exporting an XML file to my card. I selected All cards, Include Categories, Include Statistics. (I don't have any custom definition cards.) Even though I only have about 5000 cards, the during export the counter went up to about 92,000 (what does it count exactly?) and created a 62MB XML file that filled up my card(!). I've just freed a few 100MB on my card and I'm going to try it again, but I'm not sure if it's supposed to be this big!

Is there anything you can suggest that might help me, and if you'd like to try tracking down what caused this, what would you like me to do?

Thanks
 

mikelove

皇帝
Staff member
Sounds like a corrupted database, yes - B5 had the definite potential to do that with PDB-based databases, which is why B6 creates them in flash memory by default. You could try using PlecoMover to see if the database works better if it's read from internal flash memory than from a PDB, but aside from that there's not much I can suggest - if you e-mail along your flashcard database file we can take a look and see how badly damaged / potentially-recoverable it is.
 

kkuguru

秀才
Hi Mike,

I think you forgot to address the second part of my question on an earlier post below:

On another issue, when running a flashcard session and in the case that the definition is longer than 1 page, I can not use the up/down key's to scroll. I went into the preferences, under flashcards and checked the options with the buttons, but there is no option for scroll. Can this be added in or is there another way to scroll in a flashcard session, using the up/down key?

Also, the automatic language detection, while useful at times, was driving me nuts because it kept switching to the wrong language and not switching back, even when I manually clicked back (a problem when the word you are looking for is not in the dictionary, but since it keeps switching, it's hard to know for sure whether the word is not in the dictionary or just an error in the way it's inputed.) Anyway, to solve this problem, I went into the preferences to turn off automatic language detection. But then when I highlight some Chinese characters to put them into the entry field, it does not automatically switch over to Chinese dictionary, which is quite useless (how can Chinese characters, ever be used in the English dictionary). In short, can you change the coding so that when automatic language detection is turned off, it still is smart enough to know that when Chinese charactes are inputed, to use the Chinese dictionaries?

Thx,

K.
 

sych

榜眼
Hi Mike,

I'm currently doing another attempt at XML export so I can open up the XML file and see if the contents are usable, I might try moving it to hidden internal memory after that and see if it makes any difference. I've emailed you a copy of the PDB, if you can have a quick look at it I'd appreciate it.

One of the reasons I haven't switched to using the hidden internal storage by default is that I'm uncomfortable that those files don't get backed up during a hotsync.
 

sych

榜眼
Mike, in export, what does the counter count exactly? It's currently up to 134,000+! Is this unusual?
 

ldolse

状元
Ran into a very minor bug today, and a bit of feedback on the radical input screen.

BUG - If you get a notification like a new SMS or email in the middle of stroke order playback then whatever stroke should have been drawn in during that animation is not drawn in. Instead it finally gets colored in after all the other strokes.

Regarding the radical input screen, I noticed the hardware button just skipped right over the whole character selection box. Would be really nice if it went into the box let you scroll through the results.
 

mikelove

皇帝
Staff member
kkuguru - oops, sorry for missing that. Scroll button support in flashcard definitions is tricky because there's a lot of other stuff buttons can do on that screen, but we'll see if there's an easy way to add scroll text up / scroll text down to the list of assignable button commands. Tapping on the language switch button should change languages even with a word that isn't in the dictionaries, so that sounds like a bug.

There actually are scenarios when it's useful to have auto-detect not even switch on Chinese characters (pasting in input from other programs that needs to be cleaned up, accidentally highlighting the wrong thing in Copy to Input, forgetting your Chinese input system is in Pinyin-to-character instead of alphabet mode, etc), so if we wanted to auto-detect on characters but not Pinyin it would have to be as an additional option and we're really not adding any more options to 2.0 at this point - we can certainly consider it for 2.1 though.

sych - the weird counter in export sounds like a bug, now that we have the file from you we'll see if we can reproduce it here. (F everyone's I we were able to recover most of the data from it, or at least we think we were - file corruption from PDBs seems to mostly affect indexes rather than actual data and those are generally easy to rebuild)

If any tech-savvy people run into database corruption and don't feel like sending your file to us, put it in PQB using PlecoMover, open it in SQLite's command-line sqlite3 utility (available for Linux, Mac, and Windows here), use the .dump command to dump it to a SQL file, then create a new database and use the .read command to read that SQL file back into the new database and you should be all set (though you may get some single-line error messages thanks to the corrupted bits).

ldolse - thanks for finding that notification bug, tricky sort of thing to catch. Getting our custom controls working well with hardware button control selection is kind of tough, but I'll make a note to work on that for 2.1 (if not before). BTW, I'm emailing you a copy of Beta 6 updated with our attempted fix for the audio bug, let me know if you notice any improvement - we found a situation in which the old code could potentially get stuck in an infinite loop (about the only thing that can regularly force a soft reset on Pocket PC, and then only if it's in a higher-than-normal-priority process like audio playback).
 
#2 is probably happening because of the new system we use to cache profile/scorefile names (so they're not as maddeningly slow as they were in Beta 5) - does the list still populate slowly or is it faster now than it was in Beta 5?

sorry, I miss-reported this. It WAS slow in B5...its caching fine in B6

#3b are you using frequency-adjusted or repetition-spaced lately?

Exclusively Repetition-spaced. I actually have had a lock-up or soft-reset a couple times when I hadn't re-entered a saved session. The only common fact seems to be clicking "no" I don't want to resume this session later. Its also possible that a common factor is doing some memory intensive things before I enter the session, but I haven't had any other memory related issues in B6



#9 it sounds like it couldn't access the flashcard file from wherever it was accessing it before. Go into FileZ, Preferences, and make sure the "Show Hidden Volumes" option is checked, then tap the Close button and View and Edit Files again and you should see your hidden internal volume listed as the second item - if you open that up, is there another flashcards file in there?

I already delete so that Pleco 2 could find the actual flashcard file that is now stored on the SD Card (as per instruction from Beta 5

#8 here's something that might work: search for all flashcards with dictionary == your user dict, export them to XML (with user dict definitions included - use the Export button in More Options to export a list of search results), delete the cards, delete your user dictionary file, then import that exported file back into Pleco with definitions being stored in the user dictionary - that should bring back just the flashcard-linked user dict entries and no others.

Hmmmm...I tried this. Every time I tried re-importing my Flashcards from XML, it didn't work. Sometimes it did a soft-reset, sometimes it just exited to the PlecoFlash main screen, sometimes it waited for me to click done. When I did the search again for cards linked to the user dict, it only returned one card. Each time that card was 录下来. I can certainly do without that card, if thats the problem. I also tried importing cards without re-adding the definitions to the USR dictionary...same problem.

I'm doing a restore form back-up right now to start from scratch (of course I synced before I started deleting things!)
....I'm back...TERRIBLE news. Restore from backup FAILED. glancing throught the log, it appears that about half my files were restored, but then everything started failing becuase of "- Unexpected error number 16387". My palm is in shambles. I turn it on and it is obviously back to the factory setting color scheme and there is an error that says "No Font Database! You must have at least one Font installed." When I click OK, it clicks a little and soft resets and comes back to the same error.
OK, I broke down and did a hard reset...there's nothing else that could be done, but I knew what was coming--the forever looping digitizer screen.

back again (2 hours later)...I got past the digitizer screen finally, tried restoring from my backup again. some problem.

Palm is crap, but I'm mac user and I'm just trying to hold out on buying my iPhone until Pleco is released on it.

(more time elapses)
OK, I finally got it restored, and the touchscreen is at least working at about 95% of what it was before (which wasn't that great). Restore from backup seemed to fail because it was trying to restore TONS of files that I deleted from my Palm's internal memory ages ago. I just went through and handpicked the files that I wanted installed (incidentally, leaving out Pleco 1.0.3 for the first time) Things seem a little less stable now than before, and of course I still have all those duplicates, but I'm going to try to avoid another disaster.

I guess maybe my next step is to figure out if that exported USR dictionary card XML file is corrupted for some reason so it won't import...can I use the above listed SQLite utility for that? I'll give it a shot, and if anyone else has recommendations, I'm open.
 

ldolse

状元
That seems to have fixed it! Now that I can finally use it It's a really impressive addition. I can see why you're doing the re-encode - the female voice in particular is much quieter and seems fuzzier than the male voice.

The hardware button stuff is just a nice to have anyway, just highlighting it. It's really nice when it does work though.
 

mikelove

皇帝
Staff member
Fleminator - glad to hear the caching has sped up at least. I suspect the soft reset issue is tied to either the crash-on-repeat-incorrect or the generic reset-on-exit-due-to-heap-corruption issue, both of which should be covered by our fixes for those in Beta 7. (or at least I really hope they are, since we've finally just about finished dispatching with all of the other double-starred super-serious bugs)

Re the later issues - very strange that the memory would get screwed up even with the files stored on an SD card; must be a consequence either of damage done previously by Beta 5 or by the fact that we still store temporary files as PDBs (though we could easily change that, and might want to for Beta 7). Of course if the TX weren't so prone to hard resets anyway this wouldn't be a problem... for what it's worth Windows Mobile sync actually works pretty well with Missing Sync, assuming you have a compatible Pocket PC (which the iPAQ 110/210 unfortunately aren't), though I can certainly understand the desire to have something more explicitly Mac-friendly like an iPhone.

The easiest way to check if that XML file is corrupted is to just open it in an XML viewer like Firefox - any sort of significant corruption would screw it up enough so that it wouldn't validate even as an XML file, let alone as a Pleco flashcard file specifically. The utility I mentioned only works on the actual database files, not on exports/imports, so that's only helpful for fixing up an actual backed-up flashcard / user DB file.

ldolse - glad to hear it! Though unfortunately we probably won't be able to fix the fuzziness / volume mismatch problems until a later release; they're not consistent across files (they were recorded in a bunch of different sessions, and for various reasons we don't have a clear list of which files were recorded together) so we'd have to go through and fix them one-by-one (batch audio converter utilities don't seem to be reliable for this) and obviously with 30,000 audio samples that takes quite a while.
 

jinji

Member
Re comment :

for what it's worth Windows Mobile sync actually works pretty well with Missing Sync, assuming you have a compatible Pocket PC (which the iPAQ 110/210 unfortunately aren't),

I haven't bought one yet but when asking which PDA was best for Pleco iPAQ 110/210 were recommended. So which non- phone model should I buy?
 

goog1e

举人
Flashcard bug?


When using the Spaced Repetition mode, I've been getting a card that has a score of 100% every time for the past few days. It now has a score of 100, and has a record of 38-0.Just yesterday I started getting another card with a score of 100% every single session as well.

I don't know how to give you any more information about this...
 
Not completely on topic, but...

When you release beta 7, could you first post a short "reply" on the "Beta 6 is released" post? We can subscribe to updates on existing threads, which tell me (by email) that there is a new post).

Unfortunately, either you haven't enabled it or it is not possible to subscribe to an entire forum. Personally, I would prefer an RSS subscription to the entire forum. Alternatively, the subscription notice (by email) should include the text of the immediate reply. Don't know if there are any configuration options for that, however.
 

ipsi

状元
Actually, you can subscribe to an entire forum. It doesn't always work terribly well, but it can be done. Link is just below the "New Topic" Button.
 

mikelove

皇帝
Staff member
jinji - a lot of people like the Axim x51/x51v, they don't make them anymore but there are a lot of them on eBay and they seem pretty solid/reliable in general. You don't actually need Missing Sync to install Pleco from a Mac, though - just an SD card reader should be plenty.

goog1e - that's odd... so it's always the same card? Anything at all unusual about it, length or rare characters or missing Pinyin or anything like that? Is this in self-scored or some other mode? Could you send me a copy of your flashcard file by any chance? (along with letting me know which card it is)
 

gato

状元
Flashcard.
- Has the option to sort cards by card ID been added, yet? I can't seem to find it.

Reader.
- Would it be possible to reduce the page scroll by one or two lines? Seeing the last line or two on the previous page makes it easier to orient oneself on a new page.
- It takes quite a while (maybe 10 seconds) to open a 100KB PDB file. Any chance that this time can be reduced?
 
Top