Beta 7 Bug Report Thread

mikelove

皇帝
Staff member
stephanhodges - I'm not quite clear on this import bug; so 乱 mapped to Guifan instead of ABC? Were there traditional characters / Pinyin on the input line? Are you sure the remapping worked (i.e., checked back on the card and it actually reflected the ABC definition)? And you definitely should be able to export custom flashcard definitions - was the checkbox for that enabled in the Export screen?

With the keyboard tab switch, was that the first time you shifted the keyboard during that session? Good point on bounce-checking for hardware buttons, no reason we can't apply the same check to them that we do for onscreen ones.

gandq - it looks like the audio/stroke order features are included in your keyfile, so it probably is something on the installation end. Make sure you installed both sets of audio files (basic along with extended), the extended audio won't work by itself.

Ajo - both of those errors are actually present in the original Oxford dictionary too; it doesn't include a traditional translation for the 們 in 我們 nor does it include one for the 險 in 風險, both appear in simplified-only in the printed dictionary. Since the Oxford unlike our other dictionaries actually does include traditional characters in example sentences / definitions, we wanted to stick as accurately as possible to their traditional character conversions rather than re-converting the entries ourselves. I'll make a note that we should re-check them (no reason we should refuse to change something when it's clearly wrong), but it probably won't happen until after 2.0 is released since the current database is at least faithful to the printed version.

The Duelist - thanks, but we're still having no luck at reproducing this. I'm starting to think it might be a hardware issue - if you use PlecoMover to move your flashcard database to an SD card (and set the version in internal memory to Backup), does that help any? What if you move your other Pleco database files to an SD card as well? (so there's as little as possible left in your Palm's internal memory)

gaokang - well I guess this is better than not starting up at least :) But we haven't seen this problem on other devices - are you using Input Field Compatibility Mode? Any other preferences changes?

goog1e - interesting; nothing particularly unusual about that character code (unicode 0x801C), so I'm not sure what could be causing this.
 

marsch

举人
Re my previous post.
1. Saving flashcard session: in b7a, it doesn't save anything for a change :). Yes, the files in b7 were 0 bytes, no matter where I copied them. My folder is named "\Programme\Pleco" (German), but my documents folder is "\My Documents" (English). Could that be the problem? A schizophrenic OS? Where is the file supposed to be located now? I created a "\Program Files\Pleco", but to no avail. Yes, both my PDAs are WM 2003 - that's what CE 4.2 means, right?
4. Good program :). That's what I would expect as a behaviour, too.
6. Doing a session with headword displayed, then iTask, then tap back on the session. Hanzi disappears, top info line and buttons are still visible - so I can step back and return to the current flashcard and it displays normal again.
7. Thanks. Very tricky :D
8. That was Adso, not HanDe. Haven't checked for others. But found out today that "a lot" of words in Adso don't have pinyin - is that intentional? E.g. yi4guo2ta1xiang1 - 异国他乡, as you can't find it via pinyin obviously.

Re soundfiles. I'm not sure whether the difficulty I have in hearing some of the sounds are due to my poor earphones (tried all 2 sets I had available), to the hissing or are actually wrong recordings. It's sometimes hard to tell apart t/k, b/d/g, qu/xu, c/k. Should I just post all of my examples and then you blame it on my bad ears or is the some way to check beforehand?

I'm not sure where to ask this question, this may be the wrong forum, but anyway: I don't quite understand how to build up a reasonable "difficulty", i.e. get it above say 150 - at least not with repetition spacing. If I start with score 100, difficulty 100, even if I use lowest aggression level, by the time I reach difficulty 152, the score is 363552, so I'll be asked again in 10 years - I hope to have finished my language study by then. So why is there an upper limit of 670 for the difficulty - this is not just an arbitrary number, is it?
 
Minor/cosmetic bug:

Small square in main screen is not refreshed. Small square is on the right side, just under the second line of the tool bar and shows something that looks like a chinese character one. In order to refresh that part you need to invoke the system keyboard or display a long definition (the scroll bar refreshes that part of the screen)

To reproduce: Go to OS home page and then back to Pleco main page

PalmTX with list palette and input palette in the bottom of screen
 

radioman

状元
Am I missing something or is there no way to create a flashcard manually? that has like 8 hanzi characters in it. I can make the string of characters in Pleco, but just do not see how to make the flashcard itself. Version 1 had a "new flashcard" button which I am just not seeing.
 

mikelove

皇帝
Staff member
marsch - OK, we'll try reproducing #1 with WM03. The file's supposed to be in \\Program Files\Pleco, but it definitely queries the system for the localized name of the Program Files folder, so it should be putting that file in \\Programme (and I guess it was before, though with a 0 byte file). A lot of entries in Adso are missing Pinyin, yes - not something on our end but just a limitation in the original database. (CEDICT is a better bet on that front, though it only has about half as many words)

The unclear audio may partly be a function of the aggressive data compression we use to get 68,000+ recordings to fit into a manageable amount of disk space - the female samples have some issues in the original recordings too, at some point we'll go through and do a proper noise removal filter on all of them but there isn't time to do that for 2.0. The difficulty cap is just to avoid math errors, it's highly unlikely anyone will ever get close to it but just in case they do we set that limit to make sure there isn't an overflow error in calculating your new card score.

estudiando - sounds like the radical selection box from the radical palette; not sure why it's redrawing there but it should be easy enough to make sure it stays hidden.

radioman - tap the Add Card button in Manage Flashcards (rightmost in the toolbar below the card list) and that will let you create a new card manually.
 

gaokang

Member
Yes, much better! Resetting preferences to defaults solved the problem. When time allows, I'll change preferences one at a time to determine which one might have caused this bug.

Sony Clie PEG TH-55
32MB, 1GB Memory Stick PRO Duo
Palm OS 5.2.1
PlecoDict 1.03
PlecoDict 2 Beta 7 Stability-Fix
 
mikelove said:
The Duelist - thanks, but we're still having no luck at reproducing this. I'm starting to think it might be a hardware issue - if you use PlecoMover to move your flashcard database to an SD card (and set the version in internal memory to Backup), does that help any? What if you move your other Pleco database files to an SD card as well? (so there's as little as possible left in your Palm's internal memory).
Hi Mike

I've Moved The Flashcard DB To The SD Card.
Still Have Stablility Issues.
(May Be Marginally Better??? ie 1-2 More Taps Before Crashing :wink: )

At The Moment PD2 Is The Only Application That Is Causing My T5 To Crash.
PD1 Is Still Rock Solid.
If It's A Hardware Issue,
Can You Suggest An App That I Can Use To Check The Condition Of My Device's Memory With?

TIA
The Duelist
 

mikelove

皇帝
Staff member
gaokang - great! Definitely let us know if/when you figure out which preference was the problem.

The Duelist - did you try moving the other databases to the SD card as well? I know it seems odd that there'd be a hardware problem when your other software is working fine, but we've seen a similar situation on Palms here (two identical TXes with identical software loaded on them, one crashes / corrupts database files and one works perfectly even though 1.0 works fine on both), and since several of the issues you're describing don't seem to be getting reported by anyone else (and aren't reproducible here) we have to at least consider that possibility. The fact that Unihan and ABC are both loaded in internal memory (and those are the two main databases involved in Char Info and Components) further suggests that there might be a faulty flash memory chip at work here - if there's a bad cell on that memory chip then it's quite possible 2.0's much-larger-than-1.0 data files could be the first things written to that particular cell. But in that case moving the databases to an SD card would fix the problem.

If it's not a hardware problem, then there has to be some other way that your T5 differs from our T5 - we've loaded the same files on it in the same places with the same firmware, same flashcard database (and even the same size memory card), are you running any other third-party software besides CJKOS? How much free memory do you have in internal memory / internal storage / on the SD card?
 

Ajo

Member
Mike,
Two topics discussed here:

1. Regarding the 們
mikelove said:
Ajo - both of those errors are actually present in the original Oxford dictionary too; it doesn't include a traditional translation for the 們 in 我們 nor does it include one for the 險 in 風險, both appear in simplified-only in the printed dictionary. Since the Oxford unlike our other dictionaries actually does include traditional characters in example sentences / definitions, we wanted to stick as accurately as possible to their traditional character conversions rather than re-converting the entries ourselves. I'll make a note that we should re-check them (no reason we should refuse to change something when it's clearly wrong), but it probably won't happen until after 2.0 is released since the current database is at least faithful to the printed version.
I understand the situation and I understand it will be long. But I hope it will be done. For non professional translator like me (I use the dictionary to speak with the family), Pleco Dict is very useful if I do not need to buy a more expensive dictionary to get accurate information. I checked in Pleco Version 1.0 and the Oxford dictionary give the good traditional characters.

2. a new subject
in Menu-Preferences-Button Actions. If I tap on the button "other button actions..." the Palm TX reboot (Softreset). It happens each time I tried. This is not essential for my way of working with Pleco Dict. But I think you should be aware.

Thank you for your work.
Alexandre
 

ipsi

状元
The 'search dictionaries in order' preference doesn't actually seem to be working. I don't recall seeing mention of this earlier.

Also, I am very grateful for whoever decided that including the Matthews number in the Unihan dataset was a good idea! That had just saved me a huge amount of look-up time. Would be better if I had Matthews in Pleco, but that's just not going to happen... :(

EDIT: On the topic of buttons, I can't seem to get the up/down buttons on my Treo to be recognised during a flashcard session... Codes are 0132 0132 (up) and 0132 0133 (down). It does however get button 0136 013D (centre button). However, setting the calendar button (0205 0205) and the Mail button (0206 0206) both work. They also allow me to mark an answer correct without revealing it. I'm generally pretty unwilling to bind my application shortcuts to other functions in a program, so is there a particular reason why the up/down keys aren't being (can't be?) recognised?
 

radioman

状元
Here is a new one

- a simple period ".", parentheses ")" or "(" and I figure some other punctuation.

If that is the only thing in the search string, then I get an immediate and simple fatal error message.
 
mikelove said:
The Duelist - did you try moving the other databases to the SD card as well? I know it seems odd that there'd be a hardware problem when your other software is working fine, but we've seen a similar situation on Palms here (two identical TXes with identical software loaded on them, one crashes / corrupts database files and one works perfectly even though 1.0 works fine on both), and since several of the issues you're describing don't seem to be getting reported by anyone else (and aren't reproducible here) we have to at least consider that possibility. The fact that Unihan and ABC are both loaded in internal memory (and those are the two main databases involved in Char Info and Components) further suggests that there might be a faulty flash memory chip at work here - if there's a bad cell on that memory chip then it's quite possible 2.0's much-larger-than-1.0 data files could be the first things written to that particular cell. But in that case moving the databases to an SD card would fix the problem.

If it's not a hardware problem, then there has to be some other way that your T5 differs from our T5 - we've loaded the same files on it in the same places with the same firmware, same flashcard database (and even the same size memory card), are you running any other third-party software besides CJKOS? How much free memory do you have in internal memory / internal storage / on the SD card?

Hi Mike

I Finally Found The Source Of My Problem. :)
I'm Also Running The Hanwang HWR App.
I Use Hanwang To Write Memos.

Once I Disabled Hanwang, PD2 Seems To Work Fine. :)
I Guess This Kind Of Makes Sense, But It Would Be Nice To Be Able To Keep Using Hanwang. :wink:

The Duelist
 
Tapcis Palm Beta 7+

Scrolling on the radical selection page is too fast to accurately see if I am missing lines. It doesn't seem to move by "one page" if I click the main scroll area, and tap and hold the button is too fast also. For example, looking up 搭 is under stroke count 9, which has more than a full screen (in landscape mode), unless I work around the bug that prevents full width screens (reported earlier).

In this screen in particular, it would be nice to have some sort of line indicator. Perhaps just a small ">" at the far left for the top row, and then a line number above (where there's plenty of room). I know that's got to be an enhancement...

In the meantime, it would be great if the scrolling was a bit slower. Click and drag on the thumb is also to fast right now.
 

daniel123

榜眼
Beta 7+ fixed two bugs I reported before and works pretty well. Seem to be at least as well stable as Beta 7. But I noticed a new bug:

Canceling a flashcard session without saving and afterwords finishing the Flashcards Module area doesn't lead to the main application menu but leaves the application completely. When I start Pleco again the problem occurs as long until for example I go into Manage Flashcards or New Flashcard Session and leave it.

Just then it is possible to leave the Flashcard Module area without leaving PlecoDict itself.
 

mikelove

皇帝
Staff member
Ajo - well yes, it's certainly something we'd want to fix eventually, and hopefully sooner rather than later. Your problem #2 has already been fixed in the new "Experimental" release, so if you install that it should take care of it.

ipsi - thought we'd gotten search-in-order working a few betas ago, but maybe something was screwed up by the other recent dictionary search changes. Glad the Matthews codes have proven helpful, I think embedding UniHan in general has the potential to provide a lot of those little-thing-that's-a-killer-feature-for-a-small-number-of-people type additions that Pleco's gotten to be known for. Not sure what's happening with the up/down buttons, perhaps the button pad is getting set into one of its confusingly many different input modes - ironically the up/down/left/right buttons are about the only ones that work consistently in flashcard sessions on WM at present.

radioman - not seeing that crash on an E2 here, it just sits there quietly until you change something. Have you made any changes to Pleco's search / results settings from the defaults?

stephanhodges - looks like the copy command isn't converting tone marks into tone numbers in that case; the input field doesn't support tone-marked text. The radical screen seems to scroll normally here, but we'll see if there's any reason why it might be acting up on a Tapwave Zodiac specifically.

The Duelist - great! Very relieved to hear that, I'd been really worried about these T5 bugs... it might be possible to work around this by checking to see if HWPen's installed and, if so, only loading the Hanwang code in Pleco when you actually recognize a character (and unloading it immediately afterwards), so you don't have both systems running at the same time for very long; we'll try that out here and see if it makes a difference.

daniel123 - haven't seen this before, must be a lingering bug with the windowing system. (thought we'd fixed most of those in B7 but apparently not)
 
Minor:

with "search only in current dictionary" popup definition shows the previously tapped word if the word is not found in the current set dictionary for the popup screen.

Example: search word in ABC, tap character for popup definition. Set dictionary in pop-up to TL, tap another character that is not in TL, the pop up shows the previously result. This problem is fixed if "search only in current dictionary" is not selected.
 
Functionality enabled?

I was just reading the feature guide about hyperlinks. Is the feature already in B7? When I tap on the underlined words in ABC nothing happens. if I tap and hold the words get highlighted but still nothing happens.
 
Bug Description:
Inside A Flashcard Session (Multiple-Choice)
Using 'Tap & Hold -> Popup Defn' On A HyperLink Causes A Crash

Platform: Palm T5
Software: PD2b7(a)
 

ipsi

状元
estudiando:

Yep, the features in Beta 7. You'll need to go to Preferences -> Stylus Actions, and set 'Link Tap Action' to 'Copy to Input' or 'Popup Definition'. I don't know if it's supposed to have a default or not.

Also, as far as I'm aware, it only works in the main dictionary screen, not flashcards or the reader.
 
Top