2.0.2 (final) Bug Report Thread

[sych]

Mi Mike,

I noticed this in the beta stages but I was too busy with exams at the time to post a bug report.

Sometime since the original V2 that you've allowed those of us with CJKOS or similar to input flashcard category names in 汉字 which is great. It works pretty much everywhere... except that in the Card Info screen, when viewing the categories list at the bottom, the 汉字 aren't displayed.

example: I've got a category named like this:

Zonghe Ke/综合下第01课

... if I look at the card info page for a card that's in this category, it appears as "Zonghe Ke/01".

It'd be great if you could fix this!

 

[sych]

Minor issue... in the Import Flashcards screen, "Skip, keep cats" has been renamed to "Merge Cats", but the corresponding Help screen still calls it "Skip, keep cats".

(meow!)

 

[stephanhodges]

I was going through today's reading assignment and building vocabulary words to match the books list (96 in total).

When I go to the manage vocabulary screen and show the new list, I have it set to sort by created + ascending.

However, it does not sort that way. The first word in the book, for example, ends up at #7 in the list. Word #2 in the book is down around #15, etc.

So, I will have to export the list, fix it, delete the Pleco internal copy and reimport. It's important to me and the way the class progresses to be able to subgroup the words by range. (Just to give you my reasoning). I understand this is probably not important to most people.

 

[stephanhodges]

In the reader, if I have tabs in the text, it doesn't seem to calculate the proper width on the display.

(Tapcis, landscape mode) It seems to calculate a wide filling for the tab, but show it as one space.

The result is that lines with tabs are shorter on the display, often wrapping when there's actually room for several more characters.

I'm reading a play, so with formatting to show parts (names) versus dialog, etc. this becomes pretty apparent.

I've emailed the text after being unable to attach as zip or txt file

 

[mikelove]

sych - for technical reasons that would actually be really tricky to fix - basically the problem is that while most of the other text fields in the flashcard system (category headers, popup lists, etc) are drawn using Palm OS' built in text rendering system, that field on Card Info isn't, it's drawn with the system we use to render other text. So the GB or Big5 codes that CJKOS is working with in other places are padded out to 32 bits and become unusable in that place. Only fix would be to try to reimplement that field as a regular Palm OS text field, though I guess that might be possible at least...

Good point on the mixed-up documentation, couple of bugs in there actually. (that's what we get for putting so many features in minor updates)

stephanhodges - that should be fixed once we get around to more comprehensively implementing the category reordering system that just appeared in Organize Flashcards - once you can use that sort order in other places in Pleco you should able to get card lists to show up exactly how you want them.

Strange about the reader, it should actually be ignoring those tabs both in word wrap and in text rendering - I guess there's an argument to be made for making tabs render correctly in the reader at least.

 

[stephanhodges]

Strange about the reader, it should actually be ignoring those tabs both in word wrap and in text rendering

It seems to me that it is counting the tabs as "extra wide" or variable width when calculating how much text will fit on line, but actual rendering is converting tabs to spaces when displaying. Being consistent (always treat as space, etc) would be a little better than now .

 

[ar_allen]

I am running a treo 680 palm phone and just recently purchased version 2.01 bundle.
A few days ago now I saw the upgrade to 2.02 and downloaded and installed the upgrade only zip file as it was smaller insize.

After the install I checked the "About" it stated that after the upgrade I was still version 2.01

So then I downloaded the entire 2.01 deleted the old files except keycode from the SD card installed the entire new Pleco fileset.
On starting Pleco again the "About" still states the version as 2.01

How to tell if I realy have 2.02 or not ?

 

[mikelove]

stephanhodges - yeah, consistency's the main thing that makes this feel like a bug rather than a deliberate omission.

ar_allen - Sounds like you still have 2.0.1 - did you install via an SD card? If so, you may have to copy the application file into internal memory before it'll be updated; if you launch an application from an SD card and Palm OS sees an application with the same name in internal memory, it loads the one in internal memory even if it's older. So use the Palm OS' built-in "Copy" command to copy 2.0.2 from your card to your Palm's internal memory.

 

[sych]

for technical reasons that would actually be really tricky to fix - basically the problem is that while most of the other text fields in the flashcard system (category headers, popup lists, etc) are drawn using Palm OS' built in text rendering system, that field on Card Info isn't, it's drawn with the system we use to render other text.

Out of curiosity, what's the reason for using Pleco's text rendering there? Also out of curiosity, when 2.0 first came out, I didn't think it was possible to use CJKOS to enter category names in Chinese... and now it is. Was this something that was deliberately done at some point, or more of an accident?

 

[ar_allen]

stephanhodges - yeah, consistency's the main thing that makes this feel like a bug rather than a deliberate omission.

ar_allen - Sounds like you still have 2.0.1 - did you install via an SD card? If so, you may have to copy the application file into internal memory before it'll be updated; if you launch an application from an SD card and Palm OS sees an application with the same name in internal memory, it loads the one in internal memory even if it's older. So use the Palm OS' built-in "Copy" command to copy 2.0.2 from your card to your Palm's internal memory.

Yes that works. Thanks

 

[mikelove]

sych - complicated thing having to do with some text encoding hacks we use to quickly generate that category list - not unfixable, though, we just need to sanitize / convert that string to 8-bit characters. However, we'd like to add official support for Chinese-language category names in a future release, at which point we'd have to switch the control right back to what it is now (and at which point any names created the old way would be deleted), so the question is whether or not we're likely to still be supporting Palm OS then - if we are, then it probably makes sense to leave it the way it is.

ar_allen - great!

 

[Cameroon]

Hello!
I found a bug in 2.0.2, when I tap 'Send entry to Reader', and the Reader opens with the entry and its explanation, I press my joystick to the right to see the text's characters' explanation as usual, but instead of moving to the next character, the main character changes twice: for example, we have *阿* character's explanation in reader. When navigating to the right, first *阿* changes to *耀*, then to *some character I wasn't able to represent: 耛's left side on the left and 皮 on the right*, and only after that navigation goes as it should to the next character.

 

[mikelove]

Thanks - this has actually been around for a few releases, we know how to fix it but fixing it has a lot of potential to introduce new bugs so we're being very cautious about it (given that it's more annoying than seriously disruptive) - should get taken care of in a future release, though.

 

[daniel123]

Found a small bug for something that has been reported and partly fixed before...

When I open a formely saved flashcard session and press the "x" on the upper right corner of the "Saved Session" window (instead of yes or no button below) this leads again to a "x" instead of "ok" in the upper right corner of the whole flashcard module window.

This is reproducible for me. Hope we now found the last special case like that.

THX

Daniel

 

[mikelove]

Thanks - that should be pretty straightforward to fix.

 

[radioman]

Small thing - When Drawing Characters, in the autoregognition mode, you can start writing hanzi and then undo the strokes. However, when undoing the strokes, auto recognized characters are not backtracking/undoing as the strokes are being removed.

 

[mikelove]

I'm not sure about that one, actually - you might hastily undo an incorrect stroke but then realize that it actually turned up the right character (particularly with Hanwang's brilliant ability to come up with the right character from a random garble of strokes that only vaguely resembles the character you want). The question is whether that scenario or one where you had the right character, drew an additional stroke, lost it, then hit undo is more likely. Design-wise you can make a case for either, undo's a vaguely-destructive act so it shouldn't necessarily result in a new character.

 

[Stump]

Hi Mike

I had my cell formated so I installed v202
I have a problem with flashcards since I tried to import them. I tried to import the HSK flashcards and it failed after a while, then anytime I try to retry or import a different list I get "Import failed" "Unknown error" right at the beginning of the import action.
Thing is I get the same thing even after uninstalling and reinstalling pleco, since the program seems to keep some kind of memory (last dictionnary query is still there after reinstall)

The undo button on the export window didnt help either

What can I do ?

 

[mikelove]

Any chance you might be running low on memory? That would explain the initial failure and the subsequent refusal to even start another import. You can reset all of that saved info using the buttons in the "Misc" panel of "Preferences," so try the two buttons there (Reset all prefs and Clear recent search) and see if those make any difference.

 

[radioman]

Well, most of my Hanzi is a "vague" - Hanwang might be good, but I am confident my writing would win in a competition to confound Hanwang or the Chinese population at large

But, seriously, if I hit undo, it would be for a reason. The right one could have been presented at the time where I decided to back up a few strokes. However, my experience has been that on the times I have checked, my inclination to go back is well founded as I have screwed up the character at that point and the right one is not there.

It is not a big deal to go back and start over as I can use the practice anyway, and it just doesn't happen that often. Perhaps the bigger thing for me was it just felt unnatural to hit undo and the auto-recognition did not change accordingly (i.e., evaluating the current state of the drawing)

I'm not sure about that one, actually - you might hastily undo an incorrect stroke but then realize that it actually turned up the right character (particularly with Hanwang's brilliant ability to come up with the right character from a random garble of strokes that only vaguely resembles the character you want). The question is whether that scenario or one where you had the right character, drew an additional stroke, lost it, then hit undo is more likely. Design-wise you can make a case for either, undo's a vaguely-destructive act so it shouldn't necessarily result in a new character.