Beta 7 Bug Report Thread

[mikelove]

Post any bug reports on Beta 7 here. This thread is ONLY for bug reports, since we've now closed 2.0 to new feature suggestions - those should be posted in the already-running Feature Suggestions thread. Audio errors please post in the separate Beta 7 Audio Error thread (easier for people to check for duplicated ones that way).

 

[ipsi]

I see fonts 24, 48 and 96, but where's 22? Installed it from Beta 6, and everything's fine . Flashcards work just fine (backed them up just in case though!).

On the topic of that crashing bug I mentioned, it might be gone... Loaded up Pocket Tunes, started/stopped two sessions (same as before), and was able to exit Pleco no problems! Yay . I suspect that will also make my system generally more stable, as I get the feeling this heap-stack corruption thing may have affected other programs as well.

Got a reset when trying to stop a flashcard session with stuff running in the background (pTunes included), but I hadn't reset my device between removing Beta 6 and installing Beta 7, so dunno about that. Also took several goes of "Start Pleco, Start Flashcard, Stop Flashcard, Start Flashcard, Stop Flashcard, exit, repeat" before it happened.

*Sigh* Repeatable. pTunes running, no CJKOS (I think?), go "Start Flash, Stop Flash, Start Flash, Stop Flash, Start Flash, Stop Flash, Exit, Start, Start Flash, Stop Flash..." and at some point in the second set of attempts it gives me a soft reset. Made 10 attempts without exiting and had no problems starting/stopping or exiting. Problem didn't recur on second entrance (5 start/stops)... Nor on the third (another 5). Odd. What I do remember about it is that it took longer than normal to stop. Normally, it will stop the session instantly when I hit the "No" button (for saving it), but when it crashed, it spend a second or two waiting before it did so.

Wait, managed to get it again! Pocket Tunes and CJKOS turned on, did it on the third one. And it spent a couple of seconds waiting around for something before it died. Possibly memory? Either way, there may not be much that can be done about this one. Picking up a Treo Pro as soon as it comes out. I assume it'll be much less likely to occur on a WM device?

For reference, about 1900 flashcards total, with that session having 206. Never answered a question, just hit stop straight away.

Anyway, the problems with the reader are mostly fixed . I can see more than three files! Yeah! Bookmarks do indeed stick around between files, and are no longer trashed if I mess up the encoding. Good to know.

On the Tuttle dictionary, it's quite nice. I think I may be bit beyond it, but it's always nice to have example sentances.

 

[ipsi]

Oh, and I found it a tad confusing having the "Send to Reader" menu item shift from "Edit" to "Modules", but that's neither here nor there, and only because I was used to it being under "Edit".

 

[ldolse]

Launcher works well now, I can definitely see a difference when the application is already running. From a cold launch it looks like the Launcher will give you an IA popup after about 4.5 seconds instead of ~6 seconds to launch the full app.

Very happy to have all the free dicts back now, and really appreciate having stylus actions under flashcards.

Minor issue, but when I installed the new keyfile which included the Tuttle dictionary I got an error message that Pleco couldn't find all the components that I'd purchased, please check the installation. I manually installed, and the registry from Beta 6 wasn't cleared out beforehand. It looks like everything is there, so not sure what it thought was missing.

Also minor, this 2nd bug was from Beta 6; flashcard searching wasn't working well enough for me to figure it out then, B7 search is better. I'm pretty sure it should still apply. I was adding Unihan radicals to the flashcards, and some Unihan entries have multiple pronunciations. The problem is that this looks like multiple Pinyin syllables, but it appears that the multiple choice selection criteria is based off the hanzi, not the Pinyin. So I had a multiple choice test for a single character, and it one of the choices it presented me with was Si1 Mou3, which obviously can't be an answer for a single character. 厶 Si1 Mou3 was the character that the entry was pulled from. Not too big a deal because it's not too common that people will use Unihan for flashcards.

 

[llammamama]

Things are looking great so far!

The installer worked very well, however I don't think the keyfile was installed correctly. After I installed I was able to toggle through all of my dictionaries except for the new one (thanks for the free license!), however it did show up under the manage dictionaries screen. I looked into the program files/pleco folder and saw that the keyfile was 0 bytes for some reason. Everything worked properly after I manually deleted that zero-byte file and replaced it with the actual keyfile. It did give me that error saying it couldn't find everything I had payed for, but everything seems to be in working order.


Also, I noticed one small issue with the new category / subcategory selection behavior.

It seems impossible to select ("check") a top-level category if one of its sub-categories is un-checked. For example, I have a category with two sub-categories that looks like this:

Writing
->Radicals
->Puzzles

I am not able to select the "writing" and "puzzles" category unless I also select "radicals". I like that checking the writing category automatically checks the radicals / puzzles categories, however un-checking the radicals category should not automatically un-check the writing category.

 

[beirne]

Query not found

My standard query test has been to look up "Efrem Zimbalist Jr." Not that I talk about Quinn Martin productions in Chinese that much, but I was impressed that it he is in the 21st Century Dictionary. Starting with Zimbal, though, I end up at zimbabwe in the LDC and can't get to Zimbalist in the 21C. I know that this term isn't very important, but don't know if it could be a sign of problems elsewhere.

Beirne

 

[Sarevok]

I exported my 14000 or so cards to XML (not changing any settings in the exporter), but can't import them to 2.0 - the import always stops at the same number of 1955 cards. I set the import settings as follows:

Missing Entries: Skip
Ambiguous Entries: Prompt
Duplicate Entries: Prompt
Store imported definitions in user dict checkbox checked

Don't know about the previous betas, I played with the flashcards in Beta 5 and 6, but this is the first time I actually used the export/import function (as I finally wanted to start using them seriously)

 

[Sarevok]

Reduced the number of cards in 1.0 to 8035 by deleting some HSK lists (which I didn't use for more than a year, so repetition spacing-wise they are not important and I can reimport them later from the txt files) and switched all import settings to Skip with the same result - the import process stopping at 1955... is this a result of my 1.0 flashcards being corrupted somehow (it's working properly under 1.0) or do they get corrupted by the converter during the export process?

 

[mikelove]

ipsi - we went back to fonts 10/12/18/24 to keep the numbers the same as in 1.0. Glad everything's mostly working but we'll take another look at pTunes and see if we can flush out anything - definitely sounds like an out-of-memory error still. (and yes, this should be much less likely on WM) Any problems with UTF-8 so far?

ldolse - glad the launcher's making a difference. With Unihan since those syllables are always in order I think we'll probably just have it automatically pick out the first one - until we change that, though, I'd recommend against creating cards from Unihan, since the Pinyin in those won't be automatically updated when the dictionary is.

llammamama - strange about that keyfile - are you using Vista or XP? Categories are actually supposed to behave that way, because a parent category always includes all of the cards in its child categories (unless you're doing an Exact Category search in Manage Flashcards) - hence the only way to disable one of those child categories would be to disable the parent too. If you want to do something with just the cards under Writing and not the ones under Radicals/Puzzles, the best bet would probably be to create a new Writing subcategory, do an exact category search for all of the cards in the higher-level Writing category and put them in that subcategory instead.

beirne - interesting; we didn't really change the English-Chinese search system in this beta, but since LDC is much larger than NWP/OXEC/CT it's quite possible it might have flushed out a new bug we didn't see before. I'm seeing this here too anyway so it's definitely reproducible now.

Sarevok - could you e-mail me your old flashcard files and that XML export file? It's possible this could be caused by file corruption in the old files, but there certainly could be a bug in the converter/importer too.

 

[qubus]

the import process stopping at 1955... is this a result of my 1.0 flashcards being corrupted somehow (it's working properly under 1.0) or do they get corrupted by the converter during the export process?

I'm just importing my text-based HSK list with 8000+ entries and it just passed # 3000 and counting. So it seams that the importer can handle them UTF-8 files just fine (no definitions, just the hanzi + pinjin)

(passed 4000 by now..) :roll:

 

[beirne]

beirne - interesting; we didn't really change the English-Chinese search system in this beta, but since LDC is much larger than NWP/OXEC/CT it's quite possible it might have flushed out a new bug we didn't see before. I'm seeing this here too anyway so it's definitely reproducible now.

Mike,

Cool, reproducible bugs are always good. FWIW, it has never worked, but I didn't report it much in earlier betas because the emphasis was on more serious problems. I figured I'd speak up now, though.

Thanks,
Beirne

 

[ldolse]

Very minor issue - Highlight chinese text in Pleco's search box, then put Pleco in the background with the 'x' button. Now launch Pleco with the hardware button. Pleco will launch, but you'll also get a pop-up def for the text that that was previously highlighted.

One other thing - Is the title bar name 'Pleco Engine' going to stay? That name always seemed like a stop gap to solve the fact that the initial betas were called PlecoDict and it conflicted with the 1.0. If I had to vote for a new name I think I'd prefer just 'Pleco' over 'Pleco Engine'... I know you want to steer away from PlecoDict because 2.0 does so much more.

 

[mikelove]

beirne - found/fixed, it looks like some changes we'd made early on for the sake of full-text searches (which are a bit different from regular E-C ones since they need to match exactly) could on rare occasions cause regular E-C searches to land one entry too early, which "zimbalist" happened to set off. (so a fortunate choice of test word)

ldolse - found/fixed the IA bug (fixing them faster than you guys are reporting them at this point , along with the Unihan one from before, and yes, the title bar should / will be renamed to just "Pleco" as "Pleco Engine" was indeed a beta holdover. (though we're keeping that nomenclature in a few places like the About box to distinguish it from the databases / add-ons / etc which we don't own the copyright to - "Pleco" as a product is the sum of all of its Pleco-developed and non-Pleco-developed parts)

Sarevok's importer bug has been fixed too, incidentally - turns out that our flashcard exporter was ignoring part of the XML specification which states that a few specific character codes aren't allowed, and the open-source and extremely-standards-compliant XML parser we use in the importer was rejecting those invalid characters - if anyone else runs into interrupted imports, a workaround for now is to look through the file for any characters with codes < 0x20 other than 0x09/0x0A/0x0D.

 

[ipsi]

Not a bug, as such, more a comment on how full-text search works, as opposed to how I expected it to work.

When searching, I tend to associate the 中／英 signs with the language the dictionary is for, as opposed to the language I'm searching in, so I expected the full-text search to, when I entered #中文 while I had the 英 sign up, to search through the various E -> C dictionaries and list all the entries containing 中文 (if any). Conversely, entering #Locomotion would select all entries in the E -> C dictionaries where the word "Locomotion" appears (if any).

However, it works slightly differently to that, in that when 英 is present, it will only search for English words, and ignore Chinese if present. And will search all dictionaries (E -> C and C ->E/C/G) for occurrences of that word.

This confused me a bit, and I thought full-text search wasn't working, as I wanted to search the NWP dictionary for the word 活动 (well, something else, but I forget exactly), but it wasn't working. Switched to 中, and hey-presto! Life is good.

EDIT: Also, changed shortcut keys! Gah! Ahhh! Have to re-learn everything! I think they're slightly more sensible now, just have to remember that "Y" is audio, not change languages...

 

[stephanhodges]

Looking up 澴 (huan2), it does not list the ABC dictionary as available. But, it is in Wenlin. If I click on the character, for the character info box, it DOES show the definition that's in ABC (matches Wenlin), but in the normal window, only shows the ADSO one and the German one.

 

[ldolse]

Ipsi - Mike and I had an extended conversation on that behavior in the Beta 6 thread (maybe 5 too) if you go back and look. Note that the color of the 中／英 icon inverts when it's in full text mode - that's the indicator that the mode is flipped, the icon just tells you what language your currently searching for across all dicts...

It does seem backward from the 1.0 convention but it does make a bit more sense when you think about it, since you may want to search for a specific string across all the dictionaries regardless of whether they're E-C or C-E.

Try adding the toolbar toggle to switch the mode - I find that a lot easier.

 

[ipsi]

Actually, now that I think about it, I do remember reading that. Must have been one of those posts I read while half-asleep... Will go back and have a look.

 

[stephanhodges]

Crashed (needed reset on back of machine) when I I tried adding a list of classmates from a UTF8 text file, on root of external card.

It processed all 16 cards, and said Done. So, I clicked "Exit". I don't recall if it crashed right away, or if I was able to select the menu first.

Needed to press Reset button on back. Was causing bad hissing noise too.

File sent by email.

I'm pretty sure that it's because I used Chinese characters in the category name :shock: but I was really hoping that would work!

Palm tapwave

***EDIT** Actually I remember now. After pressing EXIT, I immediately pressed the "Home" button on the soft screen, because I wanted to recalibrate the screen touches. That's when it crashed.

 

[stephanhodges]

Palm Tapcis

Soft Reset when I selected options flashcards change buttons.

I had been doing a little bit of flashcard review, and aborting reviews in the middle. First time I aborted, I went to fix a couple cards, changing from custom to standard dictionaries. I cycled through dictionaries and settled one time on ABC, one time on OX, etc. Second time, I stopped to delete 4 duplicates that were in the flashcard list for the category. Category was T\DT02 with about 36 cards.

3rd time, I selected to start flashcard, but didn't even reveal answer for first one, but immediately quit the session again. Then, immediately went to the options menu, selected flashcards, then change buttons. That's when I got a reset message. This is the first time that the soft reset has actually worked, btw. All other times I've had to use the hardware button on the back.

***EDIT ***
After the reset, I started Pleco, and went to Preferences->Flashcards->Change Button and immediately got a Fatal Alert (same as before). It says "Emul68K Common.c Line: 1594, Index out of range

As a sidenote, for Preferences, I have selected: "Change btns in landscape mode" and Cmd bar = 13 and Search Bar = 4 in case that has anything to do with it.

 

[Jaybo2002]

This is my first time using this thread so I hope that this hasn't already been reported.

I just loaded Beta 7 on my AT&T Tilt. I have been using V 1 for about eight months and am familiar with the software. With Beta 7 using full screen character input, after drawing one character on the screen and having it recognized, when I try to input the second character the screen does not respond to the stylus. If I then bring up the on-screen keyboard the part of the screen above the keyboard then becomes receptive for input. (I don't have to actually use the keyboard). I can then remove the keyboard from the screen and input characters normally. I don't generally look up many words at one time, so I'm not sure whether this problem reoccurs, but when I start the program again it does happen.