That was to bring that user back from having empty schedules. Although looking back at the post made me realize I left the last few words of one of the sentences off, my intention was to refer to that damage done to those few hours of logs on the vocab end of the quizzing histories.
Okay, I'm not sure what you mean now. I thought the link was in response to my bug/problem report, so I thought that was about a (possible) solution to that.
I quizzed about 400 terms today (or so it says in the stats), but my schedules, most of which were already down to 0 except for one big one and the kanji, suddenly reverted to having terms to study.
Does your comment mean that my quizzing from today got lost/deleted, so I have to do it over? Or is this some other glitch?
My apologies for the confusion, and for the issue. The main chunk of the quizzing history got about 2 hours (vocabulary only) was lost. I still have the raw quiz logs - and if necessary, I can use those to rebuild the individual user history for that small window, but because of the way that the logs are kept, it would not be 100% accurate (historical data is timestamped with the day it was done on ...nothing on the hour or minute range), and it would most likely take a rather large amount of time to create the code needed to do that.
If it didn't work 100% correctly, it might cause additional damage to the quiz history, so I begrudgingly feel that scratching off those couple of hours is the safest thing to do for the data.
Okay, I guess that was bad timing for me! Thanks for the explanation.
The issue itself is gone now, though? Meaning if I re-do my quizzes from today, it'll stick now?
(BTW, I tried doing some kanji while I was waiting for your reply on this, but my kanji quiz keeps aborting. IDK if this is related; I can also make a separate bug report if that helps for keeping track. But ATM I never seem to get more than 1-3 questions in before the quiz aborts and I get the error message.)
The issue is fully resolved.
In regards to the second issue that you just brought up, I just (a few seconds ago) pushed a fix I believe will take care of it.
Finally. Numbers didn't quite match, but the number of questions was correct. One schedule said 19 left to study, but I ended up with 22. Not complaining since I only had to flush three schedules.
Not sure if this is related to this specific bug, but...
Was almost finished quizzing for the day, finished with my second-to-last quiz on my last kanji schedule (so all other schedules were completely finished), but at the end quiz page, 4 kanji schedules (4027, 3438, 5738, and 5183) all popped up as having 1 term to study. Going to the individual schedule pages cleared the term out, and there is a fair amount of overlap between all the schedules so there is a slight possibility that I could have the same term in all 4 that somehow decided to misbehave, but I'm not sure exactly why that happened!
[[EDIT]] and just tried to do that last quiz (only 8 terms!) and the quiz won't start. Flushing doesn't change anything. Schedule 10022
Alright - so I've added a bit of debugging info to the schedule summary page.
There are actually two ways of measuring the number of terms left for the day: the mastery level charts (the green ones to the right), and the daily 'left to study' study goals.
While they are both processed at the same time, they are processed and stored differently.
Next time you have a mismatch on the numbers, please jump to your schedule's summary page. Just below the 'daily study goals' list, you should see a debug number (basically, that number is a sum of all the 'ready to study' green bars over on the right side). I'm very interested in knowing if that number is inaccurate as well, or if it is *just* (I hope) the study goal numbers.
The debug number seems to be inaccurate as well...
Kanji schedule #9738 - I had finished studying it for the day but it still showed 1 term left (both 'left for today' and debug numbers), I flushed it and both came to 0.
I just had an odd experince. My LtS (Left to Study) count was at 143, but when I added the bar graph numbers up I got 167. That's the first time I've had the LtS [i]less[/i] than what it should have been. I flushed the schedule and it the LtS went up to 167. This was on vocab schedule #16113.
Well then, sounds like the problem isn't specific to once measurement or the other. Will need to turn the verbose logger on. I've already got shvelven's on there - anyone else want to have their quizzing slowed down a bit in the name of fixing the numbers issue?
Vocab schedules 4782 and 5233 both kept reading 1 left to study (debug#0) as did kanji 7368 but were all cleared by flushing.
Vocab 5045 gave the correct number of questions but the rts number didn't match. It said 81 lts, gave me a 50 question quiz, then said 34 lts.
Actually, right now I'm having the summary rebuild at the beginning of every quiz you start, so if you say take 5 quizzes on one schedule and there was an error in any of the first four, you would not see it as long as there is no error in the last schedule. So I still recommend anyone interested in volunteering as it might uncover slip ups even if they are 'corrected' by the time you get to the numbers.
Go ahead. Had three different schedules in need of flushing and one with mismatched numbers today. Said I had 83 lts, 50q quiz, said 32 lts, 31q quiz (5620).
