Comments/Ratings for a Single Item
@Fergus
I report another bug: in the table of the GC games I'm supposed to have played: https://www.chessvariants.com/play/pbm/logs.php?userid=timurthelenk&age=0&sort=age&stat=actfin
There are many that I have not played, at least not these recent dates (for example, I don't remember having played at Mosaic Chess), many games which are duplicated, etc.
So, the table is wrong. Maybe I'm not the only one. This may affect other statistic pages that are computed on this site.
Apart from what the records say, I wouldn’t know which games you have or haven’t played. I see no such problem with my account, and no one else has reported this. So, if you want me to look into this, I will need exact details.
Fergus:
Please take a look at this link.
All problematic logs are distinguished by having the + sign inserted in the game name, for example, Kamikaze+Mortal+Shogi instead of the correct Kamikaze Mortal Shogi.
When trying to view those games or print the list of plays, the software warns that the logfile is missing, or something is misspelled.
Dear Fergus. There are several bugs in this page:
It is rather simple. All games dated with 2022 or 2021 or even 2020 have not been played in these last days of 2023. I am quite sure of that. Here the bug seems just they have been moved in time. Strange.
There are also games which are duplicated. Like:
timurthelenk-numerist-2023-327-894
playtester-cvgameroom-2023-303-807
timurthelenk-cvgameroom-2023-259-648
makov333-cvgameroom-2022-244-501
cssdixieland-timurthelenk-2022-235-134
panther-cvgameroom-2021-304-254
catugo-cvgameroom-2021-289-757
The duplication seems to come because of the game title which is given as "A B" and "A+B": ex: Chinese Chess and Chinese+Chess.
Finally I also see this mistake where a same number is attached to 2 different games: timurthelenk-cvgameroom-2022-73-858 is once attached to Metamachy and once to Maasai Chess.
{BUG} Hey why I move in first game I should move in and in other my games all pieces are trasparent???
Temporary bug while updating some code. Technically, the pieces had a height and width of zero, but this should be fixed now.
Thx for answer)
Please provide full URLs to the duplicated logs and their duplicates and others you find problematic.
@Fergus:
I don't understand your answer. You ask URLs. Those below are not enough? Or I don't understand what you don't understand. It seems that there is a problem with the way the recorded games are labelled.
I re-post the list below, they have all the same problem.
timurthelenk-numerist-2023-327-894
playtester-cvgameroom-2023-303-807
timurthelenk-cvgameroom-2023-259-648
makov333-cvgameroom-2022-244-501
cssdixieland-timurthelenk-2022-235-134
panther-cvgameroom-2021-304-254
catugo-cvgameroom-2021-289-757
The duplication seems to come because of the game title which is given as "A B" and "A+B": ex: Chinese Chess and Chinese+Chess.
Finally I also see this mistake where a same number is attached to 2 different games: timurthelenk-cvgameroom-2022-73-858 is once attached to Metamachy and once to Maasai Chess.
This is a huge problem that is even affecting my own games from 2004. Database searches indicate that it affects FinishedGames and GameLogs, but it doesn't affect GameSettings or OngoingGames. Since it didn't affect OngoingGames, it flew under the radar for people who didn't look back at past games. The solution has two parts. One is to clean the database. I will probably have to use a script, since a mass delete might get some rows that shouldn't be deleted. The other is to figure out what caused this and stop it. In urlencoded URLs, the + sign is sometimes used to represent the space. So, it looks like it has been saving urlencoded game names into the database. Since I know which tables in the database have been affected, I know to look at the scripts that write to those tables and to compare them to those that write to the unaffected tables.
It appears that OngoingGames is not being used anymore. I used grep to search for .php files with OngoingGames in them, and none turned up. Searches of OngoingGames for Log names matching %2016% through %2023% turned up nothing. So, I probably stopped using it in 2015, the same year I took over running the site. Around that time, I modified the Logs page to use the Database instead of reading every log in the file system, because reading all the logs was using too many resources. It seems that at that time I discontinued OngoingGames in favor of GameLogs, which contained a record of every log in a single table that could be searched with an SQL query from this page.
Comment #31244 indicates that I created FinishedGames in April, 2015, and Comment #34645 indicates that I created GameLogs in July, 2017. No comments prior to today mention OngoingGames.
The other is to figure out what caused this and stop it.
I identified code where the plus sign got added back in and added a line calling urldecode on $game. So, when you click on a log for a game with a plus sign in it, the game's name should now show up without any plus signs. This has the unfortunate side effect that games cannot have plus signs in their names. Fortunately, it doesn't affect many games, because there are few games of that sort.
One is to clean the database.
I cleaned the FinishedGames and GameLogs tables, and I dropped the OngoingGames table.
@Fergus: thank you, it looks much better.
Oh, it is not a major issue but if that interests you, I see a remaining anomaly: there are 2 couples of games which have the same reference number although being related to different games.
As far as I'm concerned, I don't ask for a correction. I just show this in case it indicates something important for the website.
Thanks again
What you're calling a reference number is just the date and time in the format date("Y-z-B"). They are the same, because the invitations were issued close enough together in time. So, it's no surprise that in each case, the same person issued the two invitations with the same date and time.
Thank you for this explanation. I didn't know it was incremented that way. Thanks
@Fergus:
Paul Rapoport (Numerist) has launched an invitation to play Pemba:
numerist-cvgameroom-2024-8-677
When I try to open it for accepting (or not), I get this error message: Your userid is timurthelenk. This log is private. It may be viewed only by the players. If you are one of the players, please sign in first. You may use the menu for this.
In my experience, this has only happened to me in the past with Paul's invitations. As far as I know, he never got an invitation accepted and he finished by cancel them. Sad. But maybe, there is something wrong that he or we have not understood. If his invitation is open to anybody, how come I am refused to open it?
Could you please have a look to this particular log. Maybe you will see what is not done properly or if there is a bug.
Many thanks
When I try to open it for accepting (or not), I get this error message: Your userid is timurthelenk. This log is private. It may be viewed only by the players. If you are one of the players, please sign in first. You may use the menu for this.
I fixed that problem, but when I clicked on the link, it automatically accepted the invitation for me without giving me the option.
@ Fergus
For some reason the Finished Game Courier logs page is showing wrong results of late today. For example, a game of Pocket Shogi Copper I won vs. wdtr2 in 2023 is showing up as 'Juan Juan has won' in the result shown for it, in the Finished Games page linked from the What's New page(!):
Also from 2023, a game of Pocket Copper Shogi I won vs. Richard Milner is showing up as 'White has won'(!), which is better, at least:
In one new line, I had tested for empty($log) when $log wasn't a global variable. So I changed it to empty($GLOBALS["log"]), and that fixed both of the games you mentioned.
Thanks Fergus, although it appears some other entries near the top of the Finished games page (games I didn't play in myself, I think) have wrong results, still.
If you view these logs, they should self-correct.
Ok, I viewed those logs in question, and, from my perspective at least, they now show correct result on the Finished Games Page that's linked to from the What's New Page.
25 comments displayed
Permalink to the exact comments currently displayed.
The first one is now corrected. The second should be
https://www.chessvariants.com/play/pbm/play.php?game=Zwangkrieg&log=sesquipedalian-cvgameroom-2023-342-707