<Stephen>

I am currently running a test on the old <date bug> we investigated last year- I want to test my suspicion that the blank date fields may no longer be updating 2-3 hours after publication.

I want also to remind you that I do not consider "leaving the date bug as it is" to be an acceptable state of affairs. Even if my tests conclude that the blank fields are indeed self-correcting after around three hours, I expect you guys to repair the code on this function so that that it works properly in the first place.

Like it did before for 20 years.

============

Details on the test I am running at present:

<Stephen>

Just passed three hours on the <date bug test> and the editing date fields are still blank.

This is the test game:
A Hodges vs Daniels, 1886

========

<Test details> in this post:

Rest assured they will be examined and deleted in due course.

<Stephen>

Sadly, it is as I expected. The <date bug test> has now passed 14 hours and counting. There has been no "self-repair" like we had in days of old, and up to very recent times after that. The self-repair seems not to be working anymore.

Start date of test and other details here:

jessicafischerqueen chessforum (kibitz #37268)

=========

<Stephen> I expect you or someone else to do something about an ACTUAL FIX for this problem. I had stopped uploading tournament games a month or so ago because the "fix" seemed not to be working- I had to revisit a large number of games post-upload and manually repair the date fields for each and every game.

I cannot do any more chess history pgn game upload work until you or someone who knows what they are doing solves this problem.

Again, this is not a new problem. Well over a year ago, when you discovered that the problem "repairs itself" after around three hours, you mentioned in an email to me that a permanent fix would have to be found.

<MissScarlett>

<no trousers>

1) Not supplying full analysis.

In this game snippet:

<7.g4 Nbd7

{better is 7...a6 8.Bd3 h6 9.Nge2 Nh7 10.Ng3 Re8 11.Qf3 Bf6 ⩲ +0.57>

After blacks 7...Nbd7, the computer suggests a better line starting with a6.

Typically you can go to the previous move
(Whites move 7.g4), click on <engine> and get the full line. In this case it says:

<Sorry, no analysis for this position.>

Obviously, it can't recommend a better line,
and not have analysis for the position.

This is how it should work:

<11.Qd2 Na6
better is 11...♘c7 12.♕e2 ♗d7 13.♖g1 ♘ba6 14.h4 ♘b4 15.♗b1 ♘ba6 = +0.48 (35 ply)>

Here you go to whites move 11.Qd2, click on <engine> and get the full line of analysis:

<1) +0.48 (35 ply) 11...Nc7 12.Qe2 Bd7 13.Rg1 Nba6 14.h4 Nb4 15.Bb1 Nba6 16.a3 Ne8 17.Bd3 Nac7 18.O-O-O a6 19.a4 Nf6 20.Nd2 g6 21.f3 h6 22.h5 g5 23.b3 Rfb8 24.Kb2 b5 25.axb5 axb5 26.cxb5 Qe8 27.Ra1 Rxa1 28.Rxa1 Nxb5 29.Bxb5 Bxb5 30.Nxb5 Rxb5>

2) Not supplying any analysis at the end of the game.

I'm looking at openings, so I typically only look at 15 moves of a game. The computer will suggest what should happen at move 16.

Here we get analysis:

<15.Nf1 h5 16.♕d2 f6 17.♗e2 ♗d7 18.♘g3 ♖f7 19.f4 exf4 20.♗xf4 ♘g7 ± +1.83 (31 ply)>

...and can access it with <engine>:

<1) +1.83 (31 ply) 16.Qd2 f6 17.Be2 Bd7 18.Ng3 Rf7 19.f4 exf4 20.Bxf4 Ng7 21.O-O-O fxg5 22.Bxg5 Bf6 23.Bd3 Bxg5 24.hxg5 Rf3 25.Nge2 Rxd3 26.Qxd3 Qxg5+ 27.Kb1 Nf6 28.Rdf1 Kg8 29.b3 Qe5 30.Qg3 Qxg3 31.Nxg3 Rf8 32.Kb2 Ng4 33.Rxf8+ Kxf8 34.Nge2 Ne5 35.Nf4 Ke7>

Here we get none:

<15.Qxh5 Qe8 *>

Here we get none:

<15.Nf3 Bd7 *>

If you try to access these with <engine> you get: <Sorry, no analysis for this position.>

3) Gibberish analysis.

<12.Be3 {? = 0.00 (35 ply)>

Here it gives 12.Be3 a question mark, but instead of supplying what should be played it gives: <= 0.00>.

The same for b5 here:
<14.f4 b5 {? = 0.00 (29 ply)>

4) Large variations in search depth.

Move 10, 33 ply:
<1) +1.38 (33 ply) 10.Nge2 Ng8 11.Ng3 g6 12.Kf1 h6 13.Kg2 b6 14.Nce2 Re8 15.Bd2 Nf8 16.Rg1 g5 17.Bc3 Ng6 18.Kh1 Nf6 19.Nf5 Bxf5 20.exf5 Nh4 21.Ng3 Rb8 22.Kh2 Qd7 23.Be2 Rf8 24.b3 Kg7 25.Rb1 Rh8 26.Bb2 h5 27.gxh5 Nxf5>

Move 12, (same game) 12 ply:

<1) +1.33 (12 ply) 11.Ne2 h6 12.Ng3 Nh7 13.Nf5 Ndf6 14.O-O Bxf5 15.gxf5 Re8 16.b3 Nh5 17.Kg2>

Move 11, 35 ply:

<1) +1.41 (35 ply) 11.Nd2 Na6 12.Nf1 Bd7 13.Ne3 Qh4 14.Qe2 Nb4 15.Bb1 Na6 16.Bc2 Rb8 17.Ba4 Bxa4 18.Nxa4 Qd8 19.Nc3 Qd7 20.O-O Nec7 21.a4 Nb4 22.Kg2 g6 23.f4 exf4 24.Rxf4 Qe7 25.Raf1 Ne8 26.R4f3 Qg5 27.Kh1 a6 28.Rf4 f6 29.Ng2 Qe5 30.h4 Qd4 31.Ne3 Nd3>

Move 15, (same game) 9 ply:

<15.Nf3 ⩲ +1.08 (9 ply)>

Re Gameype:
There seems to be a script that automatically sets a gametype to the games without any type assigned( around 4:00AM UTC), I haven't familiarized myself with it yet.

<Stephen> I appreciate that, thank you.

One more thing though- the <upper case-lower case bug>.

This bug predates the switch to <https>, but since we are working on the upload fields it would be a good time to address this also.

Here's how it works- one or more of the characters in a field cannot be altered through the editing box.

Example is A Hodges vs Daniels, 1886

On this game, it is impossible to change the lower case "o" in "offhand game" on the published pgn. Check the editing box now and you will see that I entered an upper case "O". It is impossible to get that upper case "O" into the published header on the actual game page.

Who's the one running their mouth?!

The way we say round these parts is

<All hat, no cattle>

<zanzibar>