Winboard Forum

Hello all,

I am having a problem with Winboard.



Oh the horror....

In brief: I inherited a Winboard ini with over two hundred entries ages ago (loaded with some engines I wished I had) and I have been adding betas, various commercial engines, tons of UCI engines (thanks to Polyglot) and a couple of private ones over the decades.

I built up an ini file with well over three hundred entries and then things started going wrong.

When I invoke Winboard the ini file gets corrupted - entries are chopped off, pieces of the line get moved to somewhere else etc. resulting in engines not being found. This gets particularly annoying when running an automated tournament.

I spoke with a couple of other experts - sorry Guenther you were on my list next! - who were able to duplicate the problem which seems to manifest itself when an ini file has over 300 entries.

The solution seems to suggest itself (reduce the ini file to less than 250 entries or so) BUT it seems bad form to me to be limited to what the (expletive deleted) technology gives to you.

The problem is not with the Winboard version used - Scotti's version (my all time favourite), the original 4.27 and 4.26 all do the same thing. It seems to be a limitation in the Winboard code somewhere...

Then again, I coud be talking nonsense, a not impossible feat, considering that I am a technophobe and this is computer chess programming - three words I know very little (next to nothing actually) about.

I am spoilt.

I want all my 300 + engines in my ini waiting to be summoned by their master. I want Shredder version 3 and versions 6, 7 and 8 of Shredder UCI up and running in Winboard, I want.....

You get the ugly picture I assume?




Can anyone please help?


Later.

Hi Roger,

Yes I know this problem, I had it too a few years ago
IIRC I had at that time each version for each program in
my WB.ini and at one point I noticed that somehow in the
middle of my first and second entries some characters
were always missing or misplaced!

Of course I also came to the conclusion that I seemed to
have hit some mysterious limit in the ini file.
I still don't know if it is a fault of WB or of the editor I had
used for it. AFAIK since I changed the editor to an advanced one
the problem was gone, but my memory could be wrong.

I can only say that I had reduced my ini file later anyway for
having only the newest version in, as I gave up to the sheer
crowd, which I wasn't able to test anymore
It might be WIN98 related too BTW, since I never had the problem
anymore under XP and WIN2K?
Anyway my current ini file has totally 848 lines says UltraEdit and
2 * 290 are engine entries and no problems currently.

Sorry that I couldn't be of more help.

Best wishes,
Guenther

Hi Roger,
I've checked the source file and at a first glance it seems it could be a buffer overrun problem. There is an easy way to tell. I have increased the buffer by a good amount and recompiled the program:

http://www.ascotti.org/programming/chess/zip/winboard_roger.zip

Can you see if this solves your problem? If not, I need to know the size of your INI file to start looking in the proper places. (If you can just measure the size of the "firstchessprogram" part and it exceeds 20000 characters then it's probably the above overflow.)

Alessandro Scotti wrote:Hi Roger,
I've checked the source file and at a first glance it seems it could be a buffer overrun problem. There is an easy way to tell. I have increased the buffer by a good amount and recompiled the program:

http://www.ascotti.org/programming/chess/zip/winboard_roger.zip

Can you see if this solves your problem? If not, I need to know the size of your INI file to start looking in the proper places. (If you can just measure the size of the "firstchessprogram" part and it exceeds 20000 characters then it's probably the above overflow.)

Hello Alessandro,

Fantastic service!

I will check this when I get home. Perhaps Guenther could have a look at it in the meantime....

Whether it works or not, your willingness to compile something this quickly is commendable.

Later.

(Alessandro:)

Hi Roger,
I've checked the source file and at a first glance it seems it could be a buffer overrun problem. There is an easy way to tell. I have increased the buffer by a good amount and recompiled the program:

http://www.ascotti.org/programming/ches ... _roger.zip

Can you see if this solves your problem? If not, I need to know the size of your INI file to start looking in the proper places. (If you can just measure the size of the "firstchessprogram" part and it exceeds 20000 characters then it's probably the above overflow.)

(Roger:)

Hello Alessandro,

Fantastic service!

I will check this when I get home. Perhaps Guenther could have a look at it in the meantime....

Whether it works or not, your willingness to compile something this quickly is commendable.

Later.

Hi friends,

I have checked the recompiled WB version with a ini file doubled
from my original one and I can say that the problem after 20.000
chars is solved now
In my test ini file there was always a corruption after a bit more
than 20.000 chars (incl. spaces) in a section part of the engine entries.
I see though that the corruption now takes place after around 30.000
chars.
May be with a 40.000 limit even the nerdest WB freaks will be happy
for a while

Thanks for your always appreciated efforts Alessandro!
(Soon I will start Kiwi 0.6a-beta in my new test environment)

Cheers,
Guenther

Guenther Simon wrote:In my test ini file there was always a corruption after a bit more
than 20.000 chars (incl. spaces) in a section part of the engine entries.
I see though that the corruption now takes place after around 30.000
chars.
May be with a 40.000 limit even the nerdest WB freaks will be happy
for a while

...

(Soon I will start Kiwi 0.6a-beta in my new test environment)

Yup, 30000 is the new limit I set, but no problem at all to recompile with 40000!
I will put Kiwi 0.6 final on the site in a few hours, I hope it makes it for the tournament, otherwise 0.6 beta will do just fine!

Alessandro Scotti wrote:

Guenther Simon wrote:In my test ini file there was always a corruption after a bit more
than 20.000 chars (incl. spaces) in a section part of the engine entries.
I see though that the corruption now takes place after around 30.000
chars.
May be with a 40.000 limit even the nerdest WB freaks will be happy
for a while

...

(Soon I will start Kiwi 0.6a-beta in my new test environment)

Yup, 30000 is the new limit I set, but no problem at all to recompile with 40000!
I will put Kiwi 0.6 final on the site in a few hours, I hope it makes it for the tournament, otherwise 0.6 beta will do just fine!

Hello Alessandro,

I still seem to have a problem. The statistics for the engine part of my ini file is 26052 characters. The details are below. I am still having the corruption right about the R part of the ini file (Ruffian).

Perhaps I need a really massive buffer limit?

Thanks for your efforts...



Awaiting the 60,000 limit which will make this nerdy Winboard freak happy.


Later.



Different words/items counted: 651
Total Words: 3243
Total Punctuation: 1043
Total Other Text: 775
Total Characters: 26052
Total Paragraphs: 344

Hello Alessandro,

I still seem to have a problem. The statistics for the engine part of my ini file is 26052 characters. The details are below. I am still having the corruption right about the R part of the ini file (Ruffian).

Perhaps I need a really massive buffer limit?

Thanks for your efforts...

Very Happy

Awaiting the 60,000 limit which will make this nerdy Winboard freak happy.


Later.



Different words/items counted: 651
Total Words: 3243
Total Punctuation: 1043
Total Other Text: 775
Total Characters: 26052
Total Paragraphs: 344

Hi Roger,

Perhaps your tool doesn't count the spaces between the characters?
With them you might be already over the new 30.000 limit?
If you want you can send me your file and I do a count in my tool.

Best regards,
Guenther

P.S. =>60.000 wow

Hi Roger,

You can shorten the ini file yourself a bit by using the old dos format.

Add ~1 after six characters if the name is longer than 8 letters

Example:

Code: Select all

c:\program files\winboard

becomes

Code: Select all

c:\progra~1\winboa~1\

regards,
Jim.

Hi all,

a stat() on the file to get the file size, followed by a malloc() and later on a free() will remove all these funny buffer size limits from the program, assuming that the buffer will be used for reading from file only. (I did not look into the code myself, sorry.)

Then Roger might even want to create a 4 GB WinBoard.ini file if he likes (and his OS file system allows it).

Sven

Hello all,

Alessandro compiled an executable that once again allows me to enjoy all my various betas, versions, WB and UCI engines all at once without any corruption of the Winboard ini file.

Thanks much kind sir.

Now if I could only get some more hours in a day so I could play against them all....



Later.

Hi Roger

I knew you were an engine freak, but this goes beyond what I could have imagined. I wonder if "chess engine addiction" is a recognized medical condition.

Best
Dan H.

Dan Honeycutt wrote:Hi Roger

I knew you were an engine freak, but this goes beyond what I could have imagined. I wonder if "chess engine addiction" is a recognized medical condition.

Best
Dan H.

Hello Dan,


What's that you say, a recognised mental condition? Without a shadow of a doubt. Just ask persons like Heinz van Kempen, Dann Corbit, Volker Pittlik and Roger Brown if you do not believe me. Especially Dann Corbit. he collects everything.

Yes, I engaged in gratuitous name calling but I am not going down alone on this one!



I know it is so wrong but I cannot help myself. I just have to have them. Can you suggest how I may be helped?


I am not even going to mention my non-Winboard engine collection.


Sigh.


Later.

I collect the pro engines.

I collect the amateur engines.

I guess that this hobby is a bit less addictive than crack cocaine, and also not quite as expensive. Or do I have it reversed?

Here is an actual SAT question:
"A moth is to a flame as:"
A. Dann Corbit is to a chess engine with source code
B. Dann Corbit is to a chess engine with source code
C. Dann Corbit is to a chess engine with source code
D. Dann Corbit is to a chess engine with source code

They should all be so easy.

Roger Brown wrote:I know it is so wrong but I cannot help myself. I just have to have them. Can you suggest how I may be helped?

Sorry, Roger. I searched trying to find something like "Chess Engines Anonymous", but I struck out

Best
Dan H.

Dann Corbit wrote:I collect the pro engines.

I collect the amateur engines.

I guess that this hobby is a bit less addictive than crack cocaine, and also not quite as expensive. Or do I have it reversed?

Here is an actual SAT question:
"A moth is to a flame as:"
A. Dann Corbit is to a chess engine with source code
B. Dann Corbit is to a chess engine with source code
C. Dann Corbit is to a chess engine with source code
D. Dann Corbit is to a chess engine with source code

They should all be so easy.

All of the above.

Best
Dan H.