Eelco de Groot wrote:Hello Denis, someone on Rybkaforum suggested a 64-bits compile of the egbbdll.dll would be needed. I don't think we need the egbbso.so for the UCI funtionality? Daniel Shawul I believe has made the sources public. But maybe he would have a better idea. We could put a question in the Winboard Forum about it hoping Daniel finds it there, or maybe you could take a look at the code?
It seems like the best thing to try as a cure for the problems on 64-bits operating systems.
For George and other Vista owners, I just came across a small article that says the final version of Service Pack 2 for Vista is on the way, now still in Beta. http://news.bbc.co.uk/newsbeat/hi/techn ... 764371.stm
Several hundred bugfixes Final version expected early 2009 but not more precise than that. I don't think it will help with this egbbdll.dll kind of incompatibility though.
Eelco
This problem is case to case. and I pedict it's just a compiling issue.
what you mentioned solves the crashing because it doesn't initialize the dll, it's as if there's none. Even my 32-compiles won't initialize EGBBprobe. Only a 64-bit Grapefruit can call the 64-bit egbbdll.dll.
The egbbso.so is probably a Linux counterpart of Windows egbbdll.dll. The 64-bit dll was supposedly made to make use of the bitbases when using 64-bit (egbb enabled) engines like, Sloppy, Scorpio and/or 64-bit Togas.That was the reason why I requested the new sources from Daniel Shawul. CCRL testers noticed that Scorpio64-bit can't access the bitbases.
Master Jim already compiled and included it on his 64-bit Sloppies and Scorpios. I've made mine too in some thread with Ray Charles at CCRL regarding this egbb 64-bit problems of Scorpio.
I've informed Vadim about this and he'll probably resolve this on his next update.
Thanks Eelco.
Denis
Hi Denis-
What is Vadim's full name...?
I'm interested because I'm working on a chronololgy of fruit derivative releases, and i currently am listing grapefruit's author as 'ThinkingALot'.
here's a link to the page in case you're interested:
Written by France's Fabien Letouzey in object-oriented C++
Actually the Fruit code is not object oriented by the common definition of that term (not a single class is used). Furthermore it uses only very few features of C++ and might as well have been written in C.
I tend to think of the Fruit source as high quality C code.
Denis P. Mendoza wrote:CCRL testers noticed that Scorpio64-bit can't access the bitbases.
Hi Denis,
Are you sure?
At least in infinite analysis, using the 64-bit egbb.dll, I found that the bitbases worked very well with Scorpio 2.0 64-bit (look at eval. score)
Written by France's Fabien Letouzey in object-oriented C++
Actually the Fruit code is not object oriented by the common definition of that term (not a single class is used). Furthermore it uses only very few features of C++ and might as well have been written in C.
I tend to think of the Fruit source as high quality C code.
interesting point...
Fabien Letouzey likely considered it C++, - all program files are .cpp, and it was compiled (as a C++ program) with MS Visual C++.
As far as definitions of what is C and what is C++: i'm not sure if it really matters if classes are explicitly defined or not, i.e. actually implemented in the program source code. (perhaps this is a matter of interpretation).
certainly many of Fruit's included header files would have utilized the many built-in Microsoft C++ class libraries since it was compiled with MSVC++...
The two features of C++ that are used I think is the fact that you can have two functions with the same name if they have different signatures and the fact that you can omit "struct" from declarations. The second feature is purely cosmetic and the first one is used very sparingly.
It is trivial to convert Fruit or Toga into a pure C program. For example Togarouter which plays at FICS is a C version of Toga 1.4.1SE.