BNCSutil 1.3.1

Spilled · September 13, 2006, 07:02 PM

Quote from: K on September 13, 2006, 06:50 PM
Quote from: Spilled on September 13, 2006, 05:14 PM
Quote from: K on September 09, 2006, 02:18 PM
I have compiled the latest version with the fixes mentioned here.
You can download it here.

Where can I get the binaries?

From the link "bncsutil.zip" on the page I linked to. I'm hesitant to link directly to files on my server from other webpages.

I'm not sure if you misread but I need the binaries, not the only the dll but the lib and header files.

K · September 13, 2006, 07:15 PM

Quote from: Spilled on September 13, 2006, 07:02 PM
I'm not sure if you misread but I need the binaries, not the only the dll but the lib and header files.

Well you should have said "where can I get the lib and header files" then

Give me a minute, and I'll update the page/download to include both.

Spilled · September 13, 2006, 07:16 PM

Quote from: K on September 13, 2006, 07:15 PM
Quote from: Spilled on September 13, 2006, 07:02 PM
I'm not sure if you misread but I need the binaries, not the only the dll but the lib and header files.

Well you should have said "where can I get the lib and header files" then

Give me a minute, and I'll update the page/download to include both.

k Thanks alot

K · September 13, 2006, 07:22 PM

Done.

Stealth · September 13, 2006, 07:23 PM

Quote from: K on September 13, 2006, 07:22 PM
Done.

It would seem that VC7 is not properly importing yvals.h. The definition for _STD_BEGIN is not around properly when it tries to build the project. Argh.

Ersan · September 14, 2006, 08:24 AM

BNCSUtil seems to have the same problem that BNLS did (omg wonder what they're using?)

The new gigantic integers are getting truncated somewhere and I can only connect 10% of the time and checkrevision seems to fail 50% of the time, an update would be cool, or I'll try to figure it out.

Spilled · September 14, 2006, 10:59 AM

Quote from: Ersan on September 14, 2006, 08:24 AM
BNCSUtil seems to have the same problem that BNLS did (omg wonder what they're using?)

The new gigantic integers are getting truncated somewhere and I can only connect 10% of the time and checkrevision seems to fail 50% of the time, an update would be cool, or I'll try to figure it out.

Hrmmm, I'm getting Invalid Game Version response but I thought it was something I was doing wrong... Someone should double check this... Then again I could be doing something wrong... *shrugs*

K · September 14, 2006, 12:56 PM

I haven't seen this problem using BNCSUtil (actually, the old version with a patch for the new checksum algorithm) running on Ubuntu AMD64. I haven't really looked at anything that could be causing this, but since BNCSUtil uses GMP for handling large integers, I don't see why battle.net increasing the length would cause any problems.

Spilled · September 14, 2006, 01:14 PM

When i call checkRevision from the binaries you posted it returns me a bad CheckSum....

K · September 14, 2006, 01:36 PM

Quote from: Spilled on September 14, 2006, 01:14 PM
When i call checkRevision from the binaries you posted it returns me a bad CheckSum....

Actually, I'm not sure what I was talking about. GMP is used for the NLS functions only. The problem with checkRevision (this is the old version that I'm looking at, but I bet it's the same) is this:

Code Select


  long values[4]; // ...
  values[variable] = atol(token)

And the reason it works for me and not for you is that since I've compiled for AMD64, sizeof(long) on my machine is 8 instead of 4 bytes for 32bit.

I'll look into coming up with a fix.

Spilled · September 14, 2006, 01:45 PM

Quote from: K on September 14, 2006, 01:36 PM
Quote from: Spilled on September 14, 2006, 01:14 PM
When i call checkRevision from the binaries you posted it returns me a bad CheckSum....

Actually, I'm not sure what I was talking about. GMP is used for the NLS functions only. The problem with checkRevision (this is the old version that I'm looking at, but I bet it's the same) is this:

Code Select Expand
long values[4]; // ... values[variable] = atol(token)

And the reason it works for me and not for you is that since I've compiled for AMD64, sizeof(long) on my machine is 8 instead of 4 bytes for 32bit.

I'll look into coming up with a fix.

Ahh, had me worried there for awhile, thought it was something i was doing. Thanks

K · September 14, 2006, 01:59 PM

I've uploaded a new version (same place). I didn't test it because I have no way to, but you can give it a try.

Mystical · September 14, 2006, 02:03 PM

Ive tried it on using war3 doesn't work for me, will see if for anyone else.

Spilled · September 14, 2006, 02:07 PM

I'm still recieving a bad checksum =\ hrmm...

Edit:

My mistake, Version check is passing correctly everytime for me.

K · September 14, 2006, 02:40 PM

Quote from: Spilled on September 14, 2006, 02:07 PM
I'm still recieving a bad checksum =\ hrmm...
Edit:

My mistake, Version check is passing correctly everytime for me.

Sounds good. I just verified that the fix I made worked using the test values that Hdx posted in the other thread.

Using no modifications,

Code Select


SEXP
ver-IX86-5.mpq
A=3028046468 C=4177216242 B=3749337048 4 A=A-S B=B-C C=C^A A=A+B.
checksum (should be 0x6aafef39): 699a1a4b

with fix:

Code Select


SEXP
ver-IX86-5.mpq
A=3028046468 C=4177216242 B=3749337048 4 A=A-S B=B-C C=C^A A=A+B.
checksum (should be 0x6aafef39): 6aafef39