ICC + timestamp

Discussion of chess software programming and technical issues.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
bob
Posts: 20795
Joined: Mon Feb 27, 2006 6:30 pm
Location: Birmingham, AL

ICC + timestamp

Post by bob » Mon May 19, 2014 2:34 am

Anyone aware of an OS/X timestamp for ICC? My office machine is now a 4-core iMac and in trying to get things set up for ICC, I ran into this issue. I run Crafty on one node of our 12-core cluster, but I can't run xboard and such as we have none of that installed on the actual compute nodes. I generally just run xboard on my office box and have it ssh to the cluster to run Crafty. But I can not find a working time stamp. They have a very old apple os timestamp on their download page, but it is for the way older power pc chip based Macs...

Anything available?

kinderchocolate
Posts: 412
Joined: Mon Nov 01, 2010 5:55 am
Full name: Ted Wong
Contact:

Re: ICC + timestamp

Post by kinderchocolate » Mon May 19, 2014 7:36 am

Bob,

Timestamp is a closed source software developed by ICC. Only ICC knows how they encrypt their data. It's a top security from ICC, nobody other than ICC can provide you an alternative. Their security has been reverse engineered 10 years ago, but not now.

Sargon
Posts: 4
Joined: Tue Nov 26, 2013 11:18 pm

Re: ICC + timestamp

Post by Sargon » Mon May 19, 2014 11:42 am

There's a timestamp available for OSX available under http://www6.chessclub.com/resources/icc/timestamp/.

But given its age (it's from 13-Sep-2003) I don't know whether it's working for current OSX versions.

Sargon

zullil
Posts: 5936
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: ICC + timestamp

Post by zullil » Mon May 19, 2014 3:27 pm

Sargon wrote:There's a timestamp available for OSX available under http://www6.chessclub.com/resources/icc/timestamp/.

But given its age (it's from 13-Sep-2003) I don't know whether it's working for current OSX versions.

Sargon
No good:

Code: Select all

ProcyonLeo: ~/Downloads] file timestamp.macosx_6.6 
timestamp.macosx_6.6: Mach-O ppc executable

bob
Posts: 20795
Joined: Mon Feb 27, 2006 6:30 pm
Location: Birmingham, AL

Re: ICC + timestamp

Post by bob » Mon May 19, 2014 4:58 pm

Sargon wrote:There's a timestamp available for OSX available under http://www6.chessclub.com/resources/icc/timestamp/.

But given its age (it's from 13-Sep-2003) I don't know whether it's working for current OSX versions.

Sargon
That's the power PC chip version. No good for Intel...

bob
Posts: 20795
Joined: Mon Feb 27, 2006 6:30 pm
Location: Birmingham, AL

Re: ICC + timestamp

Post by bob » Mon May 19, 2014 5:00 pm

kinderchocolate wrote:Bob,

Timestamp is a closed source software developed by ICC. Only ICC knows how they encrypt their data. It's a top security from ICC, nobody other than ICC can provide you an alternative. Their security has been reverse engineered 10 years ago, but not now.
I understand. I had hoped ICC had an updated Mac version hidden away somewhere.

BTW their security is not so hot. Couple of guys wrote a paper detailing the weaknesses. They broke it easily. Also broke the "supposedly secure" protocol they use to get your credit card info...

mvk
Posts: 589
Joined: Tue Jun 04, 2013 8:15 pm

Re: ICC + timestamp

Post by mvk » Mon May 19, 2014 6:00 pm

bob wrote:BTW their security is not so hot. Couple of guys wrote a paper detailing the weaknesses. They broke it easily. Also broke the "supposedly secure" protocol they use to get your credit card info...
They changed that over 10 years ago, and, in Internet years, have relied on HTTPS since Methusalah was a toddler.
[Account deleted]

bob
Posts: 20795
Joined: Mon Feb 27, 2006 6:30 pm
Location: Birmingham, AL

Re: ICC + timestamp

Post by bob » Mon May 19, 2014 6:34 pm

mvk wrote:
bob wrote:BTW their security is not so hot. Couple of guys wrote a paper detailing the weaknesses. They broke it easily. Also broke the "supposedly secure" protocol they use to get your credit card info...
They changed that over 10 years ago, and, in Internet years, have relied on HTTPS since Methusalah was a toddler.
I've not followed what ICC does regarding IP security. I've never had the inclination to give 'em my credit card info so it didn't matter. We've ALWAYS seen hackers breaking timestamp, however, just not very widespread. I suppose the only choice is to take the linux timestamp, disassemble, and modify to work on OS/X. The changes are not significant so long as they avoided the "int" instruction to do I/O, but even that is fixable (I just did this for a library I use in my X86 asm course in fact, to make it work on Mac OSX.).

Was hoping to avoid the effort.

One good bit of news, I just used objdump and it appears to be all C with no asm, at least none referencing the int instruction. It also has procedure names included (not stripped) but there appear to be no variable names..

mvk
Posts: 589
Joined: Tue Jun 04, 2013 8:15 pm

Re: ICC + timestamp

Post by mvk » Mon May 19, 2014 6:53 pm

bob wrote:
mvk wrote:
bob wrote:BTW their security is not so hot. Couple of guys wrote a paper detailing the weaknesses. They broke it easily. Also broke the "supposedly secure" protocol they use to get your credit card info...
They changed that over 10 years ago, and, in Internet years, have relied on HTTPS since Methusalah was a toddler.
I've not followed what ICC does regarding IP security. I've never had the inclination to give 'em my credit card info so it didn't matter. We've ALWAYS seen hackers breaking timestamp, however, just not very widespread. I suppose the only choice is to take the linux timestamp, disassemble, and modify to work on OS/X. The changes are not significant so long as they avoided the "int" instruction to do I/O, but even that is fixable (I just did this for a library I use in my X86 asm course in fact, to make it work on Mac OSX.).

Was hoping to avoid the effort.

One good bit of news, I just used objdump and it appears to be all C with no asm, at least none referencing the int instruction. It also has procedure names included (not stripped) but there appear to be no variable names..
I'm afraid it is the only fast way indeed. (other than just running the front-end on a rpi or somthing). Or don't use Xboard at all, but use icsdroneng on the Linux node. For XBoard with OSX, I use xboard without timestamp ever since Rosetta was withdrawn. For FreeBSD, I'm relying on 32-bit linux emulation as well, and don't expect it will work forever. When that fails I will probably switch those systems to Linux anyway.

Reverse engineering for this purpose seems to be allowed by the DMCA, but personally I wouldn't want to try my luck with an US-based company, even ICC.

But speaking of Linux, weren't you going to switch over anyways?
[Account deleted]

bob
Posts: 20795
Joined: Mon Feb 27, 2006 6:30 pm
Location: Birmingham, AL

Re: ICC + timestamp

Post by bob » Mon May 19, 2014 10:33 pm

mvk wrote:
bob wrote:
mvk wrote:
bob wrote:BTW their security is not so hot. Couple of guys wrote a paper detailing the weaknesses. They broke it easily. Also broke the "supposedly secure" protocol they use to get your credit card info...
They changed that over 10 years ago, and, in Internet years, have relied on HTTPS since Methusalah was a toddler.
I've not followed what ICC does regarding IP security. I've never had the inclination to give 'em my credit card info so it didn't matter. We've ALWAYS seen hackers breaking timestamp, however, just not very widespread. I suppose the only choice is to take the linux timestamp, disassemble, and modify to work on OS/X. The changes are not significant so long as they avoided the "int" instruction to do I/O, but even that is fixable (I just did this for a library I use in my X86 asm course in fact, to make it work on Mac OSX.).

Was hoping to avoid the effort.

One good bit of news, I just used objdump and it appears to be all C with no asm, at least none referencing the int instruction. It also has procedure names included (not stripped) but there appear to be no variable names..
I'm afraid it is the only fast way indeed. (other than just running the front-end on a rpi or somthing). Or don't use Xboard at all, but use icsdroneng on the Linux node. For XBoard with OSX, I use xboard without timestamp ever since Rosetta was withdrawn. For FreeBSD, I'm relying on 32-bit linux emulation as well, and don't expect it will work forever. When that fails I will probably switch those systems to Linux anyway.

Reverse engineering for this purpose seems to be allowed by the DMCA, but personally I wouldn't want to try my luck with an US-based company, even ICC.

But speaking of Linux, weren't you going to switch over anyways?
I just got this iMac for the office. I might well make that switch. Only problem I have had is UAB has become a microsoft environment, and while open office / libre office are good, they are not perfect and cause some problems.

I'm not totally sold on OS/X yet. Works better on my iMac (process scheduler seems to understand hyper threading reasonably well as opposed to on my macbook. But it has enough quirks and apple certainly loves to ship broken software. I've converted my x86 class to nasm. Apple's distributed nasm is broken. macports version works just fine. Ditto for clang vs gcc 4.7. Etc.

Post Reply