Topic: Open Transactions Server: Asset/Bond/Commodity/Cryptocoin/Deed/Share/Stock Exch. - page 5. (Read 42605 times)

Yes it works pretty well now, you should be able to go ahead and try it now, the actual server and client work, next update is just migrating the latest application-level communications protocol/handshake type stuff to the scripting system so the scripts can enjoy the new robustness the client's communication with the server now enjoys.

-MarkM-

Maybe you don't remember me asking :p

I wasn't really able to spend too much time trying to get OT to work (was starting GLBSE), it's been a year since then, I'm pretty bloody sure it's matured and stabilised since then, you and others have been working on it a lot.

I still don't have much time but I'm guessing it's not going to be too much to get started this time around, and it will be cool to finally start using it.

You never asked me for any help with any problems a year ago.

Maybe there weren't problems at all.

:-)

Maybe you just configured it wrong and then didn't ask me.

-FT

This would be sooooo nice...

https://bitcointalksearch.org/topic/moneychanger-windows-builds-out-of-date-77301

Open Transactions is getting pretty robust now it seems.

-MarkM-

Hi
Is there a pre built version for windows? Ready to go? I would like to do some quick testing moneychanger i think it's called with out installing everything

Open-Transactions is network-protocol-agnostic. It uses a callback and you can run it over any protocol you want that supports request/response. (Like web protocols, or TCP, etc.)

The current prototype of OT uses ZeroMQ, which you can read about here:  http://www.zeromq.org/

Markm was referring more to the application protocol, which I have been working on lately to improve OT's auto-synchronization, and its ability to handle network problems (missed messages, timeouts, etc.) The most recent check-in was the result of a couple month's work in this area, and I've still got a few hiccups left to correct before I'll be fully satisfied with it.

-FT

What networking protocol are you using, markm? Is that an open-standard communications protocol approved by the Internet Engineering Task Force (IETF)? If not where can I read the specifications?

Soon. Still a few little problems today, but soon...

-MarkM-

Was thinking would use OT to allow people to move their assets off GLBSE (for ones that were created on GLBSE).

I'ts been a year since I've looked at OT.

Should work no problem right?

Okay, the latest round of deep networking/protocol changes are over, we are back to testing, and irning out the wrinkles in the new improved system.

Meanwhile the first round of IPOs in complete, a bunch of DeVCoin oriented Corps. Four of them, each issuing one million shares at 20 DVC per share, all snapped up.

-MarkM-

It does seem that the signatures problem is fixed, however we have now been experiencing delays while the messaging protocol is being re-done to be much more robust, since in real life on the net one can never really be sure one will hear back at any point in the message-passing.

This has meant a lot of coding deep in the guts of the thing; the next round of tests will be to make sure those changes have not introduced new bugs of their own.

Basically the actual accounting has proven to be very sound and robust but the networking, the actual communication been server and clients, was not as robust as it could be thus is being re-done.

-MarkM-

A hash that evaluates in very large integers as being a larger integer than the modulo used in the signing algorithm is an error. There is some workaround you are supposed to do for this, some kind of padding or something.

DO enough transactions, and eventually you generate a message whose signature actually needs this padding or whatever it is, so we were getting bad signatires on some of the client<-> server messages when doing large numbers of messages.

It was feared that changing the signature code to do the correct padding or whatever to prevent this error would result in all signatures being different with the new code, thus all previously signed things having to be re-signed, which in the case of things like asset contracts would change their ID, which in turn would change IDs derived from them like account IDs, so, basically, all the data would have to be done over again starting with signing all the asset contracts and even the server contract itself with the new code. It would have in effect been a different server even, as its ID, which is the hash of its signed contract, would have been different.

Luckily it turns out that all the non-error signatures already done on the past continue to work fine with the new code. It seems it only affects the cases in which the error would have happened. (Assuming, that is, that it has in fact prevented the error from happening anymore. I should go check account inboxes now that the batches of market offers I made yesdterday or earlier will have expired. In the past all the large batches of offers caused glitched inboxes due to at least some of the messages having the error. If it has indeed been fixed I should now be able to process all the offer-expired messages. So I shall go try that now...

-MarkM-

Can you make a summary about what those changes improve?

Thank you.

Wow, we now have in place a bunch of deep fixes to the crypto that we thought would result in having to create all the contracts and so on over again as we thought their signatures would change from fixes to the signature system, but all the existing data seems to be working fine!

SO grab the latest Open Transactions and Moneychanger from github and come check it out!

-MarkM-

i dont fully understand how OT works.  but perhaps you smart fellas could join in on the conversation about the p2p currency exchange.

Oh yeah... here is a development build of Moneychanger...  still very much in development... but good for playing arround.


Moneychanger_5-02-2012.7z

https://github.com/da2ce7/Moneychanger/downloads

Ok.

The first windows builds in a little while have been made:

(my master branch was used to create them)
https://github.com/da2ce7/Open-Transactions


find the OT_Bin_Windows_06-02-2012.7z file on the following page
https://github.com/da2ce7/Open-Transactions/downloads


Edit... updated pgp message.

Updated my master branch to include the work done by washort.

Everyone Please test and compile...

any MSVC guru's out there... I'm getting this weird link.exe hang... it seems to happen a bit randomly...

OK... there are two Projects that need testing now:

There has been much work in making OT more portable to windows, and lots of general code file structure changes.

Major Changes:

• Completely Remade Visual Studio Project Files,  now we have two solutions.  1.  project/open-transactions  2. library/library
• Compiling Successfully on Windows 7 with solution files.   Working for both 32bit and 64bit targets
• Reorganized the source code so that it now lives in the 'src' sub-folder.  This makes finding different parts of OT much more natural

Please checkout my git: https://github.com/da2ce7/Open-Transactions
Do some testing.... Follow the instructions in the docs/INSTALL-VisualStudio.txt guide.

If you want to use a pre-compiled boost for windows please use my upload: https://github.com/downloads/da2ce7/Open-Transactions/boost_1_48_0_precompiled.7z

Overall have fun!

And keep testing my Moneychanger master also