*bzr* under the firewall

BZR could not work well now in the intranet that has an firewall. I am working behind the firewall. I have to use the http proxy to access networking. There are some problems when I use the bzr, a distributed version control system tool which used in opencog.

The first problem is that it could not use the *lp:opencog*. When I used 'bzr branch lp:opencog' it give me a 'connection time out ERROR', even though I set the http_proxy enviroment correctly. After asking this question in the IRC channel of #bzr at freenode, I got the answer that the bzr is not supported the *lp:opencog* by using the http_proxy now . I should replace it with the URL of opencog trunk . So I used 'bzr branch http://bazaar.launchpad.net/~opencog-dev/opencog/trunk/' to make a local branch instead of using 'bzr branch lp:opencog' and it worked.

The second problem is that I could not use'bzr push lp:~xiaohui/opencog/moses' to push it to my branch of opencog on the lannchpad. So I changed it to
'bzr push https://bazaar.lauchpad.net/~xiaohui/opencog/moses' , but it also could not work. After discussing in the #bzr IRC and googling it. I know the reason is that bzr use the SSH to connect to launchpad. And my SSH could not connect to the networking because I am behind the firewall. So I need a proxy like http proxy or SOCKS proxy for SSH. After setting the proxy for SSH(it will be mentioned later) correctly, and then I use the command
'bzr push bzr+ssh://bazaar.lauchpad.net/~xiaohui/opencog/moses' to push it to the my branch for my gsoc project correctly.

There are some ways to set the proxy for the SSH, I was followed this link to set the ssh
through http_proxy by using corkscrew. And there also some other method to set the SSH proxy, the following links maybe useful for it.

1. using tsocks to tunnel
http://www.plenz.com/tunnel-everything
2. using connect.c to tunnel
http://bent.latency.net/bent/darcs/goto-san-connect-1.85/src/connect.html

Accepted!!!

I am so excited that my proposal was accepted by opencog. It means that I could spend my summer time to play with the opencog.It is really cool and meaningful, I like it .

From the beginning of April, I was very busy doing my mater projects . During these time, I have write the third patch of MOSES, but it has some bugs. I don't have enough time to fix it.Some days later, Bogdan submited the same patch of gen-mixed-table and it works well(Thanks ,Bogdan).

The next stage I need to do is following the schedule in my proposal . This is the Stage I. It should be finished before starting coding .

1. Keep learning boost
2. Reading the papers about MOSES, BBHC, and SAA, to master this algorithm
3. Reading the document of Opencog and MOSES
4. Reading the code of MOSES and BBHC and do some little change to the code, like to add the comment of the code with the format of doxygen and patches.

The second patch for MOSES

Today , I've finished the second patch, which is to generate a truth table complete or partial. After the first patch , it is easier to finish the second one.

In these two days , except for doing the school things , I spent the most of my time on reading the code of comboreduct and moses. And now , I have a general impression of them. I know how the combo program are stored and the basic datasturture and algorithm.

In the coming days, I plan to do the third patch about the mixed-table generation , to profile how moses works in detail and then to learn how to integrate the bbhc to moses.