[Openvpn-devel] NTLMv2 patch corrected

2008-01-15 Thread Mirek Zajic
I'm sending the patch again. Now it should be correctly diffed against version 2.1. Configuration is done with standard http-proxy command in config file, with 'ntlm2' setting for auth method, for. ex.: http-proxy 192.168.0.1 8080 stdin ntlm2 Username to proxy can be passed among with auth. d

Re: [Openvpn-devel] NTLMv2 patch corrected

2008-01-15 Thread Alon Bar-Lev
On 1/15/08, Mirek Zajic wrote: > I'm sending the patch again. Now it should be correctly diffed against > version 2.1. Great! But you have change the indentation, so your changes cannot be reviewed. For example, the first function create_des_keys, you did not modify it, but still it modified. P

Re: [Openvpn-devel] NTLMv2 patch corrected

2008-01-15 Thread Alon Bar-Lev
Please also don't keep changing the subject... Just reply to the same thread. On 1/15/08, Alon Bar-Lev wrote: > On 1/15/08, Mirek Zajic wrote: > > I'm sending the patch again. Now it should be correctly diffed against > > version 2.1. > > Great! > > But you have change the indentation, so your c

Re: [Openvpn-devel] NTLMv2 patch corrected

2008-01-15 Thread Mirek Zajic
Thank you for info. I wanted to use equal indentation rules in the whole file. I'm sending another patch without these indentation changes. Also, this code was originally written for x86 processors with 32-bit environment. Different wordsize or endian may cause trouble when porting to other pl

Re: [Openvpn-devel] NTLMv2 patch corrected

2008-01-15 Thread Alon Bar-Lev
On 1/15/08, Mirek Zajic wrote: > Thank you for info. I wanted to use equal indentation rules in the whole > file. I'm sending another patch without these indentation changes. > > Also, this code was originally written for x86 processors with 32-bit > environment. Different wordsize or endian may c

Re: [Openvpn-devel] NTLMv2 patch corrected

2008-01-15 Thread Mirek Zajic
The keep-alive is enabled only when ntlm or ntlm2 is used. The reason is that proxy may close the connection before ntlm messages are sent. It is the correct way to always use keep-alive when ntlm is used. I'm going to fix the architecture independency and other things and I'll send the patch