This is the mail archive of the cygwin@sources.redhat.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: cygwin with sockscap32


MADHU wrote:
> 
> helu, without any disrespect: I am not skilled in win32 api's and
> unfortunately I am not qualified enough to understand the mechanisms -
> which was why I posted the original question, with the hope that
> someone who recognized the call to the registry (where the application
> hangs) might explain what was going on.
> 
> sockscap (like runsocks) takes a binary program, and at runtime uses
> different socks-aware dlls after intercepting network calls. that is
> the extent of my understanding. 

If you read my message time-stamped 18:27:43 -0500, the problem is few
(if any) of US understood what sockscap WAS. Here's a WAG
(wild-ass-guess):

when a "new" application that sockscap has never seen before attemps to
call any WinSock functions, sockscap creates some sort of entry in the
registry FOR the application, where it will store information concerning
HOW this new application uses networking, what SOCKS protocol should be
used for it, etc.  So, in this case, socks32.dll is creating registry
keys for ssh.exe, and hangs -- looks to me like socks32.dll has the bug,
not ssh.  But, again, refer to sockscap FAQ#7 quoted in my other
message.

> I also did not understand the legal
> issues that came up. sockscap is distributed under a license, and I'm
> was not aware if it was illegal to run cygwin under sockscap.

No, it's illegal to distribute binaries that use cygwin, without (a)
distributing the source, or (b) buying a proprietary cygwin license. 
However, that doesn't apply to sockscap -- it doesn't "use" cygwin.  See
my other messages.

> I do know that f-secure and certain other products that use some
> security features (like license NTLM stuff) of windows do not run
> under sockscap: I was especially pleased that earlier cygwin did not
> complain.  From a naive "end user" point of view, I lose cool
> functionality that I had.

It appears this is because sockscap "makes assumptions about the
application and stack implementation" that are not valid for f-secure
and now, cygwin.

--Chuck

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]