Re: [pgsql-hackers-win32] Win32 lost signals open item

>> Basically, I think internal_forkexec() needs to be split up
>into two -
>> one win32 and one other. For win32 version, it needs to
>CreateProcess()
>> *before* it does write_backend_variables(), and then pass
>the process id
>> as a parameter to write_backend_vars().
>
>Huh? Why?

Because we need to write the duplicated socket structure/pipe handle to
the parameter file. I guess we could create a separate parameter file
just for these things, but that seemed a bit unnecessary.

Basically:
1) To create the new HANDLE, or in the case of a socket, the
WSAPROTOCOL_INFO structure, we need to know the new pid. We can only
know this after CreateProcess(), which sits in win32_forkexec.
2) To get the data down to the backend, we need to put it in the
parameter file, which is done using write_backend_vars().

This gives that we either:
1) Write the data out in write_backend_vars, in which case
write_backend_vars needs to know the pid of the new backend.
or
2) Write the data out in win32_forkexec, in which case we need another
parameter file.

I was thinking (1) was the cleaner approach.

//Magnus

"Magnus Hagander" <mha(at)sollentuna(dot)net> writes:
>> Huh? Why?

> Because we need to write the duplicated socket structure/pipe handle to
> the parameter file. I guess we could create a separate parameter file
> just for these things, but that seemed a bit unnecessary.

Do we actually need to pass the handle, or could the subprocess reopen
the pipe for itself?

regards, tom lane