Angus Leeming <[EMAIL PROTECTED]> writes:
| On Thursday 12 September 2002 11:08 am, Lars Gullik Bjønnes
| wrote:
>
>> the main loop in the LocalClientListener looks like this:
>>
>> sock_.listen(5);
>
| what is sock_ ? Is this the "default socket" (ie has a name
| known to the client) thro
On Thursday 12 September 2002 11:08 am, Lars Gullik Bjønnes
wrote:
> the main loop in the LocalClientListener looks like this:
>
> sock_.listen(5);
what is sock_ ? Is this the "default socket" (ie has a name
known to the client) through which this client can ask the
server to create a n
Angus Leeming <[EMAIL PROTECTED]> writes:
| On Wednesday 11 September 2002 11:21 pm, Lars Gullik Bjønnes
>
>> No, they don't share the same socket.
>>
>> it will be code "like" this:
>>
>> int sock = socket(ARGS)
>> bind(sock, ARGS)
>> while true; do
>> int new_sock = accept(sock. ARGS);
>>
On Wednesday 11 September 2002 11:21 pm, Lars Gullik Bjønnes
> No, they don't share the same socket.
>
> it will be code "like" this:
>
> int sock = socket(ARGS)
> bind(sock, ARGS)
> while true; do
> int new_sock = accept(sock. ARGS);
> // register this new_sock with a stream
> // re
Angus Leeming <[EMAIL PROTECTED]> writes:
| On Wednesday 11 September 2002 6:31 pm, Lars Gullik Bjønnes
| wrote:
|
| > well no... there should be no need to really know the fd after
| > you have passed it into the streamclass.
|
| > this exposes too much low-level for my liking...
|
| Fair en
On Wednesday 11 September 2002 6:31 pm, Lars Gullik Bjønnes
wrote:
> well no... there should be no need to really know the fd after
> you have passed it into the streamclass.
> this exposes too much low-level for my liking...
Fair enough. But somehow the lyxserver needs to know /which/
client
Angus Leeming <[EMAIL PROTECTED]> writes:
| > Close but not quite. New connections (when speaking socket)
| > must be handled with an accept. The fd returned from this
| > accept can be passed to a socket_buf, and the rest can then be
| > handled by your stream. (except that you now have to prepa
On Wednesday 11 September 2002 5:56 pm, Lars Gullik Bjønnes
wrote:
> | I think that your sockets are similar in this respect to my
> | "client" pipes. Multiple "clients" can connect to existing
> | pipes lyxpipe.in/out and communicate with lyx
> | simultaneously.
>
> As I showed with pipes... th
Angus Leeming <[EMAIL PROTECTED]> writes:
| On Wednesday 11 September 2002 5:05 pm, Lars Gullik Bjønnes
| wrote:
| > | I'll make it easy for you. Here's the "bloody code". As you
| > | can see there's a template parameter commT.
| > |
| > | commT can be
| > | pipecomm_client
| > | pipecomm_s
On Wednesday 11 September 2002 5:05 pm, Lars Gullik Bjønnes
wrote:
> | I'll make it easy for you. Here's the "bloody code". As you
> | can see there's a template parameter commT.
> |
> | commT can be
> | pipecomm_client
> | pipecomm_server
> | socketcomm_client
> | socketcomm_serv
Angus Leeming <[EMAIL PROTECTED]> writes:
| On Wednesday 11 September 2002 4:41 pm, Lars Gullik Bjønnes
| wrote:
| > Angus Leeming <[EMAIL PROTECTED]> writes:
| > | On Wednesday 11 September 2002 2:39 pm, Lars Gullik Bjønnes
| > |
| > | wrote:
| > | > This is how I think this should be done.
| >
On Wednesday 11 September 2002 4:41 pm, Lars Gullik Bjønnes
wrote:
> Angus Leeming <[EMAIL PROTECTED]> writes:
> | On Wednesday 11 September 2002 2:39 pm, Lars Gullik Bjønnes
> |
> | wrote:
> | > This is how I think this should be done.
> | >
> | > lyx<->lyxserver--localsocket--pipesocket(as a
>
Angus Leeming <[EMAIL PROTECTED]> writes:
| On Wednesday 11 September 2002 2:39 pm, Lars Gullik Bjønnes
| wrote:
| > This is how I think this should be done.
|
| > lyx<->lyxserver--localsocket--pipesocket(as a socketclient)
| >\
| > socketclient
On Wednesday 11 September 2002 2:39 pm, Lars Gullik Bjønnes
wrote:
> This is how I think this should be done.
> lyx<->lyxserver--localsocket--pipesocket(as a socketclient)
>\
> socketclient
>
> what I mean is that the localsocket is the native int
Angus Leeming <[EMAIL PROTECTED]> writes:
| On Wednesday 11 September 2002 12:03 pm, Lars Gullik Bjønnes
| wrote:
>
>> Busy now, but I will have a look later. But I do not expect
>> these changes to go into 1.3.0 anyway.
>
| Lars, I have made some progress.
>
| pipestream::eof() returns 1 after
On Wednesday 11 September 2002 12:03 pm, Lars Gullik Bjønnes
wrote:
> Busy now, but I will have a look later. But I do not expect
> these changes to go into 1.3.0 anyway.
Lars, I have made some progress.
pipestream::eof() returns 1 after the first write, so an
explicit call to pipestream.clea
On Wednesday 11 September 2002 12:03 pm, Lars Gullik Bjønnes
wrote:
> | Thus the problem lies in overflow() sync() or flush() but my
> | knowledge is obviously too limited. Can anyone help?
>
> Busy now, but I will have a look later. But I do not expect
> these changes to go into 1.3.0 anyway.
Angus Leeming <[EMAIL PROTECTED]> writes:
| Anybody know anything about the details of stream buffers?
>
| Attached is my work to date on creating a viable pipestream.
| Compiling the code results in an executable "client" that can be
| used to interact with LyX. It's still pretty clunky, but y
Anybody know anything about the details of stream buffers?
Attached is my work to date on creating a viable pipestream.
Compiling the code results in an executable "client" that can be
used to interact with LyX. It's still pretty clunky, but you get
the idea.
The design:
the pipestream buffer
19 matches
Mail list logo
