> > > Which is also a requirement. When we call namedpipe_create, we have to > > > RETURN > > > SOMETHING!
Maybe not... What if we open a namedpipe in Windows in bidirectional mode. Then we call CreateFile(pipename,...) for mode read and CreateFile(pipename, ...) for write mode. Not tested but it might work. >>>> Win32 will create a pipe handle (not the same as the read/write file > > > handle.) Every (NT/2000) machine could Create or Connect to get that > > > pipe handle. > > > But once that pipe handle is closed, the pipe evaporates, they are not > > > persistant. > > > > Not to be contrary, but can't we handle this by registering a cleanup with > > the handle? That way, the handle survives until we specifically clean it > > up, and we shouldn't have to change the API much. Not saying this is the > > right way to go, I am just offering suggestions. > > No, I don't think it's the right way. I'd expect that we add a placeholder of > apr_namedpipe_t that contains the name of the pipe on Unix (and nothing more), > or the pipe name and handle on Win32. Opaque, of course. > If the above works, then passing in an "open named pipe" flag on the apr_file_open() would enable APR to prepend the file://pipe/ stuff Bill
