Hi Greg,
On Tue, May 10, 2016 at 03:24:36PM +0000, Greg Smith wrote: > From: Sylvain Rochet [mailto:[email protected]] > > > > Not really, the last user of the SIO framework is SLIP. The SIO API > > sucks, I proposed my help to get rid of it in SLIP in > > <[email protected]<mailto:[email protected]>> > > > > but the change was not acknowledged. > > There are so many other changes in the PPP, that I, personally, have > had to make enough changes that this would not bother me if it broke > compatibility. I guess you misread. PPP is not SLIP. PPP is not using the SIO API anymore, SLIP is the last SIO user. > (It almost does anyway because sio_write() can be passed directly to > pppos_create(), for example.) Well, considering: PPP output callback signature is: u32_t ppp_output_cb(ppp_pcb *pcb, u8_t *data, u32_t len, void *ctx); sio_write signature is: u32_t sio_write(sio_fd_t fd, u8_t *data, u32_t len); If it works, that's only by luck, because you only need data and len, and because the size of a sio_fd_t is the same size of a ppp_pcb* because they are probably both pointers, that's not something that was done on purpose. > This is basically where I was headed myself, but it seems redundant; > therefore, I was figuring I was doing something wrote. Ideally, if I > could get rid of sio.h altogether, that would seem to be cleaner. Do > you think that is possible in the current state of 2.0.0RC1? Either > way, for now, I'll do much like what you've listed here. Of course it is, PPP is not using the SIO API anymore for, humm, almost a year. There is not a single reference to SIO in the whole PPP code (except in the documentation maybe to tell that we are not using it anymore :p). > Thank you for all your help! You're welcome :) Sylvain
signature.asc
Description: Digital signature
_______________________________________________ lwip-users mailing list [email protected] https://lists.nongnu.org/mailman/listinfo/lwip-users
