Enrico Granata

> Are you more concerned about iiod or in network.c? > > ``` > iiod/ops.c:256: ret = recv(pdata->fd_in, dest, len, MSG_NOSIGNAL); > iiod/ops.c:1315: ret = recv(pdata->fd_in, buf, len, > iiod/ops.c:1331:...

> Passing a NULL buffer is correct as long as no data is written - which is the case since we use MSG_TRUNC. It only works for TCP, but the...

> Just so everyone is on the same page: > > AF_UNIX creates filesystem objects and it only works between processes on the same host - so wouldn't it make...

> I'd be very cautious about this. Using stdio bite us in the past and that's the reason why we [switchted to file descriptors](https://github.com/analogdevicesinc/libiio/commit/ba809d05a06d3b81ba935c2b5152a70f71a92c3f#diff-2c45afd80e47b7d2f8b8be845be9696e). What I'd suggest to do, is...

This is reproducible via the `stat` tool Problem is that individual filesystems do not handle an attempt to open their root directory (or an empty path) correctly. For FatFS, this...

The former is done as of 3714535eba4fc80d66fac07c77e126c2056241a0 The latter is not exactly trivial to do because there is a potential issue where one could set the flag, run the callback,...

See https://github.com/egranata/puppy/issues/164 for required plumbing here

At a glance, the lexing makes sense ``` ARGUMENT env ARGUMENT FOO= QUOTED a;b TERMINATOR ``` The difference is that `FOO=a` is parsed as ``` ARGUMENT env ARGUMENT FOO=a TERMINATOR...

This is very likely to work, but needs testing

The fix for this is likely to be non-trivial and involve a fair bit of lexing magic if I want to end up sticking with `lex` Alternatives include: - rewriting...