> On Thu, Jan 31, 2019 at 04:57:47AM -0500, Frediano Ziglio wrote: > > > On Thu, Jan 31, 2019 at 03:50:07AM -0500, Frediano Ziglio wrote: > > > > > > > > > > On Wed, Jan 30, 2019 at 04:23:56AM -0500, Frediano Ziglio wrote: > > > > > > ping > > > > > > > > > > I'd drop the 'socket compatibility layer' stuff from that series to > > > > > make > > > > > it shorter, a 32 patch series with no real world use case is very low > > > > > priority to me. A shorter portability patch series is more likely to > > > > > get > > > > > a review from me. > > > > > > > > > > Christophe > > > > > > > > > > > > > Which part exactly of the "socket compatibility layer" ? If I drop > > > > everything > > > > code will end up with a lot of conditional code spread all over the > > > > code, > > > > I don't think we want to have that result. > > > > > > I'm just suggesting to start with a smaller series without the socket > > > compatibility layer, once that is reviewed and merged, you can then send > > > that socket stuff as a second series. > > > > > > Christophe > > > > > > > Sorry, still not clear. > > It seems to me you are asking for a full rewrite but you are not explaining > > how. > > The compatibility layer is to reduce the changes, if I remove the changes > > will > > be bigger but you want them smaller. > > Isn't the socket compatibility layer roughly patches 6 to 18 in the > series, with most patches before and after that not needing/being > related to the socket changes? > > Christophe > Not exactly, the patches before are not related to socket but most of the patches after the compatibility layer are still socket related. But I can move the patches not socket related before the socket ones and send them if you prefer and in the future send the others. Another change to reduce the code changes would be to continue to use straight "int" for sockets and not change the name of socket functions (like bind, listen, getsockname) but change required ones (like read, write, close). Frediano _______________________________________________ Spice-devel mailing list Spice-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/spice-devel