Interesting! I will try to sort the scattered parts in my mind and on my computers and then I will put in my two cents.
But in short: Are there some historical reasons or other resentements against using the TCP-feature in NML? IMHO this feature should be upgraded/expanded. matsche On 2014-10-24 21:55, Jeff Epler wrote: > I intended to copy this message to the developers mailing list but it > doesn't seem to have shown up. Here's a copy for posterity. > > The key paragraph is > Since so few people are using NML over TCP in LinuxCNC, I believe that > we should incorporate a version of this patch in our master branch *even > if* it totally breaks NML over TCP, because 328/395 affect ALL of our > users. > if you have a stake in this decision please speak up. > > ----- Forwarded message from Jeff Epler <[email protected]> ----- > > Date: Tue, 21 Oct 2014 13:21:45 -0500 > From: Jeff Epler <[email protected]> > To: [email protected], [email protected] > Subject: Patch for issue #328 rebased onto master, resolves #395 as well > Message-ID: <[email protected]> > MIME-Version: 1.0 > Content-Type: text/plain; charset=us-ascii > Content-Disposition: inline > User-Agent: Mutt/1.5.21 (2010-09-15) > > I have rebased the original SF#328 patch onto the tip of the LinuxCNC > master branch. > > This work can be found on the tip of the branch 'jepler/master/sf328' on > git.linuxcnc.org. > > My patch should be like 'fix-emccommand-queue-fix1.patch', though I > started with 'fix-emccommand-queue.patch' and found and re-added the > missing call myself. > > I did not attempt to incorporate any form of the tcp serial number > fixes. > > For me, this resolved the #395 bug as well, and all runtests still pass. > > If you're affected by #328 or #395 please test with my branch and report > your results. > > Since so few people are using NML over TCP in LinuxCNC, I believe that > we should incorporate a version of this patch in our master branch *even > if* it totally breaks NML over TCP, because 328/395 affect ALL of our > users. > > (I would have said 'no users' but back in April, W. Martinjak (matsche) > indicated that he was currently using it successfully between 32-bit > machines and with limited success after some changes between 32-bit and > 64-bit machines) > > In a separate e-mail I've brought EBo (who wrote those patches) and > matsche (who reported using NML-over-TCP) into the loop about the #328 > patch. I haven't tried the later patch which intends to add reliable > global serial numbers to TCP. > > Jeff > > ----- End forwarded message ----- > > ------------------------------------------------------------------------------ > _______________________________________________ > Emc-developers mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/emc-developers -- "In der Wissenschaft siegt nie eine neue Theorie, nur ihre Gegner sterben nach und nach" Max Planck ------------------------------------------------------------------------------ _______________________________________________ Emc-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-developers
