#linuxcnc-devel | Logs for 2013-10-02

[07:16:30] <skunkworks> I can see that could be a problem if you can't accelerate during a blend.. (if the segments end up being mostly blends)
[08:19:03] <skunkworks> I had a dream last night that rellenberg called me and wanted to talk about lookahead. I was trying to explain to him that I didn't understand what he was talking about.. Must have been the pizza
[08:32:39] <cradek> haha
[08:32:45] <cradek> I have no idea here have some pizza
[09:16:17] <cradek> oh no stuart nooooooo
[09:20:52] <skunkworks> what is he thinking!
[09:21:21] <cradek> rather what is he failing to remember
[09:21:36] <cradek> haven't we gone through this in recent memory?
[09:22:24] <skunkworks> that is the great thing about the internets-- you can always search history
[09:22:51] <cradek> I have lots of evidence that not everyone knows this :-)
[10:47:45] <skunkworks> so close with the 8168
[10:47:46] <skunkworks> http://www.mail-archive.com/rtnet-users@lists.sourceforge.net/msg03371.html
[10:48:05] <skunkworks> that is what I get - but don't understand what the changing 1 into 2 means...
[10:49:38] <skunkworks> and why you would be changing the r8169 driver and not the rt_r8169
[11:04:46] <jepler> skunkworks: I looked at the post and the linked code and I don't understand what change he means
[11:05:07] <skunkworks> ok - I think I am going to have to get on the mailing list and ask
[11:08:29] <mhaberler> cradek: re jwp, ja3 and Seb's mail - sorry, this is too much of a moving target, so I will stop for now until this is rebased on master. I do have the jwp code working on top of ja3, and beginnings of commanded jogs.
[11:10:09] <mhaberler> IMO it is woth considering folding teleop into coord, I dont see the upside
[11:10:12] <cradek> your work on top of ja3 will apply just fine on top of ja4
[11:10:40] <cradek> I don't know what you mean by folding
[11:11:26] <jepler> cradek: (mhaberler worked by merging master plus jwp into ja3, saying rebase was the worse choice. So if he works more on top of his wip branch, he'll have to merge ja3 into ja4 which we don't want)
[11:12:15] <cradek> oh right you don't want that merge
[11:13:17] <mhaberler> other than that, there were no surprises except a few new bugs
[11:13:21] <cradek> I think we don't want most of jwp (the new kind of jogging using hal pins) so starting with that branch is probably the wrong approach
[11:14:06] <cradek> and the new configs and panels that allow the new kind of jogging
[11:14:56] <cradek> I suggest thinking about what parts you really will want for master eventually, and don't merge all the other things
[11:15:14] <jepler> which entails necessarily working by rebase (or from scratch)
[11:15:40] <cradek> yes or a few strategic cherry picks
[11:17:30] <mhaberler> well it would be super helpful to come to a conclusion before prodding me into work
[11:18:08] <cradek> that's a weird accusation
[11:19:13] <cradek> I'm trying to help you end up with something that can be easily used on top of seb's work of cleaning up many years of history on ja3 without another big history-cleanup type problem
[11:20:25] <mhaberler> so what are we discussing here, the nature of the patch or where it's going?
[11:20:35] <cradek> merging a bunch of unwanted stuff (and I thought we both agreed it was unwanted, the other day) is the wrong approach, and I never suggested it, and I never prodded you
[11:21:08] <mhaberler> I took away something rather different
[11:21:26] <mhaberler> we agreed that integrating wheel and command jog makes sense
[11:21:31] <cradek> if I was unclear I apologize
[11:21:45] <mhaberler> well yes you were
[11:22:04] <cradek> yes, so the jogs while paused can work as expected, instead of in a new way requiring a new panel etc
[11:22:42] <mhaberler> you certainly are aware this entails UI changes too
[11:23:15] <jepler> yes, it was commented on in the meeting that this would entail UIs changing their ideas of when the jog functionality was available
[11:23:21] <jepler> a UI that doesn't take the time to do this won't be able to jog in the new conditions
[11:23:37] <jepler> this is clearly better than having *this set* of jog UI which you use sometimes, and *that set* you use at other times
[11:24:00] <mhaberler> what the guys where rather clear about, and what you dont seem to like: this is something working _now_, independently of other folks fixing their UI's task and whatnot
[11:24:08] <cradek> yes that was my objection to your existing approach since day 1
[11:24:55] <jepler> just changing when axis disables the UI elements for jogging is trivial compared to everything else
[11:25:06] <mhaberler> well first, the core part of the patch - the pause FSM - will be identical whether there are hal pins or not, the only thing which changes is how commanded pos is aquired
[11:25:13] <cradek> I don't care about that, since they can use it as-is for as long as they like. our new goal is to make something suitable for merging into master, which means it does not have the separate type of jog
[11:25:31] <cradek> what is FSM?
[11:25:36] <mhaberler> finite state machine
[11:25:58] <cradek> then cherry-pick the thing that does that part
[11:26:16] <cradek> leave out everything else that we would have to remove later anyway
[11:30:24] <cradek> it looks like you want most of http://git.mah.priv.at/gitweb?p=emc2-dev.git;a=commitdiff;h=778b620606052ed6efc881b6f0710860842cc98d
[11:30:53] <cradek> then you will remove the hal-dependent stuff (maybe use rebase -i to squash in that removal) and continue
[11:32:13] <mhaberler> I do not like how you proclaim 'our new goals' - make it a goal decision item for the IRC meeting, discuss it on the list beforehand like you want from everybody else. We have decided to not make decisions on IRC.
[11:33:13] <cradek> now you just want to argue and I decline. bbl.
[18:52:55] <KGB-linuxcnc> 03andy 05ssi-fanuc-biss-dpll 8d4dcc5 06linuxcnc 10src/hal/drivers/mesa-hostmot2/abs_encoder.c * Tweaks to Fanuc. Also no longer require a name (no dummy name for patch bits)