[LinuxCNC/linuxcnc Issue#632] V 2.9 ignoring joint.2.home-sw-in while homing About GRBL HAL

Issue #632 | 状态: 已关闭 | 作者: rodw-au | 创建时间: 2019-08-10

Here are the steps I follow to reproduce the issue:

1. Previously running Master branch v2.9
2. Compiled latest version of 2.9 as of today
3. Homed machine

This is what I expected to happen:

YYXZ gantry Machine would home normally just as it did earlier today under earlier version of master branch V 2.9

This is what happened instead:

Joint 2 home switch input was ignored and limit switch was triggered. This behavior persisted on several retries.

It worked properly before this:

Yes was proven working on an earlier version of V2.9 immediatley prior the upgrade.

Information about my hardware and software:

This is a long standing master branch (v2.8 and V2.9) gantry machine running Linux Mint 17.3 RT_PREMPT and Mesa 7i76e
Checked all signals were working for this homing switch from the 7i76e input, to the signal and onto the joint pin.

#1 – andypugh 于 2019-08-10

Is this a buildbot install or do you compile it yourself?
Homing was touched here: https://github.com/LinuxCNC/linuxcnc/commit/5ea573adde87231cdc566e3759f85bb4037869f7
I guess you don’t know the version of 2.9 that was running before? (I wouldn’t if anyone asked me)

If you are using a buildbot package then you can use “package -> force version to go back a step or two, and see at which point the problem goes away. The versions match the hashes here: https://github.com/LinuxCNC/linuxcnc/commits/master (so “5ea573a” is the same as “1:2.9.0-pre0.433.g5ea573add (stable)”. If building with git then you can check-out the specifi commit(or the one immediately before it) and build.

#2 – rodw-au 于 2019-08-10

Andy, thanks for your interest. As you say, I don’t know what I was running prior to today’s update as I just did what I thought was a routine upgrade. As I said in step 2, I compiled from source (as always). If you can advise the correct commands to checkout the last commit prior to the suspected change, I am happy to give it go. Usually when stuff like this happens its a busted sensor or something similar. It not long since I replaced that homing sensor so I was pretty methodical confirming the hardware was OK before raising the issue. I thought my next step was to roll back to V2.8

#3 – andypugh 于 2019-08-10

Rod. Not at a computer today.
But google for “git bisect” instructions.

#4 – rodw-au 于 2019-08-10

Andy, thanks I’m a it confused now. When I look at commits, I see that commit 5ea573a shows as being committed yesterday but it was committed 22 days ago. So does that mean it was changed yesterday? I won’t be near this machine until Monday.

#5 – andypugh 于 2019-08-10

I think that 5ea573a was probably committed to 2.8 (or 2.7) first and has only just been merged to 2.9.

#6 – rodw-au 于 2019-08-13

This can be closed. Now working. I don’t know if it was a later commit or some config error in my repository.

原始Issue: https://github.com/LinuxCNC/linuxcnc/issues/632

Here are the steps I follow to reproduce the issue:

This is what I expected to happen:

This is what happened instead:

It worked properly before this:

Information about my hardware and software:

评论 (6)