#### logs Catalogue

days :1663
Table logs
date_time
user
message
2017-05-30 02:26:27
(luzpaz) http://phoronix.com/scan.php?page=news_item&px=Qt-5.9-RC2-Released
2017-05-30 03:04:53
phryk
"FreeCAD is still in the early stages of development" Is this still true, 14 years after the initial release?
2017-05-30 03:44:25
(kkremitzki) Well, the planned scope is pretty huge so in some sense I would say that's true... maybe a later stage of early
2017-05-30 05:03:13
dioz
yo
2017-05-30 05:03:27
dioz
any of you cnc guys know what the constant 3.82 is?
2017-05-30 05:03:39
dioz
ie. i wanna determine spindle speed
2017-05-30 05:04:07
dioz
so i 3.82 * 80(mild steel) /( 21/64)
2017-05-30 05:04:18
dioz
what is that constant?
2017-05-30 05:05:07
dioz
21/64 would be my drill size for a 3/8 fine thead
2017-05-30 05:05:17
dioz
tap size
2017-05-30 05:05:38
dioz
2017-05-30 10:16:38
(luzpaz) @sliptonic ^
2017-05-30 14:24:51
(kkremitzki) Testing $$\begin{bmatrix} 1 & 0 & 0 \\ 0 & -1 & 0 \\ 0 & 0 & -1 \end{bmatrix}$$
2017-05-30 14:31:45
(luzpaz) $$\begin{bmatrix} 1 & 0 & 0 \\ 0 & 0 & 0 \\ 0 & 0 & -1 \end{bmatrix}$$
2017-05-30 14:37:26
(luzpaz) I fear we may need to re-do the source integration database since it malfunctioning often. I put a lot of work associating fixes to tickets. Alas it may have all been in vain
2017-05-30 14:54:52
(kkremitzki) Hopefully not, that should be a last resort
2017-05-30 15:18:13
sliptonic
dioz: It's not really a constant. It converts the SFM (surface speed of the material) into inches at the same time it converts the diameter of the tool into the circumference.
2017-05-30 15:19:16
sliptonic
SFM is Surface FEET per minute so needs to be converted to inches.
2017-05-30 21:06:52
(abdullahtahiriyo) [[1 0 0] [0 1 0] [0 0 1]]
2017-05-30 21:07:05
(abdullahtahiriyo) which syntax does it use?
2017-05-30 21:07:27
(kkremitzki) LaTeX, you enclose it in double dollar signs
2017-05-30 21:09:17
(abdullahtahiriyo) Ok, so it supports LaTeX syntax. That is cool...
2017-05-30 21:10:36
(abdullahtahiriyo)
2017-05-30 21:10:36
(abdullahtahiriyo) R^2 =
2017-05-30 21:10:36
(abdullahtahiriyo) \left({\begin{array}{cc} c & s \end{array}}\right)
2017-05-30 21:10:36
2017-05-30 21:10:48
(kkremitzki) You need to wrap it in 
2017-05-30 21:10:54
(kkremitzki) $$2017-05-30 21:11:52 FreeCADBot2 (abdullahtahiriyo)$$
2017-05-30 21:11:52
(abdullahtahiriyo) \begin{matrix}
2017-05-30 21:11:52
(abdullahtahiriyo) a & b \\
2017-05-30 21:11:52
2017-05-30 21:12:07
(abdullahtahiriyo) $$2017-05-30 21:12:07 FreeCADBot2 (abdullahtahiriyo) R^2 = 2017-05-30 21:12:08 FreeCADBot2 (abdullahtahiriyo) \left({\begin{array}{cc} c & s \end{array}}\right) 2017-05-30 21:12:08 FreeCADBot2 (abdullahtahiriyo) [full message: https://gitter.im/FreeCAD/FreeCAD?at=592dd250fa63ba2f76966062] 2017-05-30 21:13:13 FreeCADBot2 (abdullahtahiriyo)$$ R^2 =
2017-05-30 21:13:13
(abdullahtahiriyo) \left({\begin{array}{cc} c & s \end{array}}\right)
2017-05-30 21:13:13
(abdullahtahiriyo) \left(\begin{array}{cc} 1 & 0\ 0 & 1 \end{array}\right)
2017-05-30 21:13:13
2017-05-30 21:13:24
(abdullahtahiriyo) It looks nice!
2017-05-30 21:13:27
(abdullahtahiriyo) Thanks!
2017-05-30 21:13:55
(kkremitzki) Indeed, it's pretty nice to have LaTeX support in a chat
2017-05-30 21:15:08
(abdullahtahiriyo) \begin{alignedat}{3} 2017-05-30 21:15:08 FreeCADBot2 (abdullahtahiriyo) & \hskip8em &\diffp{\mathcal D}{t}& = â Ã \mathcal H & \hskip8em &\text{\rlap{(Loi de Faraday)}} \\[0.5ex] 2017-05-30 21:15:08 FreeCADBot2 (abdullahtahiriyo) & & \diffp{\mathcal B}{t}& = -â Ã \mathcal E & &\text{\rlap{(Loi d'AmpÃ¨re)}} \\[0.5ex] 2017-05-30 21:15:08 FreeCADBot2 (abdullahtahiriyo) [full message: https://gitter.im/FreeCAD/FreeCAD?at=592dd3050ba4d59763f7049b] 2017-05-30 21:15:27 FreeCADBot2 (AjinkyaDahale) are the KaTeX parse errors intentional? 2017-05-30 21:15:56 FreeCADBot2 (abdullahtahiriyo) It is me trying to push the boundaries... 2017-05-30 21:16:56 FreeCADBot2 (abdullahtahiriyo)\begin{aligned}
2017-05-30 21:16:57
2017-05-30 21:16:57
2017-05-30 21:16:57
2017-05-30 21:17:35
(kkremitzki) There's a preview tool available at https://khan.github.io/KaTeX/
2017-05-30 21:19:36
(abdullahtahiriyo) That is indeed a better way...
2017-05-30 21:50:16
XXCoder
as usual clients that add protocol makes a mess for people using proper protocol lool
2017-05-31 00:19:26
2017-05-31 01:28:21
2017-05-31 01:37:45
(sgrogan) @bblacey , if your bck from vacation can you look at this https://forum.freecadweb.org/viewtopic.php?f=4&t=17501&start=250#p175243 I think on OSX your probably detecting it from OCCT cmake but I think @abdullahtahiriyo has detected a FreeCAD flaw.
2017-05-31 01:38:30
(sgrogan) Also @vejmarie are you out theere?
2017-05-31 01:41:50
(bblacey) @sgrogan - sure, Iâll take a look. Also, Iâde be happy to add the AppImage âdeploy to Githubâ logic to Travis if you and @triplus want some help with that as well.
2017-05-31 01:42:57
(sgrogan) Well that would certainly be cool. I think but am not sure that Travis is building it sucessfully.
2017-05-31 01:44:05
(bblacey) From the logs - it looks like it is: +ls -lh ../out/FreeCAD-0.17.git201705252304.glibc2.17-x86_64.AppImage
2017-05-31 01:44:05
(bblacey)
2017-05-31 01:44:05
(bblacey) -rwxr-xr-x 1 travis travis 212M May 30 16:51 ../out/FreeCAD-0.17.git201705252304.glibc2.17-x86_64.AppImage
2017-05-31 01:44:05
(bblacey) 
2017-05-31 01:45:37
(bblacey) But we build with both clang and gcc on linux so we should modify to only deploy on one, clang recommended. It should also respect the DEPLOY flag and adhere to the naming convention, all of which should be easy to do.
2017-05-31 01:47:26
(sgrogan) That's the one I uploaded today. The scipt deploys from the Ubuntu PPA trusty debs. So gcc. We don't need to re-package. An arguement could be made for a more immediate package, but the PPA is the best supported on Linux.
2017-05-31 01:47:57
(bblacey) For TBB, not sure I am following the problem/issue totally but you have to pass -DUSE_TBB:BOOL=ON to the OpenCascade cmake.
2017-05-31 01:50:54
(sgrogan) OCCT is being built with TBB, but FreeCAD isn't using it due to the the lack of the compiler definition. If I understand. @abdullahtahiriyo has been giving a tutorial. Even if FreeCAD is passed the _DHAVE_TBB=ON is passed.
2017-05-31 01:54:50
(sgrogan) Or is it a problem with the naming? You say _DUSE_TBB, I say DHAVE_TBB. I will check.
2017-05-31 01:56:14
(bblacey) 
2017-05-31 01:56:15
(bblacey) USE_TBB Boolean flag Indicates whether TBB 3rd party is used or not. TBB stands for Threading Building Blocks, the technology of Intel Corp, which comes with different mechanisms and patterns for injecting parallelism into your application. OCCT remains parallel even without TBB product
2017-05-31 01:56:15
(bblacey) 
2017-05-31 01:56:33
2017-05-31 02:19:18
(bblacey) @sgrogan - it does look like OCCT uses HAVE_TBB internally. For example, from
2017-05-31 02:19:18
(bblacey) 
2017-05-31 02:19:31
(sgrogan) In OCCT 7.1 I see this "set (OpenCASCADE_WITH_TBB ON)" in OpenCASCADEConfigVersion.cmake
2017-05-31 02:22:21
(sgrogan) Ahh, these are the headers @vejmarie uses, so he has probably added to them and they are from OCCT, these are probably the upstream changes he's working on. I need to find his OCCT branch.
2017-05-31 02:26:19
(bblacey) Here? https://github.com/vejmarie?tab=repositories
2017-05-31 02:34:22
(sgrogan) somewhere here: https://github.com/vejmarie/occt-7/tree/parallel_step
2017-05-31 02:55:51
(sgrogan) Need to learn how this works [code}#ifdef HAVE_TBB
2017-05-31 02:55:51
(sgrogan) using namespace tbb;
2017-05-31 02:55:51
2017-05-31 06:27:56
(vejmarie) Yes i am back
2017-05-31 09:27:09
2017-05-31 12:41:14
(luzpaz) @sasobadovinac thx but i cant view it on this iPad. It returns: WebGL implementation must support at least 16 texture units
2017-05-31 12:46:14
(luzpaz) WB @vejmarie :+1:
2017-05-31 12:46:31
(vejmarie) thanks
2017-05-31 13:12:23
(luzpaz) Look at the 2nd result when throwing this in to google: qt5 touchpad pinch osx
2017-05-31 13:18:52
AndreeeCZ
hello all!
2017-05-31 13:19:05
AndreeeCZ
is there a way to produce a 'cut' or 'profile' view?
2017-05-31 13:19:26
AndreeeCZ
say you have an engine composed of various stuff and you want to see inside, how it all fits together
2017-05-31 13:21:28
AndreeeCZ
ok i found out its called 'section view'
2017-05-31 13:22:44
AndreeeCZ
is arch SectionPlane the thing?
2017-05-31 13:24:15
AndreeeCZ
http://www.khulsey.com/portfolioimages/engine-section-view-v6.jpg
2017-05-31 15:22:05
Almisuifre
Salutations \o/
2017-05-31 15:23:24
Amritpal
yorik, hi
2017-05-31 15:23:28
(luzpaz) Greets
2017-05-31 15:24:49
yorik
hi Amritpal
2017-05-31 15:25:38
Amritpal
yorik, Is there any work still pending which will be helpful in my GSoC project?
2017-05-31 15:26:59
yorik
Amritpal: from my point of view no. The task I gave you was actually more to get you fully "at home" in the internal working of the Rebar object
2017-05-31 15:27:12
yorik
so you will be able to easily modify it if needed
2017-05-31 15:27:26
(abdullahtahiriyo) @bblacey @sgrogan There seems not to be a problem at building OCCT. There seems to be a problem at building FC. The problem I think I have identified is that the OCCT include files that are used in FC use the HAVE_TBB preprocesor macro (not a cmake environment FLAG). However, the HAVE_TBB preprocesor macro is not set. There should be a way in FC's cmake to "read" from the cmake information generated by the OCCT build whether OCCT
2017-05-31 15:27:26
TBB or not, so as to set the HAVE_TBB preprocesor macro or not, so that the headers that are included in FC and the symbols that will be linked during compilation match.
2017-05-31 15:27:35
yorik
I think you did very well :) so I think you can start on your own project
2017-05-31 15:28:38
yorik
Amritpal: I would be very glad if you can pot as much as possible on the FreeCAD forum too, to let people follow the progresses
2017-05-31 15:29:06
yorik
because not many of them will go look at the BRLCAD dev log
2017-05-31 15:29:29
(abdullahtahiriyo) I do not know much of Cmake, but it would appear to me that the resulting cmake files from the OCCT compilation that are installed under the cmake library are the key. However, I have no idea how to bring those into play when compiling FreeCAD. In other words, what should be added to FC's cmake, so that this information is retrieved and the HAVE_TBB preprocesor macro set (or not) in accordance.
2017-05-31 15:29:33
(luzpaz) Yea, we love progress reports ;D
2017-05-31 15:29:43
yorik
also maybe we could setup a kind of weekly chat session with Bernd too
2017-05-31 15:31:04
Amritpal
yorik, okay I will put my all dev log on FreeCAD forum
2017-05-31 15:31:44
(luzpaz) @berndhahnebach ping
2017-05-31 15:31:54
Amritpal
yorik, by the way, I have posted the work which I have done in community bonding period.
2017-05-31 15:32:16
Amritpal
2017-05-31 15:33:42
Amritpal
yorik, I also have noticed one more thing in the FreeCAD like in the structural element having reinforcement, when I change the placement of the structural element its reinforcement is not moved with respect to the structural element. Is it a bug or I am missing something?
2017-05-31 15:34:48
(luzpaz) @abdullahtahiriyo thanks for your fantastic work on the bugtracker. Very much appreciated!
2017-05-31 15:37:01
yorik
Amritpal: No you are right. It's a problem we need to address.
2017-05-31 15:38:31
yorik
When its placement changes, the structure object should propagate the changes to all its included rebars
2017-05-31 15:40:18
(abdullahtahiriyo) @luzpaz I just fix bugs. You do the fantastic work keeping everything updated and pinging one and the other. I have said several times that FC is awesome because of the diversity of people and capabilities and the ability to work together. Let's make PDN roll ;)
2017-05-31 15:42:06
Amritpal
yorik, Yes, we need to add the placement of structural element to the rebar object placement. So, by doing this thing placement of the rebar object will also change w.r.t to structural element.
2017-05-31 15:42:15
yorik
@luzpaz @abdullahtahiriyo Agree! I am a bit behind in addressing all your pings luzpaz, but please continue! The tracker is really getting in much better shape
2017-05-31 15:45:27
yorik
Amritpal: yes that can be done in two locations: Either it is the structure that triggers a change to its rebars, or it's the rebar that triggers its own change when its host structure has changed.
2017-05-31 15:46:51
yorik
Amritpal: The thing is, we need to compare the struct placement before it is changed, with after it has changed, to get a "difference", that we can add to the rebar's placement
2017-05-31 15:49:02
Amritpal
yorik, I got it.:)
2017-05-31 15:49:11
(luzpaz) Yorik @abdullahtahiriyo thanks guys. I agree about diversity. Thanks for letting me know that the pings are not too disruptive. :+1:
2017-05-31 15:50:39
(abdullahtahiriyo) @luzpaz They are disruptive. They need be disruptive otherwise we would be happily thinking about that new feature and completely neglecting the bugs ;)
2017-05-31 15:51:32
yorik
Amritpal: I think I remember there was some beforePropertyChange() that could be used in Python Features... I'll have a look
2017-05-31 15:54:03
Amritpal
yorik, okay.
2017-05-31 15:55:46
(luzpaz) @sgrogan oinged you about https://github.com/UWP-Open-Source-Community/Documentation/issues/5#issuecomment-305209285
2017-05-31 15:58:46
(luzpaz) @abdullahtahiriyo LOL true
2017-05-31 16:13:43
(luzpaz) @yorikvanhavre ive been thinking, having a webhost that has shell access would only make things easier to make maintenance changes. And it would not add to the complexity of maintenance which i think is what folks are concerned about.
2017-05-31 16:16:17
(yorikvanhavre) @luzpaz the thing is, the webhost is handled by JÃ¼rgen, and I don't think he will want to change. Also I'm not sure I fully agree with you about "it would not add to the complexity of maintenance" ;)
2017-05-31 16:16:47
(luzpaz) Right now we have an issue with the source integration mantisbt plugin not working consistently because the initial import didnt go well. Kurt is troubleshooting it all via FTP and its difficult
2017-05-31 16:17:03
(yorikvanhavre) I know
2017-05-31 16:17:16
(yorikvanhavre) wiki maintenance is also annoying without phpbb
2017-05-31 16:17:27
(yorikvanhavre) without shell access i mean
2017-05-31 16:18:46
(luzpaz) it can get complex if we start doing more complex things voia the shell. But to read logs and execute commands all via ftp... thats prohibitive and time consuming. I understand though that there is a dependency jurgen
2017-05-31 16:20:34
(luzpaz) jurgen isnt involved though... so whats the problem with just finding a host that gives us more access ? (Beside the concern of complexity)
2017-05-31 16:21:23
(luzpaz) I sense there is a political issue
2017-05-31 16:22:50
(yorikvanhavre) well, you convince him, you convince me :)
2017-05-31 16:25:55
(luzpaz) @yorikvanhavre Sure, im willing to try :smile_cat: : How do i access him ?
2017-05-31 16:26:32
(yorikvanhavre) well I'm afraid the only way I know is sending him a PM on the forum...
2017-05-31 16:27:44
(kkremitzki) It would be nice to have shell access to the server but I think there are some intermediate steps that need to be taken before considering moving hosting providers
2017-05-31 16:28:37
(kkremitzki) Right now the FTP directories that make up freecadweb.org are in sort of a "unicorn server" state where everything working is highly dependent on, e.g. one-off fixes, configuration, etc., we saw issues related to this with the forum upgrade
2017-05-31 16:29:23
(kkremitzki) Now that all the apps that make up freecadweb are upgraded, the next stage is to make deployment repeatable, so that we could actually execute a move of hosting providers without a lot of pain
2017-05-31 16:30:01
(kkremitzki) A benefit to this is that doing it correctly means anyone who wants to contribute to the infrastructure can deploy a staging server with mostly empty test/stub databases and do things like fiddle with themes and plugins
2017-05-31 16:31:11
(kkremitzki) Even if we got a new hosting provider tomorrow, the move itself couldn't be executed without the risk of big downtime or unexpected side-effects
2017-05-31 16:31:36
(luzpaz) @kkremitzki awesome. How far are we from repeatable deployment?
2017-05-31 16:32:16
(yorikvanhavre) kkremitzki: to be totally honest with you, the more you go in that direction, the less I feel able to help and all the more all the weight of web maintenance is shifting to your shoulders only...
2017-05-31 16:32:19
(kkremitzki) BTW work on Buildbot continues, I'm testing with https://build.freecad.io right now, and also I found out the Blender project uses Buildbot as well and they have a really nice setup where you can see how much FreeCAD could do with Buildbot as well: https://builder.blender.org/
2017-05-31 16:33:13
(yorikvanhavre) I might be a bit of an old-fashion obsolete web person, though :D
2017-05-31 16:33:47
(kkremitzki) @luzpaz Well one big thing is cleaning up the backups to where we have sample DBs without, e.g. private messages, emails, etc, in them. Another thing is making a choice of technology for it all
2017-05-31 16:34:21
(kkremitzki) @yorikvanhavre I understand what you mean, but if it's done correctly, you should pretty much be able to do everything the same (i.e. just deploying with git-ftp could still work)
2017-05-31 16:34:43
(kkremitzki) another thing with that is that we would need infrastructure admin documentation. I spent a little time testing namespaces on the wiki but haven't been able to get them working
2017-05-31 16:35:04
(luzpaz) @yorikvanhavre last time jurgen interacted on the forum: Mon Oct 26, 2015 12:55 am
2017-05-31 16:35:36
(yorikvanhavre) @luzpaz: yeah i know :(
2017-05-31 16:36:00
(luzpaz) So what if he doesnt respond?
2017-05-31 16:36:05
(kkremitzki) I added a "Infrastructure" namespace on the wiki but it doesn't seem to be behaving correctly
2017-05-31 16:36:28
2017-05-31 19:24:26
(sasobadovinac) Does anyone have any knowledge on the books available in this Humble Bundle https://www.humblebundle.com/books/linux-book-bundle ?
2017-05-31 20:12:13
(sasobadovinac) @bblacey Don't you think it would make sense to use Travis Cron Jobs for AppImage builds if we do them on Travis?
2017-05-31 20:13:54
(sasobadovinac) They can be detected with "TRAVIS_EVENT_TYPE: cron", IMO running them weekly would be good enough
2017-05-31 20:14:06
2017-05-31 20:17:35
(sasobadovinac) @sgrogan What would be the commands in the terminal if you would build AppImage from PPA on a clean machine?
2017-05-31 20:37:35
(bblacey) @sasobadovinac - sounds reasonable. Iâm just offering to lend a helping hand so feel free to weigh in on the optimal deployment frequency here - https://forum.freecadweb.org/viewtopic.php?f=4&t=22606&p=176382#p176382
2017-05-31 20:45:17
(sasobadovinac) @bblacey It just sounds like it should not be that hard to do, but I did not try to make it yet :worried:
2017-05-31 21:04:50
(sgrogan) @sasa @bblacey We are already building the AppImage on Travis, look at the last lines of Travis.yml for the commands. We are not yet deploying. I think ideally we would build it if there was a PPA update from the last build, but a cron job run weekly is a good step forward.
2017-05-31 21:16:24
(bblacey) @sgrogan - I saw @triplusâ work to build the AppImage and will add the deploy logic that re-uses the travis helpers we built for the OS X deployments (very straight-forward). I will also only deploy on a cronjob as @saso recommended and schedule one cron per week unless others weigh in for a higher frequency on the forum. Sound right?
2017-05-31 21:18:50
(sgrogan) Sounds great I was deploying about every 2-3 weeks before the script broke. If we can prune the old ones that would also be cool. Right now there is an old one I've left before the PPA switched to OCCT 7.1 and broke netgen.
2017-05-31 21:19:51
(bblacey) the travis-helper scripts include a deploy. unless you want something different, I will set the prune to 1 - the same as the OS X builds.
2017-05-31 21:21:28
(sgrogan) If possible I like to leave the oldest one always, for those that need netgen. Same as I'm doing with Win. Would prune 2 do that?
2017-05-31 21:21:29
(sasobadovinac) cron jobs can be set to run monthly, weekly or daily
2017-05-31 21:21:44
(sasobadovinac) weekly seems ok to me for appimage
2017-05-31 21:22:48
(sasobadovinac) i was however thinking you would prefer to not make them from new travis builds but from ppa?
2017-05-31 21:22:54
(bblacey) prune keeps the last # builds but there is an âoptionalâ string search so we will just use a different name for the auto builds.
2017-05-31 21:25:51
(bblacey) @sasobadovinac - actually, doing it from PPA is a pretty good ideaâ¦ From a maintenance stand-point, it might be good to build and deploy from a single place. Once we have our own infrastructure, I imagine we will centralize so I can go either way. Personally, Iâm not familiar with the PPA infrastructure deploying to GitHub so maybe you could take that one if that is the direction we choose to go?
2017-05-31 21:26:46
(sasobadovinac) i was thinking it could also be faster and maybe less depended to build fails
2017-05-31 21:26:57
(sgrogan) @bblacey great. @sasso I have mixed feelings. If we use the PPA they align with Ubuntu so great for consistency. If we built from the Travis assets we could do it on every push. We'd need to add the PPA dependecies to Travis (like occt7.1 and eigen3) doing so would restrict the matrix we have now. Travis uses the oce from the trusty repo that's old. Better for CI I think.
2017-05-31 21:28:08
(sgrogan) Oh and I don't know how to build the AppImage without the existing debs
2017-05-31 21:29:22
(sasobadovinac) oooh, ok i misunderstood, i was thinking it would be just like (pseudo): if cron; install from PPA; build appimage; deploy
2017-05-31 21:29:43
(sasobadovinac) but if it is more complex then doing as we do now on travis is better
2017-05-31 21:31:19
2017-05-31 21:31:53
(sgrogan) No from the PPA that's what it is. We can build from Travis directly without the PPA. I think this means (psuedo) Build FreeCAD on Travis, create .debs, build AppImage, deploy
2017-05-31 21:33:24
(bblacey) Hmmm, I was just going to add the deploy logic for the app image that is already built on Travis - @triplusâ work with the aforementioned cron control/swtiches.
2017-05-31 21:34:46
(sgrogan) @bblacey This is all we need.:)
2017-05-31 21:38:40
(sgrogan) There is someone working on this on the OpenSuse build system too. https://forum.freecadweb.org/viewtopic.php?f=4&p=176382#p176272 The deployment form Travis is nice but I think providing the best CI for wmayer and yorik is most important. That's the main reason I don't deploy win from Appveyor until we have a working cache. To build locally for me it's just a double click.
2017-05-31 21:45:44
2017-05-31 21:46:05
(sasobadovinac) @sgrogan what do you mean by working cache?
2017-05-31 21:47:49
(sasobadovinac) IMO it never hurts to have some working backup CI methods as long as they don't represent to much trouble for us to create them
2017-05-31 21:48:46
(sgrogan) We don't use caching on Appveyor. I think keeping the build time short for the linux developers is more important than deploying from Appveyor and taking more time. Especially because it's so easy for me to do it manually. I can also look at self test failures and decide to deploy or not.
2017-05-31 21:50:44
2017-05-31 21:51:58
(sasobadovinac) crone jobs on travis run totally independent from PR / merge builds, it is why IMO they are a good fit for such things
2017-05-31 21:52:11
2017-05-31 21:53:29
(sasobadovinac) if a concern and if possible we can save that script in our git
2017-05-31 21:54:14
(sasobadovinac) as i understand it is now at appimage github?
2017-05-31 21:54:47
(sasobadovinac) i guess if they get hacked there will be bigger problems then just our deploy :smile:
2017-05-31 21:55:26
(sgrogan) I had a similar discussion with @triplus. the AppImage stuff all is on github, so it they could hack that they could hack us.
2017-05-31 21:55:36
(sasobadovinac) another way to do it is to check the yaml hash
2017-05-31 21:56:08
(sasobadovinac) but we will have to update the known hash if yaml gets updated
2017-05-31 21:57:18
(sgrogan) That would require more maintenance, recently the script broke and triplus made a pull request. You were faster to the same conclusion. I think it's OK I just value your opinion on the security stuff.
2017-05-31 22:00:10
(sasobadovinac) well the maintenance would give us the control over it
2017-05-31 22:03:33
(sgrogan) Would you mind posting your thoughts on the forum. We both seem to be arguing both sides of the same argument :smile:
2017-05-31 22:04:53
(sgrogan) I don't think we need to decide immediately
2017-05-31 22:10:27
(sasobadovinac) yes, i don't think it is an issue
2017-05-31 22:11:34
(sasobadovinac) and we can check how other projects have it setup
2017-05-31 22:12:00
(sasobadovinac) AppImage is getting quite popular
2017-05-31 22:18:07
(sgrogan) > AppImage is getting quite popular
2017-05-31 22:18:07
(sgrogan) I'm quite surprised. As a win guy I'm quite used to packing like this. Last time I counted I have 7 versions of Python on my system. I guess the convenience to overcome dependencies is just to great. From a packaging standpoint it's the best of both worlds. From the PPA we get all the sucurity updates for free and it's cross distro.
2017-05-31 22:21:31
(sgrogan) @luzpaz Thanks for the link. More OpenSource on win is a good thing.
2017-05-31 22:22:17
(sasobadovinac) As for security, personally I am waiting for all the os-es to become more like Qubes OS
2017-06-01 06:11:02
jhol
https://www.freecadweb.org/tracker/view.php?id=3061 - this is a weird one
2017-06-01 11:41:52
(abdullahtahiriyo) @kkremitzki
2017-06-01 11:42:03
2017-06-01 13:32:37
2017-06-01 16:04:18
(luzpaz) Looks like the gitter.im activity panel needs a restart
2017-06-01 16:29:38
2017-06-01 16:42:23
(abdullahtahiriyo) @luzpaz
2017-06-01 16:42:41
(abdullahtahiriyo) Not at all. That is a very interesting crash. I will give it a look
2017-06-01 16:42:51
(abdullahtahiriyo) later on
2017-06-01 16:43:41
2017-06-01 17:28:02
(abdullahtahiriyo) @kkremitzki
2017-06-01 17:28:05
2017-06-01 17:28:42
(kkremitzki) Let me take a look
2017-06-01 18:14:26
(kkremitzki) That change is interesting Abdullah, for my test case that allows mirroring w.r.t base planes but not the vertical/horizontal sketch axes
2017-06-01 18:15:12
(abdullahtahiriyo) Could you post your test (fcstd) case in the forum thread? @kkremitzki
2017-06-01 18:15:14
jhol
I encountered this bug doing some cad work on the daily build over the weekend
2017-06-01 18:15:23
jhol
really cool to see lots of interest in it
2017-06-01 18:15:54
(abdullahtahiriyo) @jhol it is the most annoying thing ever!
2017-06-01 18:16:30
jhol
abdullahtahiriyo: actually this one caused me more annoyance: https://freecadweb.org/tracker/view.php?id=3061#c9233
2017-06-01 18:16:59
jhol
-- though annoyance is the wrong word, because I'm very grateful for all the awesome work going on in FreeCAD
2017-06-01 18:19:21
(abdullahtahiriyo) @kkremitzki
2017-06-01 18:19:32
(kkremitzki) Hey
2017-06-01 18:19:35
(abdullahtahiriyo) I think I have tested your use case
2017-06-01 18:19:45
(abdullahtahiriyo) a sketch rectange
2017-06-01 18:19:49
(abdullahtahiriyo) 1x1
2017-06-01 18:19:54
(abdullahtahiriyo) coincident in the origin
2017-06-01 18:20:02
2017-06-01 18:20:03
(abdullahtahiriyo) mm
2017-06-01 18:20:31
(abdullahtahiriyo) Try trying to drag the origin coincident towards the side where your transformation is failing
2017-06-01 18:21:44
(kkremitzki) The transformation failing/being the incorrect one was actually because I had swapped the vertical/horizontal sketch lines I mentioned
2017-06-01 18:22:07
(kkremitzki) but putting it back to the way it was and mirroring works as expected.. so that's good to know more about where the bug really is
2017-06-01 18:22:38
(kkremitzki) I was kinda getting lost in the weeds so to speak with the quaternions and matrices, haha.. having heard of them but never used them before
2017-06-01 18:24:34
2017-06-01 18:24:55
(abdullahtahiriyo) With your lines swap (which I still think is a fix), it does not fail.
2017-06-01 18:25:16
(kkremitzki) Right, it wasn't failing, I just meant it was doing the incorrect mirroring
2017-06-01 18:25:46
(abdullahtahiriyo) Ah ok, but that is because we need to change the mesh not to invert and to correct X/Y, shouldn't we?
2017-06-01 18:26:45
(abdullahtahiriyo) What I mean, is that the mesh is wrong (two times wrong), not your fix with H/V. I think
2017-06-01 18:27:06
(kkremitzki) Right
2017-06-01 18:28:36
(kkremitzki) I suppose it makes sense for you to look at the sketcher/precision stuff so I could look at the viewprovider issue?
2017-06-01 18:28:45
(abdullahtahiriyo) If you feel like fixing the mesh, I will continue with the checkIntersection. I have been debugging it intermitently all the day long
2017-06-01 18:28:51
(abdullahtahiriyo) Perfect sync
2017-06-01 18:29:02
(abdullahtahiriyo) I will handle sketcher after the checkIntersection.
2017-06-01 20:37:54
(abdullahtahiriyo) @kkremitzki How do I trigger your unit test from FC Console?
2017-06-01 22:07:17
xavier_
Is there a way to make *everything* constrained to some degree?
2017-06-01 22:08:04
xavier_
Unless explicitly set - I keep accidentally moving things that I don't want to and just having that freehand move/rotate tool scares me
2017-06-01 22:18:45
(kkremitzki) @abdullahtahiriyo ./bin/FreeCAD --run-test TestPartDesignApp in build folder
2017-06-01 22:24:22
(abdullahtahiriyo) @kkremitzki ah ok, thanks!
2017-06-01 22:25:47
(kkremitzki) With Part::checkIntersection() returning true I have:
2017-06-01 22:25:48
(kkremitzki) 
2017-06-01 22:25:48
(kkremitzki) FAIL: testMirroredCase (TestPartDesignApp.PartDesignMirroredTestCases)
2017-06-01 22:25:48
2017-06-01 22:28:03
(kkremitzki) So once you have a fix that volume comparison needs to be changed to >1.0.
2017-06-01 22:29:09
(kkremitzki) Really it'd be better if 2.0 worked but I believe the mirrored volume was like 1.99....8, like 2e-16 short of 2.0 or so
2017-06-01 22:29:25
(kkremitzki) If Volume == 2.0 worked that is
2017-06-01 22:39:57
(sgrogan) @kkremitzki , would a valid case be create a cube whose center is offset from the origin by some amount in x. Then mirror about the y-axis. Volume could be anywhere between 1, if the x offset was 0, and 2 if the offset put the edge at the y-axis?
2017-06-01 22:43:24
(kkremitzki) Yeah, that's definitely on a list of test cases for mirroring, although I was gonna do a slightly different scenario with a symmetric cube and check that the volume is ~ 1.0 in one case and ~2.0 in the other
2017-06-01 22:45:15
(kkremitzki) Not sure the easiest way to describe the cube but it'd have one corner at (-0.5, 0, 0) and the opposite corner at (0.5, 1, 1), so a mirror about YZ should give a volume of 1 and a mirror about XZ should be a volume of 2
2017-06-01 22:46:34
(sgrogan) Exactly, thanks for the work on this. The mirror feature is hard to test it's so confusing.
2017-06-01 22:46:59
(kkremitzki) No kidding, not only was it failing, but it was failing to report the failure correctly
2017-06-01 22:47:37
(abdullahtahiriyo) From console:
2017-06-01 22:47:39
(abdullahtahiriyo) import PartDesign
2017-06-01 22:47:51
(abdullahtahiriyo) import unittest
2017-06-01 22:48:06
2017-06-01 22:48:14
(abdullahtahiriyo) unittest.TextTestRunner().run(mysuite)
2017-06-01 22:50:52
(kkremitzki) So I have https://build.freecad.io set up to track pushes to https://github.com/kkremitzki/freecad and attempt a build with public viewing but build control behind a login (it could also be allowed to anyone in FreeCAD Github Org) although as you may see I haven't gotten the build environment set up correctly because the builds fail almost immediately
2017-06-01 22:51:51
(kkremitzki) One thing I wanted to ask about was having build config/recipes in the main repo or a separate one. Because Buildbot allows for it to be separated which might be nice, but the recipes are all in-repo currently. Anyone have any thoughts on this?
2017-06-01 22:52:37
(kkremitzki) Buildbot also allows for in-repo config using .bbtravis.yml files
2017-06-01 22:55:46
(sgrogan) We have this https://github.com/FreeCAD/FreeCAD-ports-cache . @bblacey let me post the libpacks for the AppVeyor builds here. It has different access rights than the main repo?
2017-06-01 23:09:30
(kkremitzki) Yeah, there can be per-repo access rights, e.g. I can merge the https://github.com/freecad/homepage repo. So a future FreeCAD/Buildbot repo could be worked on in a way that doesn't require any of Yorik or Werner's time
2017-06-01 23:10:32
(kkremitzki) Buildbot workers can be ran on Windows VMs (possible currently with the freecad.io servers) and if a Mac OS build server were available, on there too
2017-06-01 23:10:53
2017-06-01 23:10:57
(sgrogan) I don't know about conflicts but maybe the ports cache repo could morph into a packaging repo?
2017-06-01 23:13:00
(sgrogan) What are the linux .tar.bz2's?
2017-06-01 23:18:19
(abdullahtahiriyo) @kkremitzki the test passes even with self.failUnless(self.Mirrored.Shape.Volume < 0.0) or self.failUnless(3 < 1), either there is some kind of cache or it is the demonstration I suck a python. I wanted to know which volume the text thinks it has but no avail. Anyway, I will go back to checkIntersection
2017-06-01 23:28:50
(sgrogan) @kkremitzki is this relevant to the tolerance issues? https://dev.opencascade.org/index.php?q=node/1056
Logs go back to 2013-03-31 18:28:57