FreeCAD Forum

C_h_o_p_i_n wrote: ↑Wed Feb 09, 2022 8:20 am and removing the "one Body" limitation in PD

Yes, the "one Body limitation" was a very important reason to use Link Branch also for me. I asked several times to remove it in the master but got very disappointing responses. This change (which would be very simple) was not accepted because, in the usual meaning, a "body" should be a single connected object. I also suggested to change the "Body" name (for example using "Part" instead) but this suggestion was also rejected because body (in the current implementation) is restricted to be a single solid. It was a sort of catch-22 situation and I had to surrender.

Being so difficult to have accepted such a simple change (which many users ask), there is no hope for the TN merging.

Just some thoughts of the top of my head.

If we create a release window for 0.20. Wouldn't it also be possible to to already start a 0.21 branch and only put the TPN fixes in it at the start. After that merge the 0.20 and other updates.
I looked at https://builder.blender.org/download/daily/ and they provide both images for the beta (3.10) and alpha (3.20) versions. That way people really can look towards a certain integration path and focus.
Another solution could look like: Make it real project and a test&release manager. But that's not really the Open Source way of working.

Make the appimages, .msi's and .dmg's available on a weekly basis. Maybe also make it's easy to switch between versions or provide a separate config directory by default. Like .config/FreeCAD-beta for Linux Then it would be easier to test, compare and also focus on it.

Like a friend of mine once said: "If a project lasts more than 2 years, it's not a project, but a subscription."

wsteffe wrote: ↑Wed Feb 09, 2022 9:56 am

C_h_o_p_i_n wrote: ↑Wed Feb 09, 2022 8:20 am and removing the "single solid per Body" limitation in PD

Yes, the "one Body limitation" was a very important reason to use Link Branch also for me. I asked several times to remove it in the master but got very disappointing responses. This change (which would be very simple) was not accepted because, in the usual meaning, a "body" should be a single connected object. I also suggested to change the "Body" name (for example using "Part" instead) but this suggestion was also rejected because body (in the current implementation) is restricted to be a single solid. It was a sort of catch-22 situation and I had to surrender.

Being so difficult to have accepted such a simple change (which many users ask), there is no hope for the TN merging.

Well - sometimes...the "purity of doctrine" seems to impede development through invisible walls that some people believe in.
In a way, - to me - it's a kind of German thinking in which rules are followed just because someone somehow set them up at some point.

jhaand wrote: ↑Wed Feb 09, 2022 10:18 am I looked at https://builder.blender.org/download/daily/ and they provide both images for the beta (3.10) and alpha (3.20) versions. That way people really can look towards a certain integration path and focus.
Another solution could look like: Make it real project and a test&release manager. But that's not really the Open Source way of working.

Make the appimages, .msi's and .dmg's available on a weekly basis. Maybe also make it's easy to switch between versions or provide a separate config directory by default. Like .config/FreeCAD-beta for Linux Then it would be easier to test, compare and also focus on it.

If you want to participate in creating such a thing check out the latest discussion on https://forum.freecadweb.org/viewtopic.php?f=10&t=52422
We're talking about making experimental builds more available

C_h_o_p_i_n wrote: ↑Wed Feb 09, 2022 1:29 pm Well - sometimes...the "purity of doctrine" seems to impede development through invisible walls that some people believe in.
In a way, - to me - it's a kind of German thinking in which rules are followed just because someone somehow set them up at some point.

Yea... OK...way off and straying in to inappropriate territory. If you want to have a dissenting opinion...open a new thread and x-post to this one. But don't start polluting this one with negativity.

wsteffe wrote: ↑Wed Feb 09, 2022 9:56 am

C_h_o_p_i_n wrote: ↑Wed Feb 09, 2022 8:20 am and removing the "one Body" limitation in PD

Yes, the "one Body limitation"

I won't argue one way or the other. But if you are going to debate this topic please use the correct terminology.

There is no "one Body" limitation. There is a "single solid" limitation per Body.

Gene

wsteffe wrote: ↑Wed Feb 09, 2022 9:56 am Yes, the "one Body limitation" was a very important reason to use Link Branch also for me. I asked several times to remove it in the master but got very disappointing responses. This change (which would be very simple) was not accepted because, in the usual meaning, a "body" should be a single connected object. I also suggested to change the "Body" name (for example using "Part" instead) but this suggestion was also rejected because body (in the current implementation) is restricted to be a single solid. It was a sort of catch-22 situation and I had to surrender.

Being so difficult to have accepted such a simple change (which many users ask), there is no hope for the TN merging.

Did you make a PR or asked someone else to work on something they don't agree with for free on their spare time? Huge difference... Plus, it's an easy change acording to you, we'll have to see if that is really the case. Note that i do agree that the single body rule is stupid and should be removed, but whatch your tone.

C_h_o_p_i_n wrote: ↑Wed Feb 09, 2022 8:20 am As time goes by, more and more user will use Link branch because the dont want to deal with the hassle, master provides by the TNP.
Just because there is no real cause to not use a branch where TNP is no such big deal any more and removing the "one Body" limitation in PD.

I have to agree here, eventually a hard fork will happen if this is not merged into master, and people will most likely choose the version with better features, we know which one that is...

wsteffe wrote: ↑Wed Feb 09, 2022 6:35 am I have understood that the main developers (those which have the last say) simply do not want to merge it. It is a perfectly legitimate decision, but please do not fool around.

The users should know that this PR will never go into master so that they may decide to use the RT branch if they want a solution (not perfect one but anyway a solution) to the TN problem.

Not the case, but I would appreciate some actual communication from wmayer who is the main guy everyone is waiting for, I dislike the supposed rumours of behind the scenes work to move the TN merge forward, it does not make sense to work on FOSS code review behind the scenes, if it's true then please change this behaviour and if it's false... please be sincere I'd much rather have a clear statement in the lines of: "I currently do not have the time and/or motivation to review such a large and complex PR, this is on hold" (which is perfectly understandable) than to be left dangling not knowing what the hell is going on or if anything is happening at all.

wmayer wrote:

Sorry to bother you but I have to ask. Is it realistic to hope this will get worked on in the foreseable future?

Kunda1 wrote: ↑Wed Feb 09, 2022 2:49 pm If you want to participate in creating such a thing check out the latest discussion on https://forum.freecadweb.org/viewtopic.php?f=10&t=52422
We're talking about making experimental builds more available

Thanks, I was already searching for info about getting such an environment in place. Only I couldn't find the link.

I will check it out.

adrianinsaval wrote: ↑Wed Feb 09, 2022 4:55 pm Plus, it's an easy change acording to you

Not only according to me. In the post https://forum.freecadweb.org/viewtopic. ... 40#p484140 davidosterberg wrote:

If you comment out those lines, then PartDesign will handle multiple bodies. I have tried it many times for debugging purposes. The limitation is a feature. I.e somebody decided it was appropriate and enforced it. It is not due to implementation difficulties. Because the underlying OCC library does not care.

So you may not say that I am asking a big work for free. It is just a matter of uncommenting a few lines.
The problem is not about work but about somebody who doesn't understand that this rule is silly (as you said) and that it should be removed.

wsteffe wrote: ↑Wed Feb 09, 2022 5:45 pm Not only according to me. In the post https://forum.freecadweb.org/viewtopic. ... 40#p484140 davidosterberg wrote:

If you comment out those lines, then PartDesign will handle multiple bodies. I have tried it many times for debugging purposes. The limitation is a feature. I.e somebody decided it was appropriate and enforced it. It is not due to implementation difficulties. Because the underlying OCC library does not care.

So you may not say that I am asking a big work for free. It is just a matter of uncommenting a few lines.
The problem is not about work but about somebody who doesn't understand that this rule is silly (as you said) and that it should be removed.

It may be just that or there might be more to it, it has to be fully checked carefully. I think most of the code in Part Design assumes it's always working on a single solid, it's likely necessary to change other things besides the checks, I remember when I tried to develop a feature for PD that there were some lines that did things that necesarily resulted in single solids. Go ahead and make the change and thoroughly test every single PD feature in their many variations and submit a PR if it's so easy to do... notice that david wrote that comment but he didn't attempt to make a PR either. Even if a change is trivial, these people are volunteers, you can't complain because they didn't do something you asked them too, everybody works on what peaks their interest.