[LDCad] Part Snapping missing/errors

[Roland Melkert]

I was thinking of something like the SLIDE property, where it snaps in one place but can then freely rotate.

You could fake that by using a square or axle SNAP_CYL pair, to force the 'teeth' inline with he nudges inside the hub connector (those connections are forced at 90 deg rotation).

The snap meta on 2907 will still need a 22.5 deg rotate to force the offset as the current official part doesn't align at rest.

[N. W. Perry]

I can't tell if this is a bug or a feature, but I noticed some studless Technic beams were missing their axlehole snaps. I figured out that this was because they use axlehole prims with unofficial updated versions (which I have set to be first in my search paths).

I added the appropriate snap_include to the unofficial primitive file, and now when I edit the part itself the snap info is there, but in the actual model (and part bin) it is not there. How do I get the snap info from unofficial prims to appear in the model?

[Max Martin Richter]

You could fake that by using a square or axle SNAP_CYL pair, to force the 'teeth' inline with he nudges inside the hub connector (those connections are forced at 90 deg rotation).

The snap meta on 2907 will still need a 22.5 deg rotate to force the offset as the current official part doesn't align at rest.

Please remember, that the official 2907 has an error with the teeth and the axle orientation. This (wrong) correction erroneously jumped through the review process and got official. As the fixed part just needs an admin cert it will surely included in the next parts update.

/Max

[Roland Melkert]

I can't tell if this is a bug or a feature, but I noticed some studless Technic beams were missing their axlehole snaps. I figured out that this was because they use axlehole prims with unofficial updated versions (which I have set to be first in my search paths).

I added the appropriate snap_include to the unofficial primitive file, and now when I edit the part itself the snap info is there, but in the actual model (and part bin) it is not there. How do I get the snap info from unofficial prims to appear in the model?

Which part number?

The bin and editor use the exact same data, they might render differently though (for snap editing purposes)

[N. W. Perry]

Which part number?

The bin and editor use the exact same data, they might render differently though (for snap editing purposes)

Part number is axlehol4.dat.

I copied this snap info into the unofficial primitive:
0 !LDCAD SNAP_INCL [ref=axlehole.dat]

It shows up fine in the parts that reference it, if opened in their own session:
   
And if I view the dependencies of those parts, it shows that they are indeed referencing the unofficial version of axlehol4.
But in the model session, the snap is missing:
   

[Philippe Hurbain]

Edited...
42862s01 is missing connectivity: Forget this one, my shadow was not up to date!

41862 tubes of vertical plate are backwards. Connectivity should be:

Code:

0 !LDCAD SNAP_CYL [gender=F] [caps=one] [secs=S 6 4] [pos=0 -10 4] [ori=-1 0 0 0 0 1 0 1 0] [grid=C 2 1 20 0]
0 !LDCAD SNAP_CYL [gender=F] [caps=one] [secs=R 6 4] [pos=0 8 0] [grid=C 2 C 2 20 20]

[Roland Melkert]

Forget this one, my shadow was not up to date!

I haven't updated this threads main version for some time, I was hoping to complete the 20-01 library first and include it in 1.7 Alpha 2, but haven't found the time for it yet

[Philippe Hurbain]

6587 (axle 3 with stud) snapping currently uses an axle meta with round section to make the stud. But this is wrong since the axle meta forces orientation of plate/tile placed on stud (moreover the hole in stud is not modeled).
Correction:

Code:

0 !LDCAD SNAP_CYL [gender=M] [caps=none] [secs=A 6 58   R 8 2] [center=true] [slide=true] [ori=0 -1 0 1 0 0 0 0 1]
0 !LDCAD MIRROR_INFO [baseFlip=Z] [counterPart=self]

instead of

Code:

0 !LDCAD SNAP_CLEAR
0 !LDCAD SNAP_CYL [gender=M] [caps=none] [secs=A 6 58   R 8 2   R 6 4] [center=true] [slide=true] [pos=2 0 0] [ori=0 -1 0 1 0 0 0 0 1]
0 !LDCAD MIRROR_INFO [baseFlip=Z] [counterPart=self]

[Philippe Hurbain]

Not really a shadow problem, but the stud of the telescope (64644) is next to impossible to insert into tube of plate 1x1 round (6141). Most of the time, the telescope jumps to connect its tube to plate stud. Moving the plate instead doesn't help. It works much better with a regular plate 1x1 (3024)!

[David Manley]

For the part "Technic Sprocket Wheel 25.4" (57520), the following places snap points for each of the sprocket notches. That in turn makes placing the chain tread (57518) somewhat less fraught.


0 !LDCAD SNAP_CLP [radius=4] [length=10] [center=true] [slide=true] [pos=0 -30 0] [ori=-1 0 0 0 0 1 0 -1 0]
0 !LDCAD SNAP_CLP [radius=4] [length=10] [center=true] [slide=true] [pos=0 30 0] [ori=1 0 0 0 0 -1 0 -1 0]
0 !LDCAD SNAP_CLP [radius=4] [length=10] [center=true] [slide=true] [pos=-25.980762 -15 0] [ori=-1 0 0 0 0 -1 0 -1 0]
0 !LDCAD SNAP_CLP [radius=4] [length=10] [center=true] [slide=true] [pos=-25.980762 15 0] [ori=1 0 0 0 0 -1 0 -1 0]
0 !LDCAD SNAP_CLP [radius=4] [length=10] [center=true] [slide=true] [pos=25.980762 -15 0] [ori=-1 0 0 0 0 -1 0 -1 0]
0 !LDCAD SNAP_CLP [radius=4] [length=10] [center=true] [slide=true] [pos=25.980762 15 0] [ori=1 0 0 0 0 -1 0 -1 0]


David

[Roland Melkert]

For the part "Technic Sprocket Wheel 25.4" (57520), the following places snap points for each of the sprocket notches. That in turn makes placing the chain tread (57518) somewhat less fraught.

Very useful, I added:

Code:

0 !LDCAD SNAP_CLP [radius=4] [length=10] [center=true] [pos=0 -30 0] [ori=-1 0 0 0 0 1 0 -1 0]
0 !LDCAD SNAP_CLP [radius=4] [length=10] [center=true] [pos=25.980762 -15 -0] [ori=-0.5 -0 -0.866025 -0.866025 -0 0.5 0 -1 0]
0 !LDCAD SNAP_CLP [radius=4] [length=10] [center=true] [pos=25.980762 14 -0] [ori=0.5 -0 -0.866025 -0.866025 -0 -0.5 0 -1 0]
0 !LDCAD SNAP_CLP [radius=4] [length=10] [center=true] [pos=0 30 0] [ori=1 0 0 0 0 -1 0 -1 0]
0 !LDCAD SNAP_CLP [radius=4] [length=10] [center=true] [pos=-25.980762 14 -0] [ori=0.5 0 0.866025 0.866025 -0 -0.5 -0 -1 0]
0 !LDCAD SNAP_CLP [radius=4] [length=10] [center=true] [pos=-25.980762 -15 -0] [ori=-0.5 0 0.866025 0.866025 -0 0.5 -0 -1 0]

to s\57520s02.dat

Only difference with yours are the clip orientations which may come into play in a future version of the snapping engine.

[David Manley]

</snip>

Only difference with yours are the clip orientations which may come into play in a future version of the snapping engine.

Thanks Roland.

When I was tutu-ing with configuring the snap information for the part, I had played about with the orientation and decided it wasn't that significant. However, it is useful to know that there may be benefits to more properly orientating the clips.

It's interesting that the Y co-ordinate of two of the points are changed from 15 to 14. I thought the sprocket would be symmetrical (not that I actually checked!), which is why I used -15 and 15. Presumably s\57520s02.dat reflects the effective Y values as being -15 and 14 hence why you used those two values.

Regards,

David

[Roland Melkert]

It's interesting that the Y co-ordinate of two of the points are changed from 15 to 14. I thought the sprocket would be symmetrical (not that I actually checked!), which is why I used -15 and 15. Presumably s\57520s02.dat reflects the effective Y values as being -15 and 14 hence why you used those two values.

No, you are right.

Something went wrong they should be at 15.

Don't know how that happened as I rotated duplicates of the top one using 0,0,0 as a rotation center.

[David Manley]

For part 41682 (Bracket 2 x 2 - 1 x 2 Up Centred), I can't seem to attach to the back of the vertical plate. The front of the vertical plate is fine.

From playing around a bit, I think that this line:

0 !LDCAD SNAP_CYL [gender=F] [caps=one] [secs=S 6 4] [pos=0 -10 0] [ori=1 0 0 0 0 1 0 -1 0] [grid=C 2 1 20 0]

should be:

0 !LDCAD SNAP_CYL [gender=F] [caps=one] [secs=S 6 4] [pos=0 -10 4] [ori=-1 0 0 0 0 1 0 1 0] [grid=C 2 1 20 0]

Regards,

David

[Roland Melkert]

For part 41682 (Bracket 2 x 2 - 1 x 2 Up Centred), I can't seem to attach to the back of the vertical plate. The front of the vertical plate is fine.

Yes, I made the same correction in response of post #88 but hadn't updated the shadow.sf mentioned in this thread for ages.

[Roland Melkert]

I've processed all thing people mailed me about and things mentioned above up to message #115.

Thanks for all whom contributed.

You can (always) download the latest seed file at:
http://www.melkert.net/action/download/shadow.sf

Just replace the old one in LDCad's seeds location  (subfolder of the exe file).

[N. W. Perry]

Part 12818 is missing its bottom snap info:

Code:

0 !LDCAD SNAP_CYL [gender=F] [caps=one] [secs=S 6 4] [pos=10 144 -50]
0 !LDCAD SNAP_CYL [gender=F] [caps=one] [secs=S 6 4] [pos=50 144 -10]
0 !LDCAD SNAP_CYL [gender=F] [caps=one] [secs=R 6 4] [pos=30 144 -50] [grid=1 2 0 20]
0 !LDCAD SNAP_CYL [gender=F] [caps=one] [secs=R 6 4] [pos=50 144 -30]

(This should also work for unofficial part 79393.)

[Magnus Forsberg]

The Grill Tile is missing snap info on the flat thing in the centre. Should stick to the inside of the hollow stud.

   

[Roland Melkert]

Not really a shadow problem, but the stud of the telescope (64644) is next to impossible to insert into tube of plate 1x1 round (6141). Most of the time, the telescope jumps to connect its tube to plate stud. Moving the plate instead doesn't help. It works much better with a regular plate 1x1 (3024)!

I've investigated this and it is caused by the 'keep sliding info sticky to last match' mechanics.

This is proven by it working more like expected after you remove all 'slide' checks from the involved part's snap info.

Thing is I'm not sure how to improve this so both scenarios work no matter the info configuration.

[N. W. Perry]

These are both very new parts, but 73825 and 79389 need bottom snap info for their 1x1 plate sections.

EDIT: Also 78256 and 79756.

[Roland Melkert]

These are both very new parts, but 73825 and 79389 need bottom snap info for their 1x1 plate sections.

EDIT: Also 78256 and 79756.

Those are from the 22-01 library, I'm still processing 20-01

[N. W. Perry]

Those are from the 22-01 library, I'm still processing 20-01

Any way we can help? These files should be very easy to copy-paste from other library parts. Or, we can limit suggestions for now to 2019 and earlier.

[Roland Melkert]

Just wanted to let people know there is a potential circular reference problem in LDCad's shadow library when used combined with the new 22-03 Library.

Normally circular references are detected during loading, but this is a 'cross reference' one where only the combined data from library and shadow will have the problem.

The bug caused these kinds of references to be missed when the parts involved are loading at the same time. If so it will crash the program.

1.7 Alpha fixes the issue:

Code:

Error    | Circular reference detected to "axl2hol8.dat" from "axl2hole.dat", reference will be cleared and rendered as "missing".

I'm hoping to release the new 1.7 version very soon.

In the meantime it can be prevented by deleting (or editing) axl2hol8.dat from LDCad's shadow library.

[Jaco van der Molen]

Part 54200 (slope brick 31 1 x 1 x 0.677 / cheese slope) no longer has snap info?

[Milan Vančura]

Part 54200 (slope brick 31 1 x 1 x 0.677 / cheese slope) no longer has snap info?

I had the same problem: I updated shadow info but main part library stayed one version behind. Updating the main ldraw library solved the problem. Isn't it your case as well?

Details: Newest shadow library matches the newest main ldraw library, where 54200 was rewritten using a subpart - and shadow info is for that subpart now.

[Jaco van der Molen]

I had the same problem: I updated shadow info but main part library stayed one version behind. Updating the main ldraw library solved the problem. Isn't it your case as well?

Details: Newest shadow library matches the newest main ldraw library, where 54200 was rewritten using a subpart - and shadow info is for that subpart now.

Ah, OK. I think I am up to date, but will check. Tnx for pointing this out.

[SNIPE]

Ah, OK. I think I am up to date, but will check. Tnx for pointing this out.

Hi,

I am slightly confused about something,

I can see that there is (for example) shadow/p/stud.dat but why does ldraw/parts/3024.dat not just use that?
instead, there is shadow/parts/3024.dat and it uses shadow/p/stud.dat, and then shadow/parts/3024.dat links to /ldraw/parts/3024.dat

Because then LDRAW part authors can just use shadow/p/ (which will almost always have the correct primative in it anyway) and not wait for ldcad to update /shadow/parts

Now I know 3024 has all of the primative snapping info intact but what about something like ldraw/parts/86996.dat which most probably have the correct primative snapping in the shadow library, but is not snapping on the bottom (antistud/clutch)

Perhaps someone can clear this up? If this is not on-topic I guess I could just PM roland.

[Roland Melkert]

Perhaps someone can clear this up?

I'm not entirely sure what you mean but:

The shadow is basically a patch to be applied to the official ldraw library.

Information can be inherited (like stud.dat) or included from another shadow file.

Parts without a shadow file will only inherit. So they usually end up without bottom connectivity because there are no useful primitives involved for shaping the antistuds.

[Marc Giraudet]

This is intended to be a catchall thread to let Roland know about missing/problematic part snapping info.

Note: make sure that the part you are reporting is official

42924 "Pin long with friction Type 2" and 61332 "Pin with friction Type 2" missing ?

[Orion Pobursky]

42924 "Pin long with friction Type 2" and 61332 "Pin with friction Type 2" missing ?

Those were just released last update and use new pin primitives. There is some lag time since Roland has to manually update his shadow lib.

[Roland Melkert]

Those were just released last update and use new pin primitives. There is some lag time since Roland has to manually update his shadow lib.

Make that a lot of lag, as I haven't had time for it at all lately.

I was hoping the github repository would help me improve on this.
https://github.com/RolandMelkert/LDCadShadowLibrary

But no pull requests so far.

[Jaco van der Molen]

Make that a lot of lag, as I haven't had time for it at all lately.

I was hoping the github repository would help me improve on this.
https://github.com/RolandMelkert/LDCadShadowLibrary

But no pull requests so far.

OK, nice, but how does one make a offLibShadow.csl file from that?

[Roland Melkert]

OK, nice, but how does one make a offLibShadow.csl file from that?

You can download the whole repository (green 'code' button top right) or use a local git installation to clone it.

Then in LDCad you point the shadow to the root directory (holding parts and p) just like the normal library locations.


If you really want a csl, just (re)zip the contents making sure the parts and p folders are the root folders inside the zip and rename it to csl.

The downloaded zip can't be used directly because at the moment LDCad doesn't compensate for the 'LDCadShadowLibrary-main' root folder inside it.

This is something that will be improved upon in the next version (It will support using the github downloaded zip as is).

[Jaco van der Molen]

You can download the whole repository (green 'code' button top right) or use a local git installation to clone it.

Then in LDCad you point the shadow to the root directory (holding parts and p) just like the normal library locations.


If you really want a csl, just (re)zip the contents making sure the parts and p folders are the root folders inside the zip and rename it to csl.

The downloaded zip can't be used directly because at the moment LDCad doesn't compensate for the 'LDCadShadowLibrary-main' root folder inside it.

This is something that will be improved upon in the next version (It will support using the github downloaded zip as is).

Ah yes, the csl is just a zip.
I will try this. Tnx.

[Jaco van der Molen]

Ah yes, the csl is just a zip.
I will try this. Tnx.

OK, this works. I downloaded, unpacked, made a re-zip, etc.
I found another brick with missing snap info (92946) and created a file for that (attached)

  92946.dat (Size: 301 bytes / Downloads: 0)
Now, can I contribute this file to the Github repos?

Is the contents correct like so:

Code:

0 LDCad shadow info for "Slope Plate 45.73  2 x  1"

0 Author: LDCad Shadow Library
0 !LICENSE CC BY-SA 4.0, see LICENSE.md

0 !HISTORY 2022-11-16 {Jaco van der Molen} Initial info for 15672.dat copy to 92946

0 !LDCAD SNAP_CYL [gender=F] [caps=one] [secs=S 6 4] [pos=0 8 0] [grid=1 C 2 0 20]

[Roland Melkert]

Is the contents correct like so:

Looks ok to me, but maybe use "Initial info for 92946.dat (copy of 15672.dat)" as the description?

[Jaco van der Molen]

Working on a model that had a few 2x2 inververted 45 degree slopes, I noticed it was a bit hard to snap these.

Doing some research, I conclude:

3660a (Slope Brick 45  2 x  2 Inverted without Inner Stopper Ring) is missing info for the front studs with hole and inner info for the 1x2 brick part.
This can be fixed adding 

Code:

0 !LDCAD SNAP_CYL [ID=studO] [gender=M] [caps=one] [secs=R 6 4] [pos=10 0 -20]
0 !LDCAD SNAP_CYL [ID=studO] [gender=M] [caps=one] [secs=R 6 4] [pos=-10 0 -20]

or use the subpart for stud2a?

For the inner bottom inside use 3004s01.dat and thus get the snap info from that

Code:

0 !LDCAD SNAP_INCL [ref=s\3004s01.dat]

The other, 3660b (Slope Brick 45  2 x  2 Inverted with Inner Stopper Ring), is missing the info inside (1x2 brick part), but does have snap info for either of the 4 studs.
The latter come from the subpart 3660bs01.

Add

Code:

0 !LDCAD SNAP_CYL [gender=F] [caps=one] [secs=S 6 20] [pos=10 24 0]
0 !LDCAD SNAP_CYL [gender=F] [caps=one] [secs=S 6 20] [pos=-10 24 0]

[Roland Melkert]

Working on a model that had a few 2x2 inververted 45 degree slopes, I noticed it was a bit hard to snap these.

Are you sure you have the latest shadow library (from github)?

As the part looks ok to me.

It has studs inherited from primitives and the bottom is inheried from s\3660as01.dat:

Code:

0 LDCad shadow info for "~Slope Brick 45  2 x  2 Inverted without Inner Stopper Ring without Front Face"

0 Author: LDCad Shadow Library
0 !LICENSE CC BY-SA 4.0, see LICENSE.md

0 !HISTORY 2014-02-07 {Roland Melkert} Initial info for s\3660s01.dat
0 !HISTORY 2022-12-29 {Roland Melkert} Moved to s\3660as01.dat

0 !LDCAD SNAP_CYL [gender=F] [caps=one] [secs=S 6 20] [pos=0 24 0] [grid=C 2 1 20 0]
0 !LDCAD SNAP_CYL [gender=M] [caps=one] [secs=R 6 4] [pos=0 0 -20] [grid=C 2 1 20 0]

[Jaco van der Molen]

Are you sure you have the latest shadow library (from github)?

If you put it that way, no: I am not sure, but will check.
Currently I am working with a local and unpacked version to easy contribute tot said library.

[N. W. Perry]

If you put it that way, no: I am not sure, but will check.
Currently I am working with a local and unpacked version to easy contribute tot said library.

I had the same problem, and indeed, I did not have the latest version of the shadow library. I guess I just need to get in the habit of updating it manually. :-)

[Jaco van der Molen]

If you put it that way, no: I am not sure, but will check.
Currently I am working with a local and unpacked version to easy contribute tot said library.

OK, I did not have the latest 

The inverted slope is OK as it is!
Tnx.

@Roland: I do have some other contributions to make that I will submit to the Github repos.

[N. W. Perry]

32064a and 32064b have different bottom snap shapes, and 32064c and 32064d have no bottom snap info. Probably, all four should inherit from the subpart 32064s01 instead?

[Magnus Forsberg]

32064a and 32064b have different bottom snap shapes, and 32064c and 32064d have no bottom snap info. Probably, all four should inherit from the subpart 32064s01 instead?

Depends, 3 of them have a bottom stud3 that should snap, but the first, 32064a have a stud blocker design.

btw,
why doesn't the stud3a have any snap info?

[N. W. Perry]

Depends, 3 of them have a bottom stud3 that should snap, but the first, 32064a have a stud blocker design.

btw,
why doesn't the stud3a have any snap info?

Well, I know that snap info is intentionally left out for offset positions on most undersides, because a) it’s too many snapping points to calculate, and b) it’s easy enough to achieve using grid snapping. (A common exception is the newer stud-holder design on many parts, such as the revised jumper plate.)

[Roland Melkert]

32064a and 32064b have different bottom snap shapes, and 32064c and 32064d have no bottom snap info. Probably, all four should inherit from the subpart 32064s01 instead?

The anti-stud shape is different on purpose, as most of them would accept a square stud (not sure if those exist )

You could add round ones to the shared subpart indeed, but as that one is lacking internals it's less future proof.

I made these for now.

32064c:

Code:

0 LDCad shadow info for "Technic Brick  1 x  2 with Axlehole with Open Sides and Bottom Stud"

0 Author: LDCad Shadow Library
0 !LICENSE CC BY-SA 4.0, see LICENSE.md

0 !HISTORY 2022-03-03 {Roland Melkert} Initial info for 32064c.dat

0 !LDCAD SNAP_CYL [gender=F] [caps=one] [secs=S 6 6] [pos=0 24 0] [grid=C 2 1 20 0]

32064d: (not sure about the include though. Might change it to a plain copy of the cyl line.)

Code:

0 LDCad shadow info for "Technic Brick  1 x  2 with Semi-Reduced Axlehole"

0 Author: LDCad Shadow Library
0 !LICENSE CC BY-SA 4.0, see LICENSE.md

0 !HISTORY 2022-03-03 {Roland Melkert} Initial info for 32064d.dat

0 !LDCAD SNAP_INCL [ref=32064c.dat]

Sorry about the lack of updates lately.

I'm working on some changes to the part editor to maintain the new github headers automatically.

Afterwards I'm hoping to extend the shadow library on a more regular basis again.

[Jaco van der Molen]

Part 27150 minifig umbrella does not snap to ie. 60897 plate 1x1 with clip.
Both do have snap info, but it doesn't seem to work?

The umbrella has snap info at the handle part, but I think it does fit a clip at the middle part too?

   

[Jaco van der Molen]

Both 15100 and 65487 [Technic pin with friction with perpendicular pin hole] are missing snap info at the pin.

This could be:

Code:

0 !LDCAD SNAP_CYL [gender=M] [caps=one] [secs=R 8 2  R 6 16  _L 6.25 2] [pos=10 0 0] [ori=0 -1 0 1 0 0 0 0 1]

[Roland Melkert]

Part 27150 minifig umbrella does not snap to ie. 60897 plate 1x1 with clip.
Both do have snap info, but it doesn't seem to work?

The 'anticlip' is too short while being capped on both ends.
Setting its length to 8 or caps to A will fix it.

The middle one would work with

Code:

0 !LDCAD SNAP_CYL [gender=M] [caps=none] [secs=R 4 5] [center=true]

[Roland Melkert]

Both 15100 and 65487 [Technic pin with friction with perpendicular pin hole] are missing snap info at the pin.

Adding it to the new pin primitive should fix it for all in one go.

Code:

0 LDCad shadow info for "Technic Friction Pin 1.0 with Base Collar and Small Rectangular Centre Hole"

0 Author: LDCad Shadow Library
0 !LICENSE CC BY-SA 4.0, see LICENSE.md

0 !HISTORY 2023-03-23 {Roland Melkert} Initial info for confric11.dat

0 !LDCAD SNAP_CYL [ID=fpin11] [gender=M] [caps=one] [secs=R 8 2   R 6 16   _L 6.25 2]

[Jaco van der Molen]

Adding it to the new pin primitive should fix it for all in one go.

Code:

0 LDCad shadow info for "Technic Friction Pin 1.0 with Base Collar and Small Rectangular Centre Hole"

0 Author: LDCad Shadow Library
0 !LICENSE CC BY-SA 4.0, see LICENSE.md

0 !HISTORY 2023-03-23 {Roland Melkert} Initial info for confric11.dat

0 !LDCAD SNAP_CYL [ID=fpin11] [gender=M] [caps=one] [secs=R 8 2   R 6 16   _L 6.25 2]