(solved) 0.18.15248 do not open files correct

[user1234]

Hello!

I made a compile now and i can not open files (assemblies). Smaller files seems to work, but larger files not. They worked 0.18.15240 (also i do not changed them). The CPU drives with all cores with 100% and the RAM goes up till FreeCAD dies. I think that a type of element will not load anymore.

Code: Select all

OS: Debian GNU/Linux buster/sid
Word size of OS: 64-bit
Word size of FreeCAD: 64-bit
Version: 0.18.15248 (Git)
Build type: Release
Branch: master
Hash: 608d42d3cdce307b585534adfdae9c9af7304436
Python version: 2.7.15+
Qt version: 5.11.2
Coin version: 4.0.0a
OCC version: 7.3.0
Locale: C/Default (C)

Greetings
user1234

[triplus]

Hi @user1234

If possible could you narrow the issue down a bit further? As you compile FreeCAD yourself. Could you compile at:

Code: Select all

git checkout fc5b6288c6eb4fe89192f63f1fbdf4a43c0fad27

To see if the issue still persist. If not likely a regression was introduced in one of the Sketcher related commits (after Nov 16, 2018).

[user1234]

this is 018.15221. But i know that worked before. Must a be one of the last 8 commits or one of my dependencies that i updated today. I will compile through the last. The funny thing, that the compiling today gave my new warnings. But i did not save the logs.

edit: i must stop today, will test this tomorrow

[abdullah]

If this is confirmed, we will need an offending file, the smallest the better.

[PrzemoF]

@user1234: did you try git bisect to get the offending commit?

https://www.preining.info/blog/2015/04/ ... made-easy/
https://www.metaltoad.com/blog/beginner ... limination

[Kunda1]

@user1234: did you try git bisect to get the offending commit?

https://www.preining.info/blog/2015/04/ ... made-easy/
https://www.metaltoad.com/blog/beginner ... limination

An aside:
probably a good idea to mention this in a dedicated section on the Github wikipage.

[abdullah]

Maybe:

ddfa51056559c70bf8698c79af69eb74a339ba81 works

ddb20468ad438f05c77b02be9b048ace92e8ad6d does not work

In those commits the XML reader was modified to throw an exception is a non-existing attribute is tried to be restored. This should be a feature in the sense that it should help discover problems that were being ignored.

If that is the case, I would really need an offending file to figure out which kind of object/attribute is the one making the mess...

[user1234]

Hello!

That thing also thought yesterday because, XML and file loading, that could me a match (whithout a spark of experience from my side). I am compiling (at first ddfa51056559c70bf8698c79af69eb74a339ba81 release mode to test if that works) now, please wait approx 1h. Then i can say more.

Greetings

[user1234]

So, thats funny. I compiled one, do not work. So just in case, deleted the userfiles, that assemlies files loads again. I know, i should do that before, i have not. But i did not set up anything since the preselecting uncheck (that slowsdown FreeCAD on assemblies extreme).

Sorry for the noise i made.

Greetings
user1234

[DeepSOIC]

You have a perfect signature of tessellation settings being screwed up. Did your slowdown on large assemblies with preselection "on" go away?