Improve DIMASSOC object #1029

michal-josef-spacek · 2024-10-10T21:45:18Z

No description provided.

rurban · 2024-10-14T07:11:10Z

As smoke/dimassoc-pr1029

rurban · 2024-10-14T21:35:42Z

test make && make -C test/unit-testing/ dimassoc && (cd test/unit-testing && ./dimassoc; cd -) in smoke/dimassoc-pr1029. osnap_pt is wrong.

unknowns see DIMASSOC.zip

rurban

Doesnt parse. see the unit-tests and the pi logs

include/dwg.h

michal-josef-spacek · 2024-10-15T12:08:31Z

I am trying to test, but I was blocked by #1032

michal-josef-spacek · 2024-10-15T18:58:39Z

Ad make && make -C test/unit-testing/ dimassoc && (cd test/unit-testing && ./dimassoc; cd -):
Yes, all >AC1018 are bad, condition with minimal length of item (35 bytes) to end of block is not working.

test/test-data/example_2018.dwg:

...
trans_space_flag: 0 [B 70]
rotated_type: 0x0 [RC 71]
dimensionobj: (4.2.399) abs:921 [H 330]
DIMASSOC_Ref.rcount1: 0
ref[rcount1].classname: "AcDbOsnapPointRef" [TU 1]
ref[0].osnap_type: 0xd [RC 72]
ref[0].num_xrefs: 1 [BL 0]
ref[rcount1].xrefs[vcount]: (4.2.398) abs:920 [H 331]
ref[0].main_subent_type: 2 [BL 73]
ref[0].main_gsmarker: 0 [BL 91]
ref[0].num_xrefpaths: 0 [BL 0]
ref[0].osnap_dist: 4.13129 [BD 40]
ref[0].osnap_pt: (0, 0, 2e+50) [3BD 10]
ref[0].has_lastpt_ref: 0 [B 75]
DIMASSOC_Ref.rcount1: 1
ref[rcount1].classname: "AcDbOsnapPointRef" [TU 1]
ref[1].osnap_type: 0x1 [RC 72]
ref[1].num_xrefs: 1 [BL 0]
ref[rcount1].xrefs[vcount]: (4.2.398) abs:920 [H 331]
ref[1].main_subent_type: 2 [BL 73]
ref[1].main_gsmarker: 0 [BL 91]
ref[1].num_xrefpaths: 0 [BL 0]
ref[1].osnap_dist: 7.44845 [BD 40]
ref[1].osnap_pt: (0, 0, 2e+50) [3BD 10]
ref[1].has_lastpt_ref: 0 [B 75]
DIMASSOC_Ref.rcount1: 2
ref[rcount1].classname: "AcDbOsnapPointRef" [TU 1]
ref[2].osnap_type: 0x3 [RC 72]
ref[2].num_xrefs: 1 [BL 0]
ref[rcount1].xrefs[vcount]: (4.2.398) abs:920 [H 331]
ref[2].main_subent_type: 2 [BL 73]
ref[2].main_gsmarker: 0 [BL 91]
ref[2].num_xrefpaths: 0 [BL 0]
ref[2].osnap_dist: 4.13129 [BD 40]
ref[2].osnap_pt: (0, 0, 0) [3BD 10]
ref[2].has_lastpt_ref: 0 [B 75]
DIMASSOC_Ref.rcount1: 3
ERROR: bit_read_TU buffer overflow at 170.6 + 112 > 172
ref[rcount1].classname: "" [TU 1]
ref[3].osnap_type: 0x44 [RC 72]
ref[3].num_xrefs: 65 [BL 0]
ERROR: Invalid xrefs size 65. Need min. 520 bits, have 3 for DIMASSOC. Set _size to 0
 object_map{3B4} = 261
 padding: 80/0 (6 bits)
crc: 3461 [RSx]
 check_CRC 121033-121226 = 193: 3461 == 3461

It need to end before last block.

rurban · 2024-12-12T21:34:17Z

I fixed the unit-test now. branch work/dimassoc-pr1029

michal-josef-spacek · 2024-12-13T11:35:34Z

@rurban

I fixed the unit-test now. branch work/dimassoc-pr1029

First, in this branch, the commit 3b55ce3 is wrong. There is 3BD, not 2BD. Or do you have some information that I don't have?

Second. The main problem with the activity to fix DIMASSOC is how to count AcDbOsnapPointRef classes. I don't see any count for a number of classes. The question is how to detect the start and end of this section. One possibility is to check the first string AcDbOsnapPointRef. Or I don't know. When we brainstormed this issue, I could finish the DIMASSOC object and test it.

rurban · 2024-12-13T17:49:37Z

@rurban

I fixed the unit-test now. branch work/dimassoc-pr1029

First, in this branch, the commit 3b55ce3 is wrong. There is 3BD, not 2BD. Or do you have some information that I don't have?

Second. The main problem with the activity to fix DIMASSOC is how to count AcDbOsnapPointRef classes. I don't see any count for a number of classes. The question is how to detect the start and end of this section. One possibility is to check the first string AcDbOsnapPointRef. Or I don't know. When we brainstormed this issue, I could finish the DIMASSOC object and test it.

As you in see in the current code, the count is computed from a number of bits

michal-josef-spacek · 2024-12-15T10:42:13Z

@rurban
If you mean this code:

  // 6 = maximum of items
  REPEAT_CN (6, ref, Dwg_DIMASSOC_Ref) // i.e. AcDbOsnapPointRef
  REPEAT_BLOCK
      // 35 is minimal size of item
      if (dat->byte + 35 > dat->size)
        {
#ifdef IS_JSON 
          ENDHASH;
#endif
          break;
        }

I wrote it.
This is not working in every use case.
I am blocked on this detection.

rurban · 2024-12-15T17:13:44Z

@rurban If you mean this code:

  // 6 = maximum of items
  REPEAT_CN (6, ref, Dwg_DIMASSOC_Ref) // i.e. AcDbOsnapPointRef
  REPEAT_BLOCK
      // 35 is minimal size of item
      if (dat->byte + 35 > dat->size)
        {
#ifdef IS_JSON 
          ENDHASH;
#endif
          break;
        }

I wrote it. This is not working in every use case. I am blocked on this detection.

Yes, this wrong. See my current code which loops on the bits

michal-josef-spacek · 2024-12-16T11:58:28Z

Ach, I rewrited your code to my, because I had issues with it. Ok, maybe it isn't wrong.

michal-josef-spacek · 2024-12-16T14:46:51Z

I have examples:
File #1 (AC1018 DXF):
associativity: 15, number of AcDbOsnapPointRef: 4
associativity: 7, number of AcDbOsnapPointRef: 3
associativity: 3, number of AcDbOsnapPointRef: 2
associativity: 2, number of AcDbOsnapPointRef: 1

File #2 (AC1018 DXF):
associativity: 3, number of AcDbOsnapPointRef: 3
associativity: 3, number of AcDbOsnapPointRef: 2
associativity: 2, number of AcDbOsnapPointRef: 2
associativity: 2, number of AcDbOsnapPointRef: 1
associativity: 1, number of AcDbOsnapPointRef: 1

rurban · 2024-12-16T18:48:41Z

I have examples: File #1 (AC1018 DXF): associativity: 15, number of AcDbOsnapPointRef: 4 associativity: 7, number of AcDbOsnapPointRef: 3 associativity: 3, number of AcDbOsnapPointRef: 2 associativity: 2, number of AcDbOsnapPointRef: 1

Here the popcounts do match

File #2 (AC1018 DXF): associativity: 3, number of AcDbOsnapPointRef: 3 associativity: 3, number of AcDbOsnapPointRef: 2 associativity: 2, number of AcDbOsnapPointRef: 2 associativity: 2, number of AcDbOsnapPointRef: 1 associativity: 1, number of AcDbOsnapPointRef: 1

Here only 3 of 5 match, hmm

michal-josef-spacek · 2024-12-16T20:44:52Z

@rurban This is definitely #1031 (comment)

lots of coverage

michal-josef-spacek · 2024-12-30T21:14:09Z

@rurban I improved PR.

Nice example output from this PR code (index 0, 1, 4, 5, 6):

--common_size: 76
unknown_bits [1578 (76,1416,0) 198 TF]: 40C008920B1A23127B9B730B82837B4B73A2932B300042028151E8E9DCBCBBC3DA7F5512416344624F736E6170506F696E745265660000A1C2C7DE4B4431150481E089F982FCF34088920B1
A23127B9B730B82837B4B73A2932B300042028151E8E9DCBCBBC3DA7F5512416344624F736E6170506F696E74526566000840502A3D1D3B9797787B4FEAA2482C688C49EE6DCC2E0A0DED2DCE8A4CACC000143858FBC9688622A0903C113F30
5F9E6812010D8C1C4CC10D8C1C490D8C1BB50D8C1BBD0D8C1BB703

associativity: 0x3 [BLx 90]
trans_space_flag: 0 [B 70]
rotated_type: 0x0 [RC 71]
dimensionobj: (4.3.630712) abs:6489874 [H 330]
DIMASSOC_Ref.rcount1: 0
ref[rcount1].classname: "AcDbOsnapPointRef" [T 1]
ref[0].osnap_type: 0x8 [RC 72]
ref[0].num_xrefs: 1 [BL 0]
ref[rcount1].xrefs[vcount]: (4.3.6306ED) abs:6489837 [H 331]
ref[0].main_subent_type: 2 [BL 73]
ref[0].main_gsmarker: 0 [BL 91]
ref[0].num_xrefpaths: 0 [BL 0]
ref[0].osnap_dist: 0.93376 [BD 40]
ref[0].osnap_pt: (0, 0, 0) [3BD 10]
ref[0].has_lastpt_ref: 1 [B 75]
DIMASSOC_Ref.rcount1: 1
ref[rcount1].classname: "AcDbOsnapPointRef" [T 1]
ref[1].osnap_type: 0x0 [RC 72]
ref[1].num_xrefs: 0 [BL 0]
ref[1].osnap_dist: 0 [BD 40]
ref[1].osnap_pt: (2.75908e+06, 81864.2, 0) [3BD 10]
ref[1].has_lastpt_ref: 0 [B 75]
DIMASSOC_Ref.rcount1: 4
ref[4].classname: "AcDbOsnapPointRef" [T 1]
ref[4].osnap_type: 0x8 [RC 72]
ref[4].num_xrefs: 1 [BL 0]
ref[4].xrefs[vcount]: (4.3.6306EF) abs:6489839 [H 331]
ref[4].main_subent_type: 2 [BL 73]
ref[4].main_gsmarker: 0 [BL 91]
ref[4].num_xrefpaths: 0 [BL 0]
ref[4].osnap_dist: 0.93376 [BD 40]
ref[4].osnap_pt: (0, 0, 0) [3BD 10]
ref[4].has_lastpt_ref: 1 [B 75]
DIMASSOC_Ref.rcount1: 5
ref[5].classname: "AcDbOsnapPointRef" [T 1]
ref[5].osnap_type: 0x8 [RC 72]
ref[5].num_xrefs: 1 [BL 0]
ref[5].xrefs[vcount]: (4.3.6306ED) abs:6489837 [H 331]
ref[5].main_subent_type: 2 [BL 73]
ref[5].main_gsmarker: 0 [BL 91]
ref[5].num_xrefpaths: 0 [BL 0]
ref[5].osnap_dist: 0.93376 [BD 40]
ref[5].osnap_pt: (0, 0, 0) [3BD 10]
ref[5].has_lastpt_ref: 1 [B 75]
DIMASSOC_Ref.rcount1: 6
ref[6].classname: "AcDbOsnapPointRef" [T 1]
ref[6].osnap_type: 0x0 [RC 72]
ref[6].num_xrefs: 0 [BL 0]
ref[6].osnap_dist: 0 [BD 40]
ref[6].osnap_pt: (2.75908e+06, 81864.2, 0) [3BD 10]
ref[6].has_lastpt_ref: 0 [B 75]
 padding: +4
 object_map{630714} = 17672
 padding: DF/F (4 bits)
crc: 506D [RSx]
 check_CRC 7452086-7452297 = 211: 506D == 506D

michal-josef-spacek · 2024-12-30T21:21:03Z

@rurban In the case of your code:

...
one_more = SUB_FIELD_VALUE(ref[rcount1], has_lastpt_ref);
...

This is not working.
one_more could be set from any first four has_lastpt_ref
Example:

--common_size: 76
unknown_bits [986 (76,856,0) 124 TF]: 40C008920B1A23127B9B730B82837B4B73A2932B300042028150E6B84AB50D97DE7F5512416344624F736E6170506F696E745265660000A131B0CDAD64D00904909281CA02B9DAC088920B1A2
3127B9B730B82837B4B73A2932B30000A02815344D323F62870DB0A52010C2681BCC10C2681B90C267B750C26772F03

associativity: 0x3 [BLx 90]
trans_space_flag: 0 [B 70]
rotated_type: 0x0 [RC 71]
dimensionobj: (4.3.9A06E) abs:630894 [H 330]
DIMASSOC_Ref.rcount1: 0
ref[rcount1].classname: "AcDbOsnapPointRef" [T 1]
ref[0].osnap_type: 0x8 [RC 72]
ref[0].num_xrefs: 1 [BL 0]
ref[rcount1].xrefs[vcount]: (4.3.99EDD) abs:630493 [H 331]
ref[0].main_subent_type: 2 [BL 73]
ref[0].main_gsmarker: 0 [BL 91]
ref[0].num_xrefpaths: 0 [BL 0]
ref[0].osnap_dist: 0.993594 [BD 40]
ref[0].osnap_pt: (0, 0, 0) [3BD 10]
ref[0].has_lastpt_ref: 1 [B 75]
DIMASSOC_Ref.rcount1: 1
ref[rcount1].classname: "AcDbOsnapPointRef" [T 1]
ref[1].osnap_type: 0x0 [RC 72]
ref[1].num_xrefs: 0 [BL 0]
ref[1].osnap_dist: 0 [BD 40]
ref[1].osnap_pt: (149131, -27364, 0) [3BD 10]
ref[1].has_lastpt_ref: 0 [B 75]
DIMASSOC_Ref.rcount1: 4
ref[4].classname: "AcDbOsnapPointRef" [T 1]
ref[4].osnap_type: 0x1 [RC 72]
ref[4].num_xrefs: 1 [BL 0]
ref[4].xrefs[vcount]: (4.3.99DCB) abs:630219 [H 331]
ref[4].main_subent_type: 2 [BL 73]
ref[4].main_gsmarker: 0 [BL 91]
ref[4].num_xrefpaths: 0 [BL 0]
ref[4].osnap_dist: 1 [BD 40]
ref[4].osnap_pt: (0, 0, 2e+50) [3BD 10]
ref[4].has_lastpt_ref: 0 [B 75]
 padding: +4
 object_map{9A070} = 19484
 padding: BF/F (4 bits)
crc: 9031 [RSx]
 check_CRC 2397786-2397923 = 137: 9031 == 9031

michal-josef-spacek mentioned this pull request Oct 12, 2024

Fix DIMASSOC object #1031

Open

3 tasks

michal-josef-spacek force-pushed the improve_dimassoc branch from a41e040 to e64aaef Compare October 13, 2024 08:26

rurban requested changes Oct 15, 2024

View reviewed changes

include/dwg.h Outdated Show resolved Hide resolved

rurban linked an issue Dec 12, 2024 that may be closed by this pull request

Fix DIMASSOC object #1031

Open

3 tasks

promote DIMASSOC to unstable

dcfe2e0

lots of coverage

michal-josef-spacek force-pushed the improve_dimassoc branch from e64aaef to a4ca4a6 Compare December 30, 2024 18:05

Improve DIMASSOC object

812ff03

michal-josef-spacek force-pushed the improve_dimassoc branch from a4ca4a6 to 812ff03 Compare December 30, 2024 18:17

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Improve DIMASSOC object #1029

Improve DIMASSOC object #1029

michal-josef-spacek commented Oct 10, 2024

rurban commented Oct 14, 2024

rurban commented Oct 14, 2024 •

edited

Loading

rurban left a comment

michal-josef-spacek commented Oct 15, 2024

michal-josef-spacek commented Oct 15, 2024

rurban commented Dec 12, 2024

michal-josef-spacek commented Dec 13, 2024

rurban commented Dec 13, 2024

michal-josef-spacek commented Dec 15, 2024

rurban commented Dec 15, 2024

michal-josef-spacek commented Dec 16, 2024

michal-josef-spacek commented Dec 16, 2024 •

edited

Loading

rurban commented Dec 16, 2024

michal-josef-spacek commented Dec 16, 2024

michal-josef-spacek commented Dec 30, 2024

michal-josef-spacek commented Dec 30, 2024

Improve DIMASSOC object #1029

Are you sure you want to change the base?

Improve DIMASSOC object #1029

Conversation

michal-josef-spacek commented Oct 10, 2024

rurban commented Oct 14, 2024

rurban commented Oct 14, 2024 • edited Loading

rurban left a comment

Choose a reason for hiding this comment

michal-josef-spacek commented Oct 15, 2024

michal-josef-spacek commented Oct 15, 2024

rurban commented Dec 12, 2024

michal-josef-spacek commented Dec 13, 2024

rurban commented Dec 13, 2024

michal-josef-spacek commented Dec 15, 2024

rurban commented Dec 15, 2024

michal-josef-spacek commented Dec 16, 2024

michal-josef-spacek commented Dec 16, 2024 • edited Loading

rurban commented Dec 16, 2024

michal-josef-spacek commented Dec 16, 2024

michal-josef-spacek commented Dec 30, 2024

michal-josef-spacek commented Dec 30, 2024

rurban commented Oct 14, 2024 •

edited

Loading

michal-josef-spacek commented Dec 16, 2024 •

edited

Loading