PES to SEW conversion not working

[BartmanEH]

fresh build of libembroiderery on mac 12.1 but conversion from PES to SEW produces invalid results

[robin-swift]

Ok, can you show me the terminal output from the conversion?

Let's work this out.

[BartmanEH]

%./libembroidery/embroider test.pes test.sew                                                       
test.pes

maybe I'm missing -t sew?

[robin-swift]

No that should work. What's the file size of the output?

[BartmanEH]

22KB (input PES is 117KB)

[robin-swift]

Does the sew file not load at all? Or is it incomplete?

[BartmanEH]

using StitchBuddy, the PES previews fine but the SEW is blank. Trying to open it in StitchBuddy yields:

[robin-swift]

Ah, that's the error message I need. Can you paste the output of:

% xxd test.sew | head

I.e. the first 256bytes in hex format.

[BartmanEH]

00000000: 6004 0000 0a00 0000 1d00 0000 2900 0000  `...........)...
00000010: 0100 0000 3a00 0000 1d00 0000 2900 0000  ....:.......)...
00000020: 0100 0000 4200 0000 1d00 0000 1e00 0000  ....B...........
00000030: 0100 0000 0c00 0000 1d00 0000 1e00 0000  ................
00000040: 0100 0000 4200 0000 1d00 0000 1e00 0000  ....B...........
00000050: 0100 0000 0100 0000 0100 0000 3200 0000  ............2...
00000060: 0100 0000 0100 0000 0100 0000 3200 0000  ............2...
00000070: 0100 0000 3200 0000 1d00 0000 3200 0000  ....2.......2...
00000080: 0100 0000 3200 0000 1d00 0000 3200 0000  ....2.......2...
00000090: 0100 0000 3200 0000 1d00 0000 3200 0000  ....2.......2...

[BartmanEH]

I tried DST conversion but it doesn't work either and I get a lot of screen output:

 libembroidery % ./embroider test.pes test.dst
test.pes
0 90.933333 19.111111 909 191 909 191
ERROR: format-t01.c encode_record(), x is not in valid range [-121,121] , x = 909
1 79.566667 16.722222 -113 -24 796 167
2 68.200000 14.333333 -114 -24 682 143
3 56.833333 11.944444 -114 -24 568 119
4 45.466667 9.555556 -113 -23 455 96
5 34.100000 7.166667 -114 -24 341 72
6 22.733333 4.777778 -114 -24 227 48
7 11.366667 2.388889 -113 -24 114 24
8 102.300000 21.500000 909 191 1023 215
ERROR: format-t01.c encode_record(), x is not in valid range [-121,121] , x = 909

[truncated]

[robin-swift]

Yeah, that's big jumps not being split properly. I know about that bug.

If you hang on 5 minutes I may work out what this colour change bug is.

[robin-swift]

Can you paste the output of

% ./libembroidery/embroider test.pes test.csv
% head test.csv

[BartmanEH]

% head test.csv      
"#","Embroidermodder 2 CSV Embroidery File"
"#","http://embroidermodder.github.io"

"#","General Notes:"
"#","This file can be read by Excel or LibreOffice as CSV (Comma Separated Value) or with a text editor."
"#","Lines beginning with # are comments."
"#","Lines beginning with > are variables: [VAR_NAME], [VAR_VALUE]"
"#","Lines beginning with $ are threads: [THREAD_NUMBER], [RED], [GREEN], [BLUE], [DESCRIPTION], [CATALOG_NUMBER]"
"#","Lines beginning with * are stitch entries: [STITCH_TYPE], [X], [Y]"

[robin-swift]

So what I want to see is slightly below that:

% head -n40 test.csv

[BartmanEH]

% head -n40 test.csv
"#","Embroidermodder 2 CSV Embroidery File"
"#","http://embroidermodder.github.io"

"#","General Notes:"
"#","This file can be read by Excel or LibreOffice as CSV (Comma Separated Value) or with a text editor."
"#","Lines beginning with # are comments."
"#","Lines beginning with > are variables: [VAR_NAME], [VAR_VALUE]"
"#","Lines beginning with $ are threads: [THREAD_NUMBER], [RED], [GREEN], [BLUE], [DESCRIPTION], [CATALOG_NUMBER]"
"#","Lines beginning with * are stitch entries: [STITCH_TYPE], [X], [Y]"

"#","Stitch Entry Notes:"
"#","STITCH instructs the machine to move to the position [X][Y] and then make a stitch."
"#","JUMP instructs the machine to move to the position [X][Y] without making a stitch."
"#","TRIM instructs the machine to cut the thread before moving to the position [X][Y] without making a stitch."
"#","COLOR instructs the machine to stop temporarily so that the user can change to a different color thread before resuming."
"#","END instructs the machine that the design is completed and there are no further instructions."
"#","UNKNOWN encompasses instructions that may not be supported currently."
"#","[X] and [Y] are absolute coordinates in millimeters (mm)."

"#","[VAR_NAME]","[VAR_VALUE]"
">","STITCH_COUNT:","3896"
">","THREAD_COUNT:","251"
">","EXTENTS_LEFT:","-3499.600000"
">","EXTENTS_TOP:","-2996.000000"
">","EXTENTS_RIGHT:","3458.800000"
">","EXTENTS_BOTTOM:","4564.200000"
">","EXTENTS_WIDTH:","6958.400000"
">","EXTENTS_HEIGHT:","7560.200000"

"#","[THREAD_NUMBER]","[RED]","[GREEN]","[BLUE]","[DESCRIPTION]","[CATALOG_NUMBER]"
"$","1","32","100","120","f. Unimplemented for now.","an the header."
"$","2","58","32","101","mb-polyline.c embArray_create(), ","memory for heapPolylineObj
"
"$","3","0","0","0","Unknown",""
"$","4","58","32","37","d
Name  : %s

","on  : %s
Description: %s
Reade"

[robin-swift]

Oh dear, that's a new one. I've not cleared out the memory for the strings so it's got whatever rubbish is there.

I'm writing a patch that should help with this testing and fix some minor bugs, hang on.

[robin-swift]

Ok, git pull and rebuild with cmake --build ..

[BartmanEH]

oh man... I'm not a dev, I'm a hack... this is gonna go horribly wrong :-)

EDIT: how do I git pull for just libembroidery?

[robin-swift]

You've followed me to here I've not been confused by your replies at any point.

This is a good real world test of the software, so I'm enjoying it.

[robin-swift]

You do git pull from inside the code directory.

So from where your terminal appears to be:

cd libembroidery
git pull

[BartmanEH]

% ./libembroidery/embroider test.pes test.sew
test.pes
Debugging Information
number of colors = 251
number of stitches = 3896
Color: 0 1
Color: 1 1
Color: 2 1
Color: 3 1
Color: 4 1
Color: 5 1
Color: 6 1

[truncated]

same size SEW results file (22KB) and still 'blank' or invalid

[robin-swift]

Huh, now that's surprising.

How similar to your design is

number of colors = 251
number of stitches = 3896

I doubt it really is a 251 thread color pattern.

[BartmanEH]

lol, indeed not: 2 colors

[robin-swift]

and how many stitches?

[BartmanEH]

According to StitchBuddy

[robin-swift]

That's useful. I'm thinking...

[robin-swift]

Ok, so can I see the same hexdump for your input?

xxd test.pes | head

[BartmanEH]

% xxd test.pes | head
00000000: 2350 4553 3030 3031 802f 0100 0100 0100  #PES0001./......
00000010: 0100 ffff 0000 0700 4345 6d62 4f6e 6500  ........CEmbOne.
00000020: 0000 0000 0000 0000 0000 0000 0000 0000  ................
00000030: 0080 3f00 0000 0000 0000 0000 0080 3f00  ..?...........?.
00000040: d80e 4480 eba5 4401 0000 0000 0059 038e  ..D...D......Y..
00000050: 0200 0000 0000 0000 0062 00ff ff00 0007  .........b......
00000060: 0043 5365 7753 6567 0100 1d00 0200 ac01  .CSewSeg........
00000070: b9fe 0400 51ff 0380 0000 1d00 872c 0400  ....Q........,..
00000080: 51ff 0000 52ff 0400 51ff 0000 52ff 0400  Q...R...Q...R...
00000090: 51ff 0000 52ff 0a00 5cff 1500 67ff 2300  Q...R...\...g.#.

[robin-swift]

Ok, so I think I can offer your workflow a solution quicker than fixing this. Then fix this at a later date.

Can you output from your design software in Tajima .dst? Then do the conversion to .sew.

[BartmanEH]

OK, I'll try that but it will take me some time to wrestle Inkstitch into submission, again. Thanks for your support. I'll post my results here hopefully later tonight.

[robin-swift]

I'll open an new issue... use #191