gerbolyze convert generates... nothing, or a bunch of identical broken files?

[transplier]

I'm trying to get the protoboard generator working again, and I've mostly succeeded, but I'm running into something I haven't managed to chase down in the bowels of gerbolyze: the gerber files made seem to be broken or blank. I need help pinpointing the issue. For an example, I will use a file generated by generate_protoboards.py, called output_protoboards/svg/tht_normal_pitch100mil_large_holes/no_mounting_holes/tht_normal_pitch100mil_large_holes_30x40.svg (referred from now on as proto.svg)

When I run the command that export_protoboards.py would run, I get a pile of tiny, identical gerber files:

(venv) g@hamilton:~/programs/gerbolyze$ python3 -m gerbolyze convert --log-level=info --pattern-complete-tiles-only --use-apertures-for-patterns proto.svg proto_out
Warning (in fontdb:207): Failed to load a font face 0 from '/home/g/.fonts/Sleepy Hollow 3.ttf' cause font doesn't have a family name.
(venv) g@hamilton:~/programs/gerbolyze$ md5sum proto_out/*
ba41d73a010d9e5e62ac383554027b38  proto_out/proto-B.Cu.gbr
ba41d73a010d9e5e62ac383554027b38  proto_out/proto-B.Mask.gbr
ba41d73a010d9e5e62ac383554027b38  proto_out/proto-B.Paste.gbr
ba41d73a010d9e5e62ac383554027b38  proto_out/proto-B.SilkS.gbr
ba41d73a010d9e5e62ac383554027b38  proto_out/proto-Edge.Cuts.gbr
ba41d73a010d9e5e62ac383554027b38  proto_out/proto-F.Cu.gbr
ba41d73a010d9e5e62ac383554027b38  proto_out/proto-F.Mask.gbr
ba41d73a010d9e5e62ac383554027b38  proto_out/proto-F.Paste.gbr
ba41d73a010d9e5e62ac383554027b38  proto_out/proto-F.SilkS.gbr
ba41d73a010d9e5e62ac383554027b38  proto_out/proto.nonplated.drl
ba41d73a010d9e5e62ac383554027b38  proto_out/proto.plated.drl
(venv) g@hamilton:~/programs/gerbolyze$ cat proto_out/proto-B.Cu.gbr 
G04 Gerber file generated by Gerbonara*
%MOMM*%
%FSLAX46Y46*%
%IPPOS*%
G75
%LPD*%
%AMGNC*
1,1,$1,0,0,-57.29578X$4*
1,0,$2,0,0,0*
21,0,$2,$3,0,0,$4X-57.29578*
%
%AMGNR*
21,1,$1,$2,0,0,$5X-57.29578*
1,0,$3,0,0,0*
21,0,$3,$4,0,0,$5X-57.29578*
%
%AMGNO*
21,1,$1,$2,0,0,$5X-57.29578*
1,1,$2,$1/2,0,$5X-57.29578*
1,1,$2,(0-$1)/2,0,$5X-57.29578*
1,0,$3,0,0,0*
21,0,$3,$4,0,0,$5X-57.29578*
%
%AMGNP*
5,1,$2,0,0,$1,$3X-57.29578*
1,0,$4,0,0,0*
%
%ADD10C,0.05*%
M02*

This looks incorrect. I'd at least expect some of the files to be different from each other...

The source svg file is here:

Various dependency versions I've arrived at by trial and error:

(venv) g@hamilton:~/programs/gerbolyze$ pip3 show gerbonara
Name: gerbonara
Version: 0.13.0
Summary: Tools to handle Gerber and Excellon files in Python
Home-page: https://gitlab.com/gerbolyze/gerbonara
Author: jaseg, XenGi
Author-email: gerbonara@jaseg.de
License: 
Location: /home/g/programs/gerbolyze/venv/lib/python3.11/site-packages
Requires: click
Required-by: 
(venv) g@hamilton:~/programs/gerbolyze$ usvg --version
0.27.0
(venv) g@hamilton:~/programs/gerbolyze$ resvg --version
0.27.0
(venv) g@hamilton:~/programs/gerbolyze$ python --version
Python 3.11.4

Could use a pointer or two! Even if it's just a known-good-set of python, usvg, resvg, gerbonara, and whatever else dependency version numbers.

[jaseg]

I suspect that your issue stems from some breakage that happened in usvg's command-line API this year. Try updating both resvg and usvg, their respective latest releases are both v0.36.0.

Gerbolyze right now also has some trouble with more recent KiCad gerbers, expect a fix for that in the next week or two.