bug with lines for large values on x axis when using xlims!

[NilsNiggemann]

When a line is plotted with relatively large values on the x axis it sometimes displays an erronous line that seems to start at x=0. Here is a MWE:

x = [100000.,1000000.] # does not contain 0.
y = [0.,0.]

lines(x,y,color =  :red) # this line shouldnt be visible!
xlims!(0,1)
current_figure()

• Scattering the same data points produces the expected behavior (no data visible in plot)
• The bug depends on the values of x chosen, for example multiplying x by 2 gives correct results (for some reason unclear to me).

---

• [x] are you running newest version (version from docs) ?
• [x] can you reproduce the bug with a fresh environment ? (]activate --temp; add Makie)
• • I am on the current version of CairoMakie v0.10.7
• What platform + GPU are you on?

  Platform Info:
  OS: Linux (x86_64-linux-gnu)
  CPU: 4 × Intel(R) Core(TM) i5-4690 CPU @ 3.50GHz
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-14.0.6 (ORCJIT, haswell)
  Threads: 4 on 4 virtual cores

  -GPU

     *-display                 
     description: VGA compatible controller
     product: Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller
     vendor: Intel Corporation
     physical id: 2
     bus info: pci@0000:00:02.0
     version: 06
     width: 64 bits
     clock: 33MHz
     capabilities: vga_controller bus_master cap_list rom
     configuration: driver=i915 latency=0
     resources: irq:32 memory:f7800000-f7bfffff memory:e0000000-efffffff ioport:f000(size=64) memory:c0000-dffff
  WARNING: output may be incomplete or inaccurate, you should run this program as super-user.

  (I don't have root permissions here, but I don't expect the GPU to be relevant anyway)

[jkrumbiegel]

Interesting, this seems to be a bug in Cairo. Because I checked what positions get computed for the line vertices in CairoMakie and that's projected_positions = Vec{2, Float32}[[7.36f7, 284.0], [7.36f8, 284.0]]. But in the svg that gets rendered, I get d="M 6491184 300 L -2197456 300 " as a path, and the L -2197456 suggests that some integer wraparound has happened. I'm not sure which part of Cairo it is that cannot deal with very large points far outside the canvas boundaries.

[ffreyer]

This specific case is fine for me on breaking-0.21, but it might have just moved to different values