Starting point for fix is likely RendererPgf.draw_path(). The linewidth is available via gc.get_hatch_linewidth(). Now we only need somebody who understands pgf.

Note: mpl.rcParams['hatch.color'] is also broken, but on a more fundamental level. It does not work for other renderers either.

Calling set_rasterized(True) on each bar element (individually) allows for me to use hatching without any issue. Not ideal but a good workaround for me.

Looks like this is fixed by

diff --git i/lib/matplotlib/backends/backend_pgf.py w/lib/matplotlib/backends/backend_pgf.py
index d68c743b7..38589ad65 100644
--- i/lib/matplotlib/backends/backend_pgf.py
+++ w/lib/matplotlib/backends/backend_pgf.py
@@ -487,6 +487,10 @@ class RendererPgf(RendererBase):
                     r"{\pgfqpoint{0in}{0in}}{\pgfqpoint{1in}{1in}}")
             writeln(self.fh, r"\pgfusepath{clip}")
             scale = mpl.transforms.Affine2D().scale(self.dpi)
+            lw = (mpl.rcParams["hatch.linewidth"]
+                  * mpl_pt_to_in * latex_in_to_pt)
+            writeln(self.fh, r"\pgfsetlinewidth{%fpt}" % lw)
             self._print_pgf_path(None, gc.get_hatch_path(), scale)
             self._pgf_path_draw(stroke=True)
             writeln(self.fh, r"\end{pgfscope}")

Feel free to pick up the patch :)

Has this issue been fixed in some later PR ? If yes we should close this

Likely not, otherwise the PR would be linked here. Have you any reason to believe this has been fixed?

The change needed seems to be a few lines , too low for a dedicated PR . If needed I can make a PR and close this issue.

Developer resources are limited. Therefore also seemingly simple fixes may sometimes lie around or fall through the cracks.

You're are very welcome to provide a PR. Please provide a test if possible as well.

This issue has been marked "inactive" because it has been 365 days since the last comment. If this issue is still present in recent Matplotlib releases, or the feature request is still wanted, please leave a comment and this label will be removed. If there are no updates in another 30 days, this issue will be automatically closed, but you are free to re-open or create a new issue if needed. We value issue reports, and this procedure is meant to help us resurface and prioritize issues that have not been addressed yet, not make them disappear. Thanks for your help!

There is an open PR for this.

Just adding enthusiastic support behind this issue being resolved. It's continuing to cause problems.