Repository navigation
[Bug]: Inconsistent rendering between backends when rendering Mathtext horizontal rule #23763
Description
Bug summary
When the last item in a Mathtext Vlist is an Hrule, the AGG and SVG backends render the image differently. The difference is not obvious at small font sizes.
To see this happen, we have to dive into Mathtext internals. An easy way is to modify fractions to end with the horizontal rule.
Code for reproduction
import matplotlib.pyplot as plt
import matplotlib._mathtext as _mathtext
plt.rcdefaults()
def _genfrac(self, ldelim, rdelim, rule, style, num, den):
state = self.get_state()
thickness = state.get_current_underline_thickness()
vlist = _mathtext.Vlist([num, _mathtext.Vbox(0, 2 * thickness), _mathtext.Hrule(state, thickness)])
return [_mathtext.Hlist([vlist])]
_mathtext.Parser._genfrac = _genfrac
(fig, ax) = plt.subplots()
plt.text(.5, .3, r'$\dfrac{up}{down}$', size=100, ha='center', va='baseline')
plt.axhline(.3)
plt.savefig('Hrule_at_bottom.png')
plt.savefig('Hrule_at_bottom.svg')
plt.show()Actual outcome
PNG output
SVG output
Expected outcome
Both images should be identical.
Additional information
Why does it matter if it's related to a private module in Matplotlib?
This is one of the things holding up #23616. Vlists ending with an Hrule are not being positioned correctly.
Operating system
Ubuntu 20.04 LTS Focal Fossa on WSL
Matplotlib Version
3.5.0.dev5855+gee2f68bdc9
Matplotlib Backend
WXAgg
Python version
Python 3.8.10
Jupyter version
6.0.3
Installation
git checkout
Activity
github-actions commented on Dec 22, 2025
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!
anntzer commented on Jan 31, 2026
Closed by #30059, I believe. Feel free to reopen if I missed something.
tfpf commented on Feb 1, 2026
Verified! I cannot visually discern any alignment differences in the two images now.
anntzer commented on Feb 1, 2026
@tfpf Do you know if this being fixed has any implications regarding your comment at #31046 (comment)?
(i.e. is the extra +/- still relevant?)
tfpf commented on Feb 1, 2026
My first thought was that it might, but I think I was wrong! I have added the details in that thread.