Control axes padding:xpad & ypad

[grantmcdermott]

Closes #662

This PR adds the new top-level tinyplot(..., x/ypad) args for fine-grained control of the how much padding is added to ends of the axes.

If left as NULL, then behaviour falls back to standard base R conventions, e.g. governed by par("xaxs"). In most cases this translates to a 0.04 (4%) padding exanpsion on each end of the axes.

(Aside: I initially wanted to call them xaxp and yaxp, but those arguments are already reserved by par.)

MWE

pkgload::load_all("~/Documents/Projects/tinyplot")
#> ℹ Loading tinyplot
plt(c(0,1), c(1,1)) # default x/ypad = 0.04

plt(c(0,1), c(1,1), xpad = 0.1)

plt(c(0,1), c(1,1), xpad = 0.5)

Etc.

^{Created on 2026-09-12 with reprex v2.1.1}

FYI @strengejacke.

P.S. I'm also considering changing (increasing) the default padding for categorical axes, but that will come in a follow-up PR.

[Copilot]

🟡 Changes recommended

Flipped boxplots and flipped plots using tpar() defaults currently apply padding to the wrong physical axis.

Get a fresh assessment by requesting another Copilot review.

Pull request overview

Adds configurable x/y-axis padding while preserving base R defaults when unset.

Changes:

• Adds xpad/ypad arguments and tpar() support.
• Applies padding across standard, flipped, logged, and faceted plots.
• Adds documentation and coordinate-based tests.

File summaries

File	Description
R/tpar.R	Registers and validates global padding settings.
R/tinyplot.R	Exposes and propagates padding arguments.
R/tinyAxis.R	Accounts for pre-expanded limits.
R/lim.R	Calculates padded axis limits.
R/flip.R	Swaps padding when axes are flipped.
R/facet.R	Applies padding to facet windows.
NEWS.md	Announces the feature.
man/tpar.Rd	Documents global settings.
man/tinyplot.Rd	Documents public arguments.
man/facet.Rd	Updates internal function usage.
inst/tinytest/test-axis-pad.R	Tests padding and validation.

Review details

Files not reviewed (3)

• man/facet.Rd: Generated file
• man/tinyplot.Rd: Generated file
• man/tpar.Rd: Generated file

Suppressed comments (1)

R/lim.R:63

• This validation accepts Inf, but expand_lim() then returns the unexpanded limits for a non-finite pad while downstream code still forces xaxs/yaxs = "i". Thus xpad = Inf silently produces a tight axis instead of either padding or rejecting the value. Require finite padding values explicitly.

  assert_numeric(xpad, len = 1, lower = 0, null.ok = TRUE, name = "xpad")
  assert_numeric(ypad, len = 1, lower = 0, null.ok = TRUE, name = "ypad")

• Files reviewed: 8/11 changed files
• Comments generated: 5
• Review effort level: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.