Bug report: Warning/Error when 'pos' argument is a vector

[NourEdinDarwish]

Hi,

I ran into a small issue when using grid.echo() on a base plot that uses text() with a vectorized pos argument (like pos = c(2, 4)).

It seems to trigger this warning from R:
Warning: number of items to replace is not a multiple of replacement length

Reprex

Here is a minimal example that reproduces it:

library(gridGraphics)
#> Loading required package: grid

# 1. Base plot with mixed text positions (left=2, right=4)
plot(1:5, 1:5, main = "Test")
text(1:5, 1:5, labels = LETTERS[1:5], pos = c(2, 4, 2, 4, 2))

# 2. Convert to grid
# This triggers the warning
grid.echo()
#> Warning in xx[pos == 2] <- xx - convertWidth(offset, "native", valueOnly =
#> TRUE): number of items to replace is not a multiple of replacement length
#> Warning in xx[pos == 4] <- xx + convertWidth(offset, "native", valueOnly =
#> TRUE): number of items to replace is not a multiple of replacement length

^{Created on 2026-01-19 with reprex v2.1.1}

Possible Cause

I looked into R/text.R (around lines 29-32 in C_text), and it looks like the code is trying to update xx using a subset index (e.g. pos == 2), but convertWidth(offset, ...) returns the full vector of offsets.

So it ends up trying to assign the full vector (length 5) into the subset slots (length 3), which causes the mismatch warning.

Suggested Fix

I was able to fix it locally by calculating the converted offsets before the subset assignment, and then subsetting the offset vector to match.

Something like this seems to work:

# Calculate full offsets first
cw_offset <- convertWidth(offset, "native", valueOnly=TRUE)
ch_offset <- convertHeight(offset, "native", valueOnly=TRUE)

# Then apply to the specific indices
if (any(pos == 2)) xx[pos == 2] <- xx[pos == 2] - cw_offset[pos == 2]
if (any(pos == 4)) xx[pos == 4] <- xx[pos == 4] + cw_offset[pos == 4]
if (any(pos == 1)) yy[pos == 1] <- yy[pos == 1] - ch_offset[pos == 1]
if (any(pos == 3)) yy[pos == 3] <- yy[pos == 3] + ch_offset[pos == 3]

Let me know if this makes sense!

[pmur002]

Thanks for the report and the suggested solution. I see the problem and will take a look at the solution.

[pmur002]

A fix has been pushed (package version 0.5-2, not yet submitted to CRAN)