Correctly avoid writing to the input file when there are no preserved…

… HTML chunks (#2535) Co-authored-by: Yihui Xie <[email protected]> Co-authored-by: Jiří Moravec <[email protected]>
rstudio · Jan 18, 2024 · 07e2a99 · 07e2a99
1 parent 65916c3
commit 07e2a99
Show file tree

Hide file tree

Showing 4 changed files with 18 additions and 2 deletions.
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,7 +1,7 @@
 Type: Package
 Package: rmarkdown
 Title: Dynamic Documents for R
-Version: 2.25.2
+Version: 2.25.3
 Authors@R: c(
     person("JJ", "Allaire", , "[email protected]", role = "aut"),
     person("Yihui", "Xie", , "[email protected]", role = c("aut", "cre"), comment = c(ORCID = "0000-0003-0645-5666")),

diff --git a/NEWS.md b/NEWS.md
@@ -9,6 +9,8 @@ rmarkdown 2.26
 
 - For the output format option `fig_crop: auto`, it will now use the same logic as in **knitr** to decide if cropping is possible (yihui/knitr#2246).
 
+- Avoid corrupting input files by accident (thanks, @J-Moravec, #2534).
+
 
 rmarkdown 2.25
 ================================================================================

diff --git a/R/html_document_base.R b/R/html_document_base.R
@@ -251,7 +251,12 @@ html_document_base <- function(theme = NULL,
 }
 
 extract_preserve_chunks <- function(input_file, extract = extractPreserveChunks) {
-  input_str <- read_utf8(input_file)
+  # Don't try to modify the input file if it's not .md, otherwise the original
+  # input could be corrupted (#2534). In theory, preserved chunks should only
+  # exist in the intermediate .md file from knit(). If the .md file is not
+  # intermediate but original, this processing should be harmless.
+  if (!xfun::file_ext(input_file) %in% c('md', 'markdown')) return()
+  input_str <- one_string(read_utf8(input_file))
   preserve <- extract(input_str)
   if (!identical(preserve$value, input_str)) write_utf8(preserve$value, input_file)
   preserve$chunks

diff --git a/tests/testthat/test-render.R b/tests/testthat/test-render.R
@@ -32,3 +32,12 @@ test_that("file_scope split correctly input file", {
   expect_snapshot_file(splitted[1])
   expect_snapshot_file(splitted[2])
 })
+
+test_that("Input file is not corrupted", {
+  input <- tempfile()
+  on.exit(unlink(input), add = TRUE, after = FALSE)
+  saveRDS("test", input)
+  # try() is necessary for Pandoc <= 2.14.2
+  try(render(input, quiet = TRUE))
+  expect_equal(readRDS(input), "test")
+})