quarto release candidate breaks on an image url

[apsteinmetz]

Bug description

This code renders successfully with Quarto stable

---
title: "example of broken link"
---
![](http://r-statistics.co/screenshots/ggplot_masterlist_2.png){width="488"} [SOURCE](http://r-statistics.co/Top50-Ggplot2-Visualizations-MasterList-R-Code.html)

$ quarto check

[>] Checking Quarto installation......OK
Version: 1.2.475
Path: C:\Program Files\Quarto\bin
CodePage: 1252

[>] Checking basic markdown render....OK

[>] Checking Python 3 installation....(None)

  Unable to locate an installed version of Python 3.
  Install Python 3 from https://www.python.org/downloads/

[>] Checking R installation...........OK
Version: 4.2.1
Path: C:/PROGRA1/R/R-421.1
LibPaths:
- C:/Users/Apste/AppData/Local/R/win-library/4.2
- C:/Program Files/R/R-4.2.1/library
rmarkdown: 2.16

---

It does not with the Quarto release candidate. I think it has something to do with 'deno,' whatever that is.

pandoc
to: html
output-file: temp.html
standalone: true
section-divs: true
html-math-method: mathjax
wrap: none
default-image-extension: png

metadata
document-css: false
link-citations: true
date-format: long
lang: en
title: example of broken link

ERROR: The filename, directory name, or volume label syntax is incorrect. (os error 123), stat 'C:\Users\Apste\Documents\R Projects\http:\r-statistics.co\screenshots\ggplot_masterlist_2.png'

Error: The filename, directory name, or volume label syntax is incorrect. (os error 123), stat 'C:\Users\Apste\Documents\R Projects\http:\r-statistics.co\screenshots\ggplot_masterlist_2.png'
at Object.statSync (deno:runtime/js/30_fs.js:322:9)
at createWalkEntrySync (file:///C:/PROGRA1/Quarto/bin/quarto.js:2754:23)
at expandGlobSync (file:///C:/PROGRA1/Quarto/bin/quarto.js:2874:25)
at expandGlobSync.next ()
at expandGlobs (file:///C:/PROGRA1/Quarto/bin/quarto.js:5786:24)
at resolveGlobs (file:///C:/PROGRA1/Quarto/bin/quarto.js:5862:26)
at resolvePathGlobs (file:///C:/PROGRA1/Quarto/bin/quarto.js:5798:12)
at resolveFileResources (file:///C:/PROGRA1/Quarto/bin/quarto.js:76176:23)
at resourceFilesFromFile (file:///C:/PROGRA1/Quarto/bin/quarto.js:76191:37)
at resourceFilesFromRenderedFile (file:///C:/PROGRA1/Quarto/bin/quarto.js:76185:12)

$ quarto check

[>] Checking versions of quarto binary dependencies...
Pandoc version 3.1.1: OK
Dart Sass version 1.55.0: OK
[>] Checking versions of quarto dependencies......OK
[>] Checking Quarto installation......OK
Version: 1.3.290
Path: C:\Program Files\Quarto\bin
CodePage: 1252

[>] Checking basic markdown render....OK

[>] Checking Python 3 installation....(None)

  Unable to locate an installed version of Python 3.
  Install Python 3 from https://www.python.org/downloads/

[>] Checking R installation...........OK
Version: 4.2.1
Path: C:/PROGRA1/R/R-421.1
LibPaths:
- C:/Users/Apste/AppData/Local/R/win-library/4.2
- C:/Program Files/R/R-4.2.1/library
knitr: 1.42
rmarkdown: 2.16

[>] Checking Knitr engine render......OK

RStudio 2023.03.0+386 "Cherry Blossom" Release (3c53477afb13ab959aeb5b34df1f10c237b256c3, 2023-03-09) for Windows
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) RStudio/2023.03.0+386 Chrome/108.0.5359.179 Electron/22.0.3 Safari/537.36

Checklist

• Please include a minimal, fully reproducible example in a single .qmd file? Please provide the whole file rather than the snippet you believe is causing the issue.
• Please format your issue so it is easier for us to read the bug report.
• Please document the RStudio IDE version you're running (if applicable), by providing the value displayed in the "About RStudio" main menu dialog?
• Please document the operating system you're running. If on Linux, please provide the specific distribution.
• Please provide the output of quarto check so we know which version of quarto and its dependencies you're running.

[cscheid]

Thanks for the report.

I can't reproduce this on macOS:

The error message makes it look like it's a path resolution issue ultimately from resolvePathGlobs in src/core/path.ts. The failure itself seems to be coming from inside deno's expandGlobSync call. I think this is, ultimately, another problem when our filenames include colons on windows.

@cscheid @cderv Is this a regression? If so, do we know what regressed it? I'm wondering if we have an opportunity for adding a missing test somewhere.

[cderv]

Sharing here my findings shared yesterday on slack based on investigation

Doing a git bisect, I found that this is the commit that have make this unhidden: 2f3507a

Before this commit the http resource was transformed to http:\r-statistics.co\screenshots\ggplot_masterlist_2.png (see backslash instead of http://r-statistics.co/screenshots/ggplot_masterlist_2.png and then it passed through resolveGlobs() as a smartGlob and got ** appended. What was going through expandGlobSync() was **/http:\\r-statistics.co\\screenshots\\ggplot_masterlist_2.png which did not caused an issue.

Whereas now the resource url http://r-statistics.co/screenshots/ggplot_masterlist_2.png is passed unmodified to expandGlobSync() and, as it does not support remote https file I believe, the "path" is seen as non absolute, so it gets prepended by root. This leads to a complete unexisting path to be checked e.g. C:\Users\chris\Documents\http:\r-statistics.co\screenshots\ggplot_masterlist_2.png

So @dragonstyle commit did make this problem unhidden, but it was not found until now, because it seems we did not have a test with remote image (which i am surprised by that...)

My proposal is that we filter out http(s) from the resource to expand as glob. As discussed with @dragonstyle, it is probably easier to specifically target this for now, instead of covering all the possible remote protocol to filter out.

However @dragonstyle I am seeing we remove the data URI filtering to deal with it in Lua - We filter out http there...
Is that supposed to prevent the issue and is it not working ? I am referring to 009c833

So @allenmanning we are working on it :)

[cderv]

However @dragonstyle I am seeing we remove the data URI filtering to deal with it in Lua - We filter out http there...
Is that supposed to prevent the issue and is it not working ? I am referring to 009c833

It seems related to preview only based on the comment.

I wondering if the fix should be as you've done in Lua, and just be sure that http resource are not "discovered" and register as resource to handle 🤔

[cderv]

@dragonstyle I think I did not go deep enough yesterday to understand how resource discovery works. My comment above is not correct. We are discovering resources through Lua.

I think we just have an issue in or Lua pattern matching as the regex only targets https and not http. Doing this solve the issue

diff --git a/src/resources/filters/quarto-pre/resourcefiles.lua b/src/resources/filters/quarto-pre/resourcefiles.lua
index 10144d357..d23059428 100644
--- a/src/resources/filters/quarto-pre/resourcefiles.lua
+++ b/src/resources/filters/quarto-pre/resourcefiles.lua
@@ -11,7 +11,7 @@ function resourceFiles()
     -- around this
     Image = function(el)
       local targetPath = el.src
-      if not targetPath:match('^http[s]:') and not targetPath:match('^data:') then
+      if not targetPath:match('^https?:') and not targetPath:match('^data:') then
         -- don't include this resource if it is a URL, data file or some not file path
         if pandoc.path.is_relative(targetPath) then
           local inputDir = pandoc.path.directory(quarto.doc.input_file)

I think that is probably the fix to do here, as in the recent commits 92c9bbf you were purposely trying to ignore remote url as resource it seems (no issue link so I don't have the context)

[dragonstyle]

You are absolutely correct- this is what that pattern was supposed to be doing!