Datasets:

hamel

/

quarto-issues

like 0

Currently, the matplotlib figures are included as `pdf` files in latex, but matplotlib allows to export graphs as `pgf` code with the [`backend_pgf`](https://matplotlib.org/stable/api/backend_pgf_api.html). Using this backend would allow to integrate matplotlib figures using the documents fonts as described in this [blog post](https://jwalton.info/Matplotlib-latex-PGF/). However getting the available `\textwidth` from latex could be tricky.

In version v0.2.306, and earlier, it was legal, and effective, to give a list for the `bibliography` entry in `_quarto.yml`, e.g.: ``` bibliography: [refs1.bib, refs2.bib] ``` In current main, which is v0.2.346, I get this error on `quarto render`: ``` ERROR: Project _quarto.yml validation failed. (line 12, columns 15--37): Field /bibliography must be a string 11: 12: bibliography: [refs1.bib, refs2.bib] ^^^^^^^^^^^^^^^^^^^^^^ 13: ERROR: Error: Project _quarto.yml validation failed. ``` See: https://github.com/matthew-brett/mysite I do find it useful to use more than one bibliography in that entry, and it is allowed in standard Pandoc. Could the validation be relaxed or modified to allow this?

I suspect it is an option clash of some sort. If I create a presentation and change to the following YAML, slides do not have numbers ```yaml --- title: "test" format: revealjs: multiplex: true self-contained: true hash-type: number slide-number: true controls: auto from: markdown+emoji menu: width: full --- ``` When numbers are used (so, not with these options), I saw the first slide is still "title-slide" while I would expect it to be 1 (or 0). Docs https://quarto.org/docs/presentations/revealjs/presenting.html#navigation-options

- When using self-contained: false I get ``` DevTools failed to load SourceMap: Could not load content for file:///home/maelle/Documents/blablabla/presentation_files/libs/revealjs/plugin/multiplex/socket.io.js.map: System error: net::ERR_FILE_NOT_FOUND ``` - When setting it to true I get ``` DevTools failed to load SourceMap: Could not load content for file:///home/maelle/Documents/blablabla/reveal.js.map: System error: net::ERR_FILE_NOT_FOUND DevTools failed to load SourceMap: Could not load content for file:///home/maelle/Documents/blablabla/socket.io.js.map: System error: net::ERR_FILE_NOT_FOUND ``` (from the non speaker file)

1. I create a Quarto presentation from RStudio new file menu. 2. I add ```yaml format: revealjs: multiplex: id: 'ID' secret: 'SECRET' ``` to the YAML. 3. I render. 4. In presentation.html I can see the secret. The [docs](https://quarto.org/docs/presentations/revealjs/presenting.html#multiplex) state > Note that the secret value will be included in only the speaker version of the presentation.

At the moment the PDF export names the PDF file `The Exact Same Words As The Title.pdf`. Should it be transformed to remove space, special characters etc? (And by the way, the PDF export mode is so useful! :pray: )

Running through the "hello world" example for quarto books shows an error message now, whereas it didn't a few weeks ago. Everything still seems to work fine though. ``` quarto create-project mybook --type book cd mybook quarto preview ``` gives output: ``` Creating project at /Users/tomasz/mybook: - Created _quarto.yml - Created .gitignore - Created .gitattributes - Created index.qmd - Created intro.qmd - Created summary.qmd - Created references.qmd - Created cover.png - Created references.bib Preparing to preview [1/4] index.qmd [2/4] intro.qmd [3/4] summary.qmd [4/4] references.qmd Watching files for changes Browse at http://localhost:6733/ ERROR: Error: Executed file not found for book item: intro.qmd GET: / ERROR: Error: Executed file not found for book item: index.qmd GET: /intro.html ERROR: Error: Executed file not found for book item: index.qmd GET: /summary.html ERROR: Error: Executed file not found for book item: index.qmd GET: /references.html ```

In https://quarto.org/docs/output-formats/hugo.html, the resulting md file is named differently from the original qmd file?

Looks like the version number is hard coded in some places, for example here: https://github.com/quarto-dev/quarto-cli/blob/c0ca049fb47210859048525a19a2f065ece496a1/package/scripts/common/quarto#L33-L36 Should these be the version defined here?: https://github.com/quarto-dev/quarto-cli/blob/de78b925c71f682d532307424aa2b457e88ac69f/configuration#L41

I admit this might be a bit niche. I like using this theme: https://github.com/dzello/revealjs-themes#sunblind To get the bold border, > 1. Add 4 divs to your HTML page: ```html <div class="line top"></div> <div class="line bottom"></div> <div class="line left"></div> <div class="line right"></div>+ ``` How do I override the basic template to add those? If I add them from the qmd file, they're not at the right level (they have to come after the div "slides").

Hi! I am super hyped about this project and I am pretty sure a homebrew formula would make it even more popular and easy to install. This is the macOS mirror of #148. I have never tried making a homebrew formula, but I can dig into that and help, of course!

Build fails: ``` Check file:///disk-samsung/freebsd-ports/misc/quarto-cli/work/quarto-cli-0.2.312/src/quarto.ts error: TS2801 [ERROR]: This condition will always return true since this 'Promise<boolean>' is always defined. if (exists(yamlFile)) { ~~~~~~~~~~~~~~~~ at file:///disk-samsung/freebsd-ports/misc/quarto-cli/work/quarto-cli-0.2.312/src/config/metadata.ts:54:9 TS2773 [ERROR]: Did you forget to use 'await'? if (exists(yamlFile)) { ~~~~~~~~~~~~~~~~ at file:///disk-samsung/freebsd-ports/misc/quarto-cli/work/quarto-cli-0.2.312/src/config/metadata.ts:54:9 TS2345 [ERROR]: Argument of type '9' is not assignable to parameter of type 'Signal'. serverLogProcess.kill(9); ^ at file:///disk-samsung/freebsd-ports/misc/quarto-cli/work/quarto-cli-0.2.312/src/execute/jupyter/jupyter-kernel.ts:143:29 TS2801 [ERROR]: This condition will always return true since this 'Promise<string | undefined>' is always defined. } else if (which("git")) { ~~~~~~~~~~~~ at file:///disk-samsung/freebsd-ports/misc/quarto-cli/work/quarto-cli-0.2.312/src/project/project-gitignore.ts:44:14 TS2773 [ERROR]: Did you forget to use 'await'? } else if (which("git")) { ~~~~~~~~~~~~ at file:///disk-samsung/freebsd-ports/misc/quarto-cli/work/quarto-cli-0.2.312/src/project/project-gitignore.ts:44:14 TS2801 [ERROR]: This condition will always return true since this 'Promise<boolean>' is always defined. if (installed()) { ~~~~~~~~~~~ at file:///disk-samsung/freebsd-ports/misc/quarto-cli/work/quarto-cli-0.2.312/src/command/tools/tools/tinytex.ts:98:7 TS2773 [ERROR]: Did you forget to use 'await'? if (installed()) { ~~~~~~~~~~~ at file:///disk-samsung/freebsd-ports/misc/quarto-cli/work/quarto-cli-0.2.312/src/command/tools/tools/tinytex.ts:98:7 TS2339 [ERROR]: Property 'signal' does not exist on type 'typeof Deno'. const streams = signos.map(Deno.signal); ~~~~~~ at https://deno.land/[email protected]/signal/mod.ts:35:35 TS2345 [ERROR]: Argument of type 'unknown' is not assignable to parameter of type 'AsyncIterableIterator<void>'. mux.add(stream); ~~~~~~ at https://deno.land/[email protected]/signal/mod.ts:38:13 TS2571 [ERROR]: Object is of type 'unknown'. stream.dispose(); ~~~~~~ at https://deno.land/[email protected]/signal/mod.ts:44:7 TS2801 [ERROR]: This condition will always return true since this 'Promise<boolean>' is always defined. isPortAvailable({ ^ at https://deno.land/x/[email protected]/mod.ts:182:5 TS2773 [ERROR]: Did you forget to use 'await'? isPortAvailable({ ^ at https://deno.land/x/[email protected]/mod.ts:182:5 TS2801 [ERROR]: This condition will always return true since this 'Promise<boolean>' is always defined. isPortAvailable({ ^ at https://deno.land/x/[email protected]/mod.ts:201:5 TS2773 [ERROR]: Did you forget to use 'await'? isPortAvailable({ ^ at https://deno.land/x/[email protected]/mod.ts:201:5 TS2339 [ERROR]: Property 'Signal' does not exist on type 'typeof Deno'. 'Deno.Signal' is an unstable API. Did you forget to run with the '--unstable' flag? onSignal(Deno.Signal.SIGINT, abend); ~~~~~~ at file:///disk-samsung/freebsd-ports/misc/quarto-cli/work/quarto-cli-0.2.312/src/quarto.ts:83:21 TS2339 [ERROR]: Property 'Signal' does not exist on type 'typeof Deno'. 'Deno.Signal' is an unstable API. Did you forget to run with the '--unstable' flag? onSignal(Deno.Signal.SIGTERM, abend); ~~~~~~ at file:///disk-samsung/freebsd-ports/misc/quarto-cli/work/quarto-cli-0.2.312/src/quarto.ts:84:21 Found 11 errors. *** Error code 1 ``` Command: > deno cache --reload ${WRKSRC}/src/quarto.ts --unstable --importmap=${WRKSRC}/src/import_map.json Version: 0.2.312 deno-1.16.1 FreeBSD 13

I just noticed (from my projects tests) that Quarto no longer writes out the header numbers into the documents that become the inputs to custom filters. Checking with an old version of Quarto (v0.2.183), I see this in the input JSON for one document with a heading: ``` { "t": "Span", "c": [ [ "", [ "header-section-number" ], [] ], [ { "t": "Str", "c": "2.1" } ] ] }, ``` but - this content is missing from the output of current Quarto, though I see the heading number does reach the output HTML. The heading number was useful, because we are using Pandoc filters to write out notebook versions of the pages, and it's convenient to have the heading numbers in the notebook, so the reader can cross-reference with the web or PDF document. Is there any way of getting the heading numbers now? How do they reach the final document?

Hi and first thanks for this awesome project. I have a problem since version [0.2.232](https://github.com/quarto-dev/quarto-cli/commit/fd6d8cd9709c23f38bc76617fab29fc578358033) introduced an error on my metadata parsing: ```markdown --- matter: "NSI" classe: "Terminale" partie: "P1&P4" sectionPrefix: "Partie~" subsectionPrefix: "Exercice~" format: latex: papersize: a5 classoption: - fontsize=10pt - DIV=calc --- # Devoir 3 ``` When i run quarto I go onto this: ``` ERROR: YAMLError: can not read a block mapping entry; a multiline key may not be an implicit key at line 6, column 7: format: ``` However this is valid `yaml` and pandoc can parse it. `pandoc -t json devoir3.qmd | jq '.meta' `.

Subfigures defined in markdown can be given unique captions/labels while belonging to a larger figure with its own caption and label. Can this be done for subfigures produced through computations, such as in the example explored here in the [documentation](https://quarto.org/docs/authoring/figures-and-layout.html#computations)? If so, it might be worth providing some documentation how. If not, it would be a desirable feature.

I use SVG images in my quarto document, like so: ``` {layout-align=center} ``` it works fine for HTML, but for PDF it says: ``` LaTeX Error: Unknown graphics extension: .svg. ``` I know that latex do not support svg, and that the usual workflow is to convert svg to PDF using inkscape or similar. Is there a way to automate this with quarto?

We are using nested divs, and v0.2.261 seems to have broken our build. With the following page: ``` # A title ::: {.callout-note} ::: {.another-div} Some text ::: ::: ``` https://github.com/matthew-brett/mysite `quarto render` gives me: ``` Error running filter /Volumes/zorg/mb312/dev_trees/quarto-cli/src/resources/filters/quarto-pre/quarto-pre.lua: ...cli/src/resources/filters/quarto-pre/../common/table.lua:13: attempt to get length of a Block value (local 'values') stack traceback: ...cli/src/resources/filters/quarto-pre/../common/table.lua:13: in function 'tprepend' .../quarto-cli/src/resources/filters/quarto-pre/callout.lua:236: in function 'calloutLatex' (...tail calls...) ...li/src/resources/filters/quarto-pre/../common/pandoc.lua:122: in function <...li/src/resources/filters/quarto-pre/../common/pandoc.lua:116> ERROR: Error at Object.onComplete (file:///Volumes/zorg/mb312/dev_trees/quarto-cli/src/project/types/book/book-render.ts:245:15) at async renderFiles (file:///Volumes/zorg/mb312/dev_trees/quarto-cli/src/command/render/render.ts:222:12) at async renderProject (file:///Volumes/zorg/mb312/dev_trees/quarto-cli/src/command/render/project.ts:157:25) at async Command.fn (file:///Volumes/zorg/mb312/dev_trees/quarto-cli/src/command/render/cmd.ts:164:24) at async Command.execute (https://deno.land/x/[email protected]/command/command.ts:997:7) at async Command.parse (https://deno.land/x/[email protected]/command/command.ts:918:16) at async Command.parse (https://deno.land/x/[email protected]/command/command.ts:870:16) at async quarto (file:///Volumes/zorg/mb312/dev_trees/quarto-cli/src/quarto.ts:71:3) at async file:///Volumes/zorg/mb312/dev_trees/quarto-cli/src/quarto.ts:89:5 ``` There is no error if I change the initial div to something generic like `{.some-div}`. v0.2.260 gives no error.

``` $ tar xf ../quarto-0.2.291-linux-amd64.tar $ ls bin COPYING.md COPYRIGHT share ``` It is good practice for an archive to start with a folder, otherwise people may run into errors like this one: ``` $ tar xf quarto-0.2.291-linux-amd64.tar tar: .: Cannot utime: Operation not permitted tar: .: Cannot change mode to rwxr-xr-t: Operation not permitted ``` And, if they do it in their home directory, they may pollute that directory instead of having the files extracted into a subfolder.

Could you please consider supporting armv7l and arm64 architectures?, even a script for manual installation on those architectures would be great (or at least one that doesn't use a download of Deno if it is already present, letting us compile it with Cargo ourselves). Just to clarify, my goal with this is to be able to compile RStudio on ARM (Raspberry Pi OS) which latest version requires Quarto

Commit https://github.com/quarto-dev/quarto-cli/commit/b055f78d7d6bf78f88428534b6637a8bfcd5d59c introduced a regression for me (first time using `git bisect`, this is pretty cool!). I am serving a website project with `quarto preview` and when I change a file I get the following output / error: ``` ERROR: Error: Directory not empty (os error 39) at deno:core/01_core.js:106:46 at unwrapOpResult (deno:core/01_core.js:126:13) at Object.opSync (deno:core/01_core.js:140:12) at Object.removeSync (deno:runtime/js/30_fs.js:148:10) at removeIfExists (file:///hits/fast/mbm/share/mbm-wiki/_dev/quarto-cli/src/core/path.ts:23:10) at file:///hits/fast/mbm/share/mbm-wiki/_dev/quarto-cli/src/command/render/freeze.ts:201:23 at Array.map (<anonymous>) at pruneProjectFreezerDir (file:///hits/fast/mbm/share/mbm-wiki/_dev/quarto-cli/src/command/render/freeze.ts:201:9) at freezeLibDir (file:///hits/fast/mbm/share/mbm-wiki/_dev/quarto-cli/src/command/render/project.ts:253:13) at renderProject (file:///hits/fast/mbm/share/mbm-wiki/_dev/quarto-cli/src/command/render/project.ts:258:11) ``` Since I have the quarto repo in the project as well we get the full stack trace for the typescript code. I noticed that after the error occurs there is a rogue `site_libs` folder in the root directory of the project that I assume should have been copied into the `_site` folder and then be delete but this step failed due to some file permissions? When I then cancel with `ctrl+c` and run the `preview` command again the changes propagate and the folder disappears.

I recently encountered a peculiar situation where cached execution in qmd causes error during render. Example: ```` --- title: Test toc: false --- ```{python} print("HELLO") ``` ```` Running quarto render produces the following error: The execution_count just keeps incrementing.  Meanwhile, specifying: ``` execute: cache: false ``` produces no error. `cache-refresh` does not affect either run.

When whe try to generate a pdf from quatro's default we get error on generating because the Blue color is not definined. This is due to the hypersetup's option in which the `citecolor` and the `urlcolor` are `Blue`, because LaTeX can be case sensitive on some systems `blue` must be used.

Use of callout's type `tip` is not possible in PDF. The `\faLightbulbO` command which is used in the `tcolorbox` defintion is a fontawesome's package command but with the fontawesome5 package the `\faLightbulb[regular]` command must be used.

From @cscheid: I think I just ran into a bug with quarto preview, the python daemon, and an admittedly funky python chunk. If you have a python cell that runs a long-running subprocess.call , and you ctrl-c the preview while it’s running, the subprocess keeps going. I don’t know if there’s a not-horrible way to fix it, but it felt bad to have to manually kill -9 some stuff.

Not sure if this is intentional: doing some tests for https://github.com/quarto-dev/quarto-cli/issues/175 I saw the value of `lang` as seen by a Lua filter is inconsistent with vanilla pandoc (the same happens for other metadata variables). The examples below use the following `index.qmd`: ```markdown --- lang: fr --- Text. ``` and the following `checklang.lua` filter: ```lua local function dump(o, level) level = (level or 0) + 1 indent = ' ' if type(o) == 'table' then local s = '{\n' for k,v in pairs(o) do if type(k) ~= 'number' then k = '"'..k..'"' end s = s .. indent:rep(level) .. '['..k..'] = ' .. dump(v, level) .. ',\n' end return s .. indent:rep(level-1) .. '}' else return tostring(o) end end function Meta(meta) print('@@@ meta.lang:', dump(meta.lang)) return nil end ``` **Observations** 1. Running `pandoc index.qmd -L checklang.lua` shows: ``` @@@ meta.lang: { [1] = { ["text"] = fr, }, } ``` 1. Running `quarto render index.qmd -L checklang.lua` shows ``` @@@ meta.lang: { [1] = Str "fr", } ``` Now I remove the metadata header in `index.qmd` and use `-M lang` instead: 1. Running `pandoc index.qmd -L checklang.lua -M lang=fr` shows: ``` @@@ meta.lang: fr ``` 1. Running `quarto render index.qmd -L checklang.lua -M lang=fr` shows the same thing: ``` @@@ meta.lang: fr ``` So pandoc itself is not consistent (maybe that's expected?) but quarto introduces a third way of storing the metadata value. This looks like something that will be annoying for filter writers?

With quarto 0.2.232 I could set the `lang` metadata by calling `quarto render -M lang=fr` (see https://github.com/quarto-dev/quarto-cli/issues/146#issue-1033671251 for a description of the use case). This doesn't work anymore in 0.2.258. For example with ``` # File _quarto.yml project: type: book book: title: Title chapters: - index.qmd filters: - checklang.lua ``` ``` -- File checklang.lua function Meta(meta) print('@@@ Metadata lang:', meta.lang) return nil end ``` Calling `quarto render -M lang=fr` with the older version would print `@@@ Metadata lang: fr`. Now it prints `@@@ Metadata lang: nil`.

I see we can use and render R codes or Python codes in separeate qmd files. Can we use R and Python codes in a single qmd file? Will we need to use 'reticulate' in qmd file?

It will benefit the academic community if you can elaborate on how to produce academic articles and theses using quarto book project like templates included in 'rticles', 'papaja' and 'thesisdown' . It is also very useful if we can adapt these templates in quarto because quarto is more userfriendky and versatile in choice of programming language used in data analysis.

Let's say I want to create this presentation ````markdown --- title: Speaker Note format: revealjs --- # Content --- Slide without title # Another --- Slide without title ```` This should create a presentation with 1 title slide two section title slides which each section using one slide. However, there is an issue with YAML parsing by Quarto ```` ERROR: YAMLError: can not read a block mapping entry; a multiline key may not be an implicit key at line 8, column 1: # Another ^ at generateError (https://deno.land/[email protected]/encoding/_yaml/loader/loader.ts:163:10) at throwError (https://deno.land/[email protected]/encoding/_yaml/loader/loader.ts:176:9) at readBlockMapping (https://deno.land/[email protected]/encoding/_yaml/loader/loader.ts:1187:16) at composeNode (https://deno.land/[email protected]/encoding/_yaml/loader/loader.ts:1509:13) at readDocument (https://deno.land/[email protected]/encoding/_yaml/loader/loader.ts:1695:3) at loadDocuments (https://deno.land/[email protected]/encoding/_yaml/loader/loader.ts:1757:5) at load (https://deno.land/[email protected]/encoding/_yaml/loader/loader.ts:1787:21) at parse (https://deno.land/[email protected]/encoding/_yaml/parse.ts:19:10) at readYamlFromMarkdown (file:///C:/Users/chris/DOCUME~1/DEV_R/QUARTO~2/src/core/yaml.ts:67:22) at readYamlFromMarkdownFile (file:///C:/Users/chris/DOCUME~1/DEV_R/QUARTO~2/src/core/yaml.ts:78:10) ```` The parser should take into account `---` as slide delimiter, right ?

Web outputs include navigational elements that are always in English, for example the title "On this page" of the right margin menu, the "Footnotes" header at the bottom of pages, and various strings in the search bar GUI. It would be great to have these translated according to the `lang` variable (as already done by babel for the PDF format).

If I were to use the Julia or R IPython kernels in Jupyter Notebooks, would that work or is notebook support specific to Python?

I love building my workflow above quarto,I used it to write my paper,.convert my markdown to pdf,word or anything need when submitting.It is a great work. However,I am now thinking use quarto to build my own Academic blog website(just like https://hugo-blog-theme.netlify.app/ ).I have known that quarto can also build a website.But is it have a proper way to build the blog like hugo or something else.

Tested with 0.2.222. Some of my testing has had me temporarily creating Quarto project directories within the directory structure of a large Git repository. Quarto can run `git status` when attempting to determine if it should write a `.gitignore` file. This happens any time a project is rendered and does not already have a `.gitignore` in place. Running `git status` can be slow if the access to the working directory is not snappy. This can happen when the filesystem is networked or, in my case, macOS filesystem mounted into a Docker container. Consider replacing `git status` with `git rev-parse`, which is much lighter weight and can still confirm if the project is within a Git repository. Here is some timing taken from within my Docker container: ```bash # initial render without .gitignore time quarto render #> real 0m14.693s #> user 0m3.648s #> sys 0m1.761s # subsequent render (having .gitignore) time quarto render #> real 0m2.037s #> user 0m2.022s #> sys 0m0.680s # now, timing of some git commands time git status #> real 0m11.965s #> user 0m1.521s #> sys 0m1.037s time git rev-parse #> real 0m0.021s #> user 0m0.002s #> sys 0m0.003s ```

Documentation indicates you can use quarto preview for both HTML and PDF output of both books and individual documents. But how do you choose which, particularly when your YAML includes options for either format? Hard to find this guidance.

After pulling the updates yesterday, the background images from a local directory for slides are no longer copied into the user specified output directory for a project. The images were rendered after manually copying the images into the output directory. Local background image are rendered well for non-project qmd files; it also works if the image url is an external link, even when the qmd is inside a project. It only fails when the background image is from a local directory, and the qmd file is inside a project with a user specified output directory.

In the latest Quarto (v0.2.232), `Pandoc` v2.14.2 is bundled. Users may want to render through different versions of `Pandoc` on system at their risk. Do you have any options like `rmarkdown::find_pandoc()`?

As of 0.2.222, providing either `-h` or `--help` to the `render` command ends up passing that option to pandoc and then presenting the pandoc help. If the document contains Shiny components or other executable code, that code is run before calling pandoc. For comparison, Git and Go take different approaches to the same problem: * `git pull -h` - shows usage information and an option summary. * `git help pull` - shows the full man page for `git pull`, including details for each option. * `go build -h` - shows usage line and instructions to use `go help build`. * `go help build` - detailed information about the `go build` command.

Do you have any plans to develop VSCode extension tailored for Quarto? Please correct me if I am wrong, but currently, users have to manually type required commands in terminal within VSCode. Another idea is to support qmd files via [REditorSupport/vscode-R](https://github.com/REditorSupport/vscode-R).

For terminal users on Windows, could you please consider supporting [Chocolatey](https://chocolatey.org/) installation? This would help streamline the cli installation on Windows.

Is there a way to set `output-dir` from the command line? I tried with `-M output-dir=...` but it doesn't seem to work. The reason I'm asking is that I have a book in two languages, with both languages sharing the same source files (so they can share non-text elements). For example a source can look like this: ``` [Some text]{lang=fr} [Some text]{lang=de} $some shared math equation$ ... ``` For rendering, a Makefile runs quarto once with `-M lang=fr` and once with `-M lang=de`. A Lua filter is used to keep only the text for the selected language. To keep things cleanly separated it would be nice to have each run working in a different directory (e.g. `_book/fr` and `_book/de`).

I spent ages today trying to get quarto to produce a PDF with A4 paper. The documentation was all very clear, but it didn't work and it took me a long time to realise there was a formatting error in the yaml. With the error in the yaml text, the whole yaml block was silently ignored, so I didn't know there was an error and hence couldn't understand why it was being ignored. Example document: ```markdown --- title: "Test Document" author: "Me" format: pdf: documentclass: article papersize: a4 crossref: title-delim: : --- # Title Text ``` Note that `title-delim` is set to (the default value of) `:`, which without any quotation marks is a yaml error. If you pass this to quarto, it does this: ``` > quarto render .\test_2.qmd pandoc to: html from: markdown output-file: test_2.html standalone: true section-divs: true html-math-method: mathjax filters: - crossref metadata document-css: false link-citations: true Output created: test_2.html ``` Note that the output isn't a PDF and there's no error message to say that the yaml was ignored.

The python library sympy uses latex as one of its output formats. It would be lovely if there was a way to integrate this with Quarto such that sympy spits out raw latex and Quarto passes it onto pandoc as latex. For example, consider this `.qmd` document: ````markdown --- title: "Test Document" author: "Me" format: pdf: documentclass: article papersize: a4 code-background: true geometry: - margin=20mm --- # Section 1 Just some basic text to start it off. A beautifully rendered equation: $$ a = \frac{2 \pi f_c}{\lambda_r} \sin(\theta - \alpha) $$ Now we load sympy... ```{python} #| output: false from sympy import * init_printing() var('a, f_c, lambda_r, theta, alpha') ``` Then we can use sympy to create the equation and work with it. ```{python} a_eq = (2 * pi * f_c) * sin(theta - alpha) / lambda_r Eq(a, a_eq) ``` but it looks horrible! It's useful because we can do integration etc all in the document: ```{python} integrate(a_eq, theta) ``` We can change printing style, but it doesn't really help: ```{python} print(latex(Eq(a, a_eq))) ``` It would be great if I could add `#| output=latex` or something so that it gets put in the document and processed by latex. ```` This is what it looks like: 

Currently this type of document will output resulting text without comment ````markdown --- title: "Title" format: html: default --- ```{r} 1 + 1 ``` ````  This is different than with R Markdown but it stay readable. For reference, we have that with Rmd  But when we add `collapse = TRUE`, I think we should add back a comment line. We get this with Quarto:  And this with rmarkdown  In quarto, source code is not really differentiable from resulting line. This is causing issue with highlighting too as with `collapse = TRUE`, everything is in one chunk. What about setting `comment = "#>"` or `comment = "##"` using an option hook when `collapse = TRUE` ? I can do a PR if that is ok. Otherwise, we should maybe rethink the way `collapse = TRUE` is supported or not for Quarto.

First of all, thank you for the remarkable work around quarto and for your responsiveness. It is really an extremely useful tool. I am setting up a scientific journal that relies on quarto for the final production of documents (working with NelleV by the way, so thank a lot for the ipynb output from Issue #133, this helps a lot). I was wondering if there was a way to use the metadata defined in the qmd header or in the yaml files (_variables.yml or _quarto.yml) in some HTML files integrated by the include-before-body command (like shortcodes). It seems complicated to me, but I'm just asking. Or a way to get a similar result. The idea would be to be able to reuse metadata (e.g. doi, affiliation, github repo address, etc.) for the definition of the banner of the final HTML document. Thanks.

This is probably a bit more of an RStudio integration issue, but I wasn't sure of a better place to open this. I've been playing around with Quarto a bit, and it has been absolutely fantastic. For the most part, the experience moving from `.Rmd` has been pretty smooth. However, one thing that I really miss is the ability to use tab-completion for chunk options. I can never remember the full set of chunk options, so going between the curly brackets and hitting tab was my go-to move. This still works if you go into the curly brackets, but it seems like the `#|` is going to be the best practice moving forward, and having some tab-completion functionality would make it a lot more accessible.

For example given the following `test.md`: ``` # Test a $$x$$ b. ``` calling `quarto render test.md` generates an html page without content (just the title). Using pandoc directly with `pandoc test.md -s -o test.html` gives a correct output.

It would be nice to support executing several languages in the same file. My use case is that I'm using "panel-tabsets" to show solutions implemented in different languages (e.g. Octave and Python). Currently only one of the tabs can show the result of running of the code.

Our book needs a small suite of R code executed at the beginning of each page, in our case called `_common.R`. At the moment we are doing this by putting this chunk at the top of each page, as suggested in #38: ~~~ ```{r setup, include=FALSE} source("_common.R") ``` ~~~ This issue is to ask for a way to specify such a file to run by default. Bookdown allows you to do this using the `before_chapter_script` YaML configuration option : https://bookdown.org/yihui/bookdown/configuration.html

Hello, Collaborators and I are using quarto for reproducible scientific publication purposes. As part of our pipeline, we would like not only to compile to HTML/PDF, but also to have a jupyter-notebook version (in order to then link to binder). I believe execution of the code is already performed using the jupyter ecosystem. Is it possible to add as output format a jupyter-notebook? Many thanks, Nelle

Is there an easy way to update my quarto client? Now I have to reinstall it when it is updated.

Curious if there is likely to be some support/documentation for emacs [polymode](https://polymode.github.io/) at some point? (That's my current workflow for Rmarkdown files: see also [here](https://www.thetopsites.net/article/59750668.shtml) Or maybe it already works with the right definitions ... ? (I am *not* an Emacs guru, so I would be cargo-culting my way through trying to make it work ...)

I have run into a strange error rendering in Quarto that appears to be a scope problem. I do not see the same error when rendering in `rmarkdown::render`. Consider the following document (also at <https://github.com/matthew-brett/quarto-test-book/blob/main/intro.Rmd>): ~~~ ```{r} source('setup.R') # Contains only the following function: # get_var <- function(name) { # return (get(name)) # } ``` ```{r} # Redefining the exact same function here fixes error below. # get_var <- function(name) { # return (get(name)) # } ``` ```{r} p <- 1 ``` ```{r} # No error get('p') ``` ```{r} # Error here: # Error in get(name) : object 'p' not found # Calls: .main ... withCallingHandlers -> withVisible -> eval -> eval -> get_var -> get get_var('p') ``` ~~~ See the noted error in the last chunk, generated by `quarto render`. There is no error for `Rscript -e "rmarkdown::render('intro.Rmd')"`, and I get the expected output (`[1] 1`). Is there something different about the scope or environment when executing via Quarto?

Quarto looks very neat. Is it possible to contract the experts to do a half-day(?) workshop?

The help for `quarto render` mentions that extra arguments can be given to forward to pandoc. And for `quarto preview` it says > You can also include arbitrary command line arguments to be forwarded to quarto render. This made me hope the following would work: ``` quarto preview --default-image-extension=svg ``` but it says `Error: Unknown option "--default-image-extension"`. For `quarto render` I get the same error, but I can get it to work with e.g. `quarto render --to html --default-image-extension=svg`. I couldn't find a workaround to make it work for `quarto preview`. (Thanks for this tool, it is truly awesome)

Use case: I needed to uninstall tinytex install by Quarto and I wondered how to do that. + So I ran ```sh $ quarto help tools Usage: quarto tools [command:string] [name:string] Version: 0.2.188 Description: Manages quarto tools. Tools that can be installed include: tinytex Options: -h, --help - Show this help. --log <level> - Path to log file --log-level <level> - Log level (info, warning, error, critical) --log-format <format> - Log format (plain, json-stream) --quiet - Suppress console output. Commands: help [command] - Show this help or the help of a sub-command. Examples: Install TinyTex: quarto tools install tinytex ``` `quarto tools --help` gives the same result. There is no mention of available commands. Only an example for install. I would expect the `Commands:` part to contain the available commands for `quarto tools` that I can use. For my need, I tried `quarto tools uninstall tinytex` and it works ! Wouldn't it be interesting to document available commands ? That would help know what is possible. From here: https://github.com/quarto-dev/quarto-cli/blob/4c69e93457a49c6530b67550dc46c75d45b44294/src/command/tools/cmd.ts It seems we can also `update`, and `list` (but the latter is the default action)

Currently Quarto is distributed as a binary, packaged for Debian, Windows and Mac systems. Since the releases are updated very often, would it be possible to offer delta updates that do not neccessarily couple the runtime, in case it's not updated? This is natively implemented in RPM and AppImage, and could speed up the update process, while reducing the download size and thus sparing bandwith, and ultimately physical resources. --- Found via #121

Quarto is a pretty exciting project with many new features that greatly enhance the scientific writing of markdown. However, the amount of templates(especially in journal) much smaller Rmarkdown. I would like to ask if it is possible to migrate Rmarkdown templates to it, and if you have any idea to increase the number of templates.😉

The search index contains page JavaScript.    The text that is same for each is: ``` window.document.addEventListener(\"DOMContentLoaded\", function (event) {\n const icon = \"\";\n const anchorJS = new window.AnchorJS();\n anchorJS.options = {\n placement: 'right',\n icon: icon\n };\n anchorJS.add('.anchored');\n const clipboard = new window.ClipboardJS('.code-copy-button', {\n target: function(trigger) {\n return trigger.previousElementSibling;\n }\n });\n clipboard.on('success', function(e) {\n // button target\n const button = e.trigger;\n // don't keep focus\n button.blur();\n // flash \"checked\"\n button.classList.add('code-copy-button-checked');\n setTimeout(function() {\n button.classList.remove('code-copy-button-checked');\n }, 1000);\n // clear code selection\n e.clearSelection();\n });\n function tippyHover(el, contentFn) {\n window.tippy(el, {\n allowHTML: true,\n content: contentFn,\n maxWidth: 500,\n delay: 100,\n interactive: true,\n interactiveBorder: 10,\n theme: 'quarto',\n placement: 'bottom-start'\n }); \n }\n const noterefs = window.document.querySelectorAll('a[role=\"doc-noteref\"]');\n for (var i=0; i<noterefs.length; i++) {\n const ref = noterefs[i];\n tippyHover(ref, function() {\n const id = new URL(ref.getAttribute('href')).hash.replace(/^#/, \"\");\n const note = window.document.getElementById(id);\n return note.innerHTML;\n });\n }\n var bibliorefs = window.document.querySelectorAll('a[role=\"doc-biblioref\"]');\n for (var i=0; i<bibliorefs.length; i++) {\n const ref = bibliorefs[i];\n const cites = ref.parentNode.getAttribute('data-cites').split(' ');\n tippyHover(ref, function() {\n var popup = window.document.createElement('div');\n cites.forEach(function(cite) {\n var citeDiv = window.document.createElement('div');\n citeDiv.classList.add('hanging-indent');\n citeDiv.classList.add('csl-entry');\n var biblioDiv = window.document.getElementById('ref-' + cite);\n if (biblioDiv) {\n citeDiv.innerHTML = biblioDiv.innerHTML;\n }\n popup.appendChild(citeDiv);\n });\n return popup.innerHTML;\n });\n }\n}); ```

Building `qmd` files through Quarto does not work here, as modules are looked up in the wrong location. We see an error that reads: ```shell # yala @ pons in ~/src/lab.libreho.st/libernetes/documentation on git:main x [10:36:34] C:1 $ quarto check [✓] Checking Quarto installation......OK Version: 0.2.183 Path: /home/yala/Downloads/quarto-0.2.183/bin [✓] Checking basic markdown render....OK [✓] Checking Python 3 installation....OK Version: 3.9.7 Path: /usr/bin/python3 Jupyter: 4.7.1 Kernels: python3 (|) Checking Jupyter engine render.... An error occurred while executing the following cell: ------------------ # imports import os import sys import types import json # figure size/format fig_width = 7 fig_height = 5 fig_format = 'retina' fig_dpi = 96 # matplotlib defaults / format try: import matplotlib.pyplot as plt plt.rcParams['figure.figsize'] = (fig_width, fig_height) plt.rcParams['figure.dpi'] = fig_dpi plt.rcParams['savefig.dpi'] = fig_dpi from IPython.display import set_matplotlib_formats set_matplotlib_formats(fig_format) except Exception: pass # plotly use connected mode try: import plotly.io as pio pio.renderers.default = "notebook_connected" except Exception: pass # enable pandas latex repr when targeting pdfs try: import pandas as pd if fig_format == 'pdf': pd.set_option('display.latex.repr', True) except Exception: pass # output kernel dependencies kernel_deps = dict() for module in list(sys.modules.values()): # Some modules play games with sys.modules (e.g. email/__init__.py # in the standard library), and occasionally this can cause strange # failures in getattr. Just ignore anything that's not an ordinary # module. if not isinstance(module, types.ModuleType): continue path = getattr(module, "__file__", None) if not path: continue if path.endswith(".pyc") or path.endswith(".pyo"): path = path[:-1] kernel_deps[path] = os.stat(path).st_mtime print(json.dumps(kernel_deps)) # reset state %reset def ojs_define(**kwargs): import json from IPython.core.display import display, HTML # do some minor magic for convenience when handling pandas # dataframes def convert(v): try: import pandas as pd except ModuleNotFoundError: # don't do the magic when pandas is not available return v if type(v) == pd.DataFrame: j = json.loads(v.T.to_json(orient='split')) return dict((k,v) for (k,v) in zip(j["index"], j["data"])) else: return v v = dict(contents=list(dict(name=key, value=convert(value)) for (key, value) in kwargs.items())) display(HTML('<script type="ojs-define">' + json.dumps(v) + '</script>'), metadata=dict(ojs_define = True)) globals()["ojs_define"] = ojs_define ------------------ --------------------------------------------------------------------------- FileNotFoundError Traceback (most recent call last) <ipython-input-1-b55edcc50804> in <module> 53 if path.endswith(".pyc") or path.endswith(".pyo"): 54 path = path[:-1] ---> 55 kernel_deps[path] = os.stat(path).st_mtime 56 print(json.dumps(kernel_deps)) 57 FileNotFoundError: [Errno 2] No such file or directory: '/usr/lib64/python3.9/pydoc_data/topics.py' FileNotFoundError: [Errno 2] No such file or directory: '/usr/lib64/python3.9/pydoc_data/topics.py' [✓] Checking Jupyter engine render....OK # yala @ pons in ~/src/lab.libreho.st/libernetes/documentation on git:main x [10:36:44] C:1 $ pipenv shell Launching subshell in virtual environment... . /home/yala/.local/share/virtualenvs/documentation-EDOFPCMD/bin/activate # yala @ pons in ~/src/lab.libreho.st/libernetes/documentation on git:main x [10:37:14] $ . /home/yala/.local/share/virtualenvs/documentation-EDOFPCMD/bin/activate (documentation) # yala @ pons in ~/src/lab.libreho.st/libernetes/documentation on git:main x documentation-EDOFPCMD [10:37:14] $ quarto check [✓] Checking Quarto installation......OK Version: 0.2.183 Path: /home/yala/Downloads/quarto-0.2.183/bin [✓] Checking basic markdown render....OK [✓] Checking Python 3 installation....OK Version: 3.9.7 Path: /home/yala/.local/share/virtualenvs/documentation-EDOFPCMD/bin/python3 Jupyter: 4.8.1 Kernels: python3, zsh (-) Checking Jupyter engine render.... An error occurred while executing the following cell: ------------------ # imports import os import sys import types import json # figure size/format fig_width = 7 fig_height = 5 fig_format = 'retina' fig_dpi = 96 # matplotlib defaults / format try: import matplotlib.pyplot as plt plt.rcParams['figure.figsize'] = (fig_width, fig_height) plt.rcParams['figure.dpi'] = fig_dpi plt.rcParams['savefig.dpi'] = fig_dpi from IPython.display import set_matplotlib_formats set_matplotlib_formats(fig_format) except Exception: pass # plotly use connected mode try: import plotly.io as pio pio.renderers.default = "notebook_connected" except Exception: pass # enable pandas latex repr when targeting pdfs try: import pandas as pd if fig_format == 'pdf': pd.set_option('display.latex.repr', True) except Exception: pass # output kernel dependencies kernel_deps = dict() for module in list(sys.modules.values()): # Some modules play games with sys.modules (e.g. email/__init__.py # in the standard library), and occasionally this can cause strange # failures in getattr. Just ignore anything that's not an ordinary # module. if not isinstance(module, types.ModuleType): continue path = getattr(module, "__file__", None) if not path: continue if path.endswith(".pyc") or path.endswith(".pyo"): path = path[:-1] kernel_deps[path] = os.stat(path).st_mtime print(json.dumps(kernel_deps)) # reset state %reset def ojs_define(**kwargs): import json from IPython.core.display import display, HTML # do some minor magic for convenience when handling pandas # dataframes def convert(v): try: import pandas as pd except ModuleNotFoundError: # don't do the magic when pandas is not available return v if type(v) == pd.DataFrame: j = json.loads(v.T.to_json(orient='split')) return dict((k,v) for (k,v) in zip(j["index"], j["data"])) else: return v v = dict(contents=list(dict(name=key, value=convert(value)) for (key, value) in kwargs.items())) display(HTML('<script type="ojs-define">' + json.dumps(v) + '</script>'), metadata=dict(ojs_define = True)) globals()["ojs_define"] = ojs_define ------------------ --------------------------------------------------------------------------- FileNotFoundError Traceback (most recent call last) /tmp/ipykernel_10306/4071903457.py in <module> 53 if path.endswith(".pyc") or path.endswith(".pyo"): 54 path = path[:-1] ---> 55 kernel_deps[path] = os.stat(path).st_mtime 56 print(json.dumps(kernel_deps)) 57 FileNotFoundError: [Errno 2] No such file or directory: '/usr/lib64/python3.9/pydoc_data/topics.py' FileNotFoundError: [Errno 2] No such file or directory: '/usr/lib64/python3.9/pydoc_data/topics.py' Exception ignored in: <function BaseEventLoop.__del__ at 0x7fbc9c1b30d0> Traceback (most recent call last): File "/usr/lib64/python3.9/asyncio/base_events.py", line 683, in __del__ self.close() File "/usr/lib64/python3.9/asyncio/unix_events.py", line 58, in close super().close() File "/usr/lib64/python3.9/asyncio/selector_events.py", line 92, in close self._close_self_pipe() File "/usr/lib64/python3.9/asyncio/selector_events.py", line 99, in _close_self_pipe self._remove_reader(self._ssock.fileno()) File "/usr/lib64/python3.9/asyncio/selector_events.py", line 277, in _remove_reader key = self._selector.get_key(fd) File "/usr/lib64/python3.9/selectors.py", line 191, in get_key return mapping[fileobj] File "/usr/lib64/python3.9/selectors.py", line 72, in __getitem__ fd = self._selector._fileobj_lookup(fileobj) File "/usr/lib64/python3.9/selectors.py", line 226, in _fileobj_lookup return _fileobj_to_fd(fileobj) File "/usr/lib64/python3.9/selectors.py", line 42, in _fileobj_to_fd raise ValueError("Invalid file descriptor: {}".format(fd)) ValueError: Invalid file descriptor: -1 [✓] Checking Jupyter engine render....OK ``` The task `[✓] Checking Jupyter engine render` can therefore also not be considered to be `....OK`. The same error appearing in both virtual environments suggests there is more at work here that I don't comprehend at the moment.

I have a shortcode `myvar` defined as "my variable" in `_variables.yml`. This works as I would expect for the top-level index file, and in sub-headings, but not in the top-level heading for a not-top-level file, where the shortcode results in the page title being truncated to the part before the shortcode - "Before shortcode" in the example below, instead of "Before short code my variable after". This is the text of the input `.qmd` page: ~~~ # Before shortcode {{< var myvar >}} after Start my text. ## Sub-heading with {{< var myvar >}} inside And some text. ~~~ If the heading starts with the short code, Quarto ignores that heading, and uses the second-level heading instead. <https://github.com/matthew-brett/quarto-test-book>

There's a peculiar behaviour with `quarto render` where repeated use of the command create another output directory (for resources) inside the pre-existing output directory. E.g., 1st round of using `quarto render`: Inside `docs/`(output-dir): index.html about.html subdirectory/index.html subdirectory/another.html subdirectory/resources/a.png subdirectory/resources/b.png 2nd round: Inside `docs/`(output-dir): index.html about.html subdirectory/index.html subdirectory/another.html subdirectory/resources/a.png subdirectory/resources/b.png **docs/subdirectory/resources/a.png** **docs/subdirectory/resources/b.png** (The last 2 are recreated multiple times further nesting in the output-dir)

Hello, there's possibly a bug in the sitemap generation. (or perhaps I configured the _quarto.yml incorrectly). One of the slash after the `http://` seem to have been removed. Sitemap result: ```xml <?xml version="1.0" encoding="UTF-8"?> <urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"> <url> <loc>https:/ludwighoon.github.io/about.html</loc> <lastmod>2021-09-29T15:05:26.300Z</lastmod> </url> <url> <loc>https:/ludwighoon.github.io/index.html</loc> <lastmod>2021-09-29T15:06:46.730Z</lastmod> </url> ``` What I defined in _quarto.yml: ```yml project: type: site output-dir: docs site: site-url: https://ludwighoon.github.io repo-url: https://github.com/ludwigHoon/ludwighoon.github.io repo-actions: [edit, issue] page-navigation: true .... ```

In the `render targets` yaml it is possible to specify targets using wild cards (https://quarto.org/docs/projects/quarto-projects.html#render-targets). This would also be very handy for specifying content layouts in the sidebar, e.g. ``` sidebar: - id: hello style: floating title: Hello contents: - section: Hello contents: - hello/*.md ```

As Vscode is gaining so much popularity (for obvious reasons) is it planned to enhance vscode's support as it is for Rstudio ? The work done with Rstudio looks very nice, quarto could grow its user base with equivalent features for vscode. Anyway, you've made a very good work with quarto. Thanks !

The default pandoc html template can take an abstract text from the metadata, but for quarto this currently only works for pdf and docx output, and there it omits the "abstract" heading. This would be key for things like research papers (see https://github.com/quarto-dev/quarto-cli/issues/111).

I have just found out the updated version of the Quarto website, nice effort. This is only a small suggestion - the "Projectys" tab could contain one more entry - the very basic use case why this project is of interest - a simple equivalent of a quarto paper. By that I mean the set up of a project equivalent of a Latex "article" class (no chapters and complex elements like TOC) showing a "reproducible report" setup. An amazing bonus would be to show the equivalent of what is possible with RMD where a Latex template can be specified. I am not sure whether this is even possible ( yet), but especially in the Python world, the ability to transform a ipynb to a camera-ready result based on a Latex class template is a massive gap in capabilities.

Hi there and thank you for your amazing work! My issue might be out of scope but I would be over the moon if it worked. I am using quarto to build an internal wiki for my research group. We have a shared file system with multiple (linux) users from which I am serving the site through quarto preview as a systemd service. When I, the user who started the service, change a file, the website updates almost instantly. But when another user changes a file, quarto seems to not recognize this change, but I does appear once I open and save the file myself. Is there some way to make quarto "more alert" to file changes?

On frameworks like mdBook, the user can toggle the table of contents menu and view the reading material out in wide if they wish to. Here's an example of that from the Rust language's book:  One can also control the width of the table of contents menu using the mouse, as well. On quarto, it looks such a feature hasn't been implemented yet and would be lovely to have.

cc @dragonstyle as discussed. Here is what I have tried ````powershell # gh auth must be setup gh release download --repo quarto-dev/quarto-cli --pattern '*.msi' v0.2.144 # will install interactively with user interaction msiexec /i "quarto-0.2.144-win.msi" # will suppress the interaction but some windows open msiexec /i "quarto-0.2.144-win.msi" /passive # using quiet mode does not seem to work msiexec /i "quarto-0.2.144-win.msi" /quiet # using special /qn flag either msiexec /i "quarto-0.2.144-win.msi" /qn ```` I don't really know what is supposed to work to install silently on Windows. Are all the usual flag from MSI activated in this bundle ? I found this while working on https://github.com/quarto-dev/quarto-actions/tree/main/install-quarto, as I tried to install Quarto MSI release on Windows GHA runner and I had trouble with it. Simple test tried on GHA: * Workflow file: https://github.com/cderv/test-actions/actions/runs/1217636260/workflow * `quarto` not found but not issue at installation - it seems it is doing nothing: https://github.com/cderv/test-actions/runs/3557169919?check_suite_focus=true It could be interesting to check that the MSI file can be installed remotely using script. Downloading the release file and installing it on GHA windows runner seems the usual way forward. How can we install MSI bundle on remote windows host like GHA runner ? <details> <summary>About the new `quarto-dev/quarto-actions/install-quarto` </summary> For the sake of a working action on all OS, I ended using Scoop to install the binary from [r-bucket](https://github.com/cderv/r-bucket). Scoop does not seem to have issue with the MSI, but maybe because it 'unbundles' it and not really install it. So for now, new action `quarto-dev/quarto-actions/install-quarto` is working, but with delay for the availability of the release in the scoop bucket compared to Github release here. I could probably improve this with an action in this repo that would trigger the update of the manifest in the bucket. Also, for GHA, Windows runner have Chocolatey package manager installed by default. Maybe offering a quarto bundle in Chocolatey would be interesting. It would help for Github action but with the same drawback as current scoop bucket. </details>

When I use `quarto render test.qmd -t pdf` on the test document in `tests/docs/` I get a `missing \begin{document}` error because the intermediate `.tex` file created looks like this: ``` %PDF-1.5 %ÐÔÅØ 5 0 obj << /Length 131 /Filter /FlateDecode >> stream xڅΫ1P߯ي^zo_[Kx$›àê‚@A±ü=…•Ԍ™“q¸Âa«ÜŸ\VµØÈÎ$êù’r¢äõŒ½Þ½ŒõNôên¬$}zÞÚdzy˜C»PÀL%F™Ê.†BûSëDŸ|‰A·ÏòçϺª7x=(ˆ endstream endobj 14 0 obj << /Length 225 /Filter /FlateDecode >> stream xÚő=OÃ0@÷üŠíÁîÏvìDAÌÞ*K1I—´Jóÿ…CZ¤”…¡ӓN÷ýz@xmðʧÔìöŽÁX݆– }Bp9‚GÒL
Rñ|%“˜ËŠ‹üH﻽quôÆ/edu´”#Яeo%wDZ—Ê -•C|lŒïù¿øóEõû?᠈«ÚD¿¥V%l½Y쨛6•wz~ËT¨ÙP¦Ê4´f§áx‘Š™Åyç<å~ÊçaÎCž·yS»2•înqC¡v¯S8íuÌöŽ—Ô|
üoM endstream endobj 28 0 obj << /Length1 1772 /Length2 20418 /Length3 0 /Length 21540 /Filter /FlateDecode >> stream ``` [test.zip](https://github.com/quarto-dev/quarto-cli/files/7134654/test.zip) I noticed this on another document of mine, so I went back to the simplest test possible. I thing it was one of the more recent commits (last 3 days).

The corresponding jupyter notebook option is called `--no-browser`, matching it might help some users not to get confused. `ERROR: Error: Unknown option "--no-browser". Did you mean option "--no-browse"?`

Hi, I have started to use Quarto with RStudio and really like its functionalities. I am currently creating a Quarto file to show the analysis of some data and I have noticed that if I use a Windows 10 Desktop, when rendering the "intermediate files" (PNG files in this case from the R figures), those files can be accessed as they are in a "-files" folder. This is useful because I can use the PNG files for a presentation or something else, and the quality is good. But when I render using a laptop with Windows 10, the folder that hosts those files is deleted upon completion of the rendering. And I am using the exact same Quarto file in each case, which is puzzling. Could you provide some support on why this is happening? In both computers I am using the RStudio Preview version (2021.09.0)

When including shiny components with `server: shiny`, `quarto run` (and maybe other deployment options, I didn't test them) doesn't serve the root directory. So when including images in the rest of the document: ```   ``` doesn't work, but moving the image to a `www` folder ```   ``` works. Images or plots that are included via a link or as output from code chunks (thus e.g. with `src=demo_files/unamed-chunk-1.png`) also still work. This might be something worth implementing to eliminate differences between regular qmd documents and those with shiny. PS: I realize I am playing around with lots of features and breaking things in the process, so do let me know if those issues are premature because you are working on it anyways!

I'm currently working on a re-working of The Carpentries lesson infrastructure to be [more portable and easier to work with](https://carpentries.org/blog/2020/08/lesson-template-design/) by separating the components for building, styling, and validation so that the tools could be updated independently of the lesson contents. One of the methods I've used to do this is [a {pkgdown}-style approach that allows for alternative HTML, CSS, and JS bundled in a separate package that can be independently updated](https://zkamvar.github.io/user2021/#p28), aka a remote theme. I would really like to use quarto under the hood for our infrastructure, but the theming seems to be limited to modifying CSS. Is it possible to include HTML templates from a remote source?

With `quarto render` working both on quarto projects as well as with a file argument on individual files I wonder if the same could be achieved to make `serve` and `preview` the same (or one) command as well i.e. so that both work on projects and single documents.

**tldr**: Regression from Rmd to qmd. `results='asis'` doesn't work in some cases. I was attempting something similar to what is described in https://quarto.org/docs/computations/running-code.html#jupyter: generating markdown from code cells. Say we have a folder with pictures and want to generate a revealjs slideshow with those. First, we get all the the image-paths and create markdown that has a heading (to make a new slide) and the image: ``` # Images ```{python} from pathlib import Path from IPython.display import display, Markdown paths = Path('.').rglob('./*.png') texts = [] for n,p in enumerate(paths): text = f'## Image {n}\n\n\n\n' texts.append(text) display(Markdown(text)) ``` However, if we look at the intermediate markdown file, we see that the output is enclosed in a div (plus one inner div per display call): ``` ::: {.cell} ::: {.cell-output-display} ## Image 0  ::: ... more images ::: ``` Unfortunately, due to the surrounding divs, revealjs no longer recognizes those headings as new slides, though they get formatted as headings and the images show up, just all on one slide. With Rmarkdown there was the chunk option: `results = 'asis'` which would include the chunk ouput as plain text in the markdown file, no questions asked. If I do the same in Rmarkdown and add this chunk: ``` ```{r test, results='asis'} for (x in py$texts) { cat(x) } ``` it works, but the same chunk rendered by quarto (with the knitr engine) still encloses the output in a `{.cell}` div. PS: love what you are doing with quarto!

Looking at the [quarto.org](https://quarto.org) site we see that the left sidebar has a scrollbar (both x and y), even though it should not need on:  I tried adding: ``` /*-- scss:rules --*/ .overflow-scroll { overflow: auto !important; } ``` to a custom scss theme, but this seems to get overridden by bootstrap once loaded: 

Hello, I noticed the following behaviour: starting with `document_02.ipynb` only :  `document_02.ipynb` is converted to `qmd`  then `document_02.qmd` is rendered  it looks like `document_02.ipynb` has now disapeared  Is that expected behaviour ? If yes I think that would be great to have a option to keep both files. Regards

Hello, I was trying `quarto` on the following example: ```` --- title: document_01 author: statquant toc: true toc-depth: 2 format: html: max-width: 800px fontsize: 18px html-math-method: katex jupyter: ir --- That's a test ```{r} #| echo: false plot(1:5) ``` ```` and running `quarto convert document_01.qmd` `document_01.ipynb` got generated but opened in `jupyterlab` as follows:  Is there additional yaml I should be adding ? PS: - Sorry for the missing ticks in `document_01.qmd` I could not have them displayed properly - The document mentions `quarto convert document.qmd --to xxx` while on the version I got it should say `quarto convert document.qmd --output xxx` (I think as the first errors while the second passes). Happy to PR this documentation change if that's any useful

Any plan to also add to quatro a distill theme that makes it possible to write distill type articles and blogs using formats besides rmarkdown.

Using Quarto v0.2.94 Given a `_quarto.yml` file that declares a site but without a `site` block: ```yaml project: type: site ``` Quarto will err when `quarto render`: ``` ERROR: TypeError: Cannot read property 'title' of undefined at title (file:///Users/aron/dev/rstudio/quarto-cli/src/project/types/website/website-navigation-md.ts:113:9) at Object.getUnrendered (file:///Users/aron/dev/rstudio/quarto-cli/src/project/types/website/website-navigation-md.ts:132:25) at file:///Users/aron/dev/rstudio/quarto-cli/src/project/types/website/website-navigation-md.ts:38:8 at Array.forEach (<anonymous>) at createMarkdownEnvelope (file:///Users/aron/dev/rstudio/quarto-cli/src/project/types/website/website-navigation-md.ts:36:12) at websiteNavigationExtras (file:///Users/aron/dev/rstudio/quarto-cli/src/project/types/website/website-navigation.ts:237:9) at async Object.formatExtras (file:///Users/aron/dev/rstudio/quarto-cli/src/project/types/website/website.ts:114:11) at async runPandoc (file:///Users/aron/dev/rstudio/quarto-cli/src/command/render/pandoc.ts:165:10) at async renderPandoc (file:///Users/aron/dev/rstudio/quarto-cli/src/command/render/render.ts:498:24) at async Object.onRender (file:///Users/aron/dev/rstudio/quarto-cli/src/command/render/render.ts:678:26) ``` Running `quarto serve`, the command displays the same error and loops until killed. The error goes away with an empty `site` block; the `title` is unnecessary: ```yaml project: type: site site: {} ```

Feature request: It would be amazing if quarto could include native collaborative writing tools, with the capabilities of {trackdown} https://github.com/claudiozandonella/trackdown

The documentation talks about "website", yet the CLI only supports "site". Should the CLI support "website" as equivalent to "site"? ``` quarto create-project --type website ERROR: Error: Project type must be one of default, site, book, but got "website". at Object.value (file:///Users/aron/dev/rstudio/quarto-cli/src/command/create-project/cmd.ts:35:17) at parseFlags (https://deno.land/x/[email protected]/flags/flags.ts:168:34) at Command.parseFlags (https://deno.land/x/[email protected]/command/command.ts:1113:22) at Command.parse (https://deno.land/x/[email protected]/command/command.ts:896:58) at Command.parse (https://deno.land/x/[email protected]/command/command.ts:870:33) at quarto (file:///Users/aron/dev/rstudio/quarto-cli/src/quarto.ts:59:64) at file:///Users/aron/dev/rstudio/quarto-cli/src/quarto.ts:76:11 ```

Hello, I just discovered Quarto by accident and think it can be a perfect tool for academic writing, reporting and publishing. Thanks for the great idea! The support for subfigures and cross-references seems better comparing to R Markdown. While playing with it, I notice several aspects could potentially be made more friendly to the users. 1. **Merge sub-captions to main caption** Currently the sub-captions are shown below each sub-figures as `(a) sub-caption1` and `(b) sub-caption2`. It will be nice if the sub-captions can be included into the main caption like: `Figure 1. (a) sub-caption1 and (2) sub-caption2`, but keep the sub-figure indicators `(a)` and `(b)` in the figure. It seems that this is actually the default setup of the [pandoc-crossref](https://lierdakil.github.io/pandoc-crossref/) but modified in Quarto. It will be nice if the user can switch to different modes easily.   2. **Automatically adjust sub-figures to fit to designated width (e.g., textwidth)** This is a dream feature to me. I used to use latex to include sub-figures in articles or beamer slides. It was a pain to adjust the sub-figures (with different original aspect ratios and sizes) to fit one page or slide. I had to manually adjust the percents for each sub-figure until the final result looks good. I could play with the numbers in the custom layout feature in Quarto, but it is still cumbersome. The following two figures compares results using the default `layout-ncol=2` option and the custom layout.   It will save a lot of time if the adjustments can be automatically done by Quarto. For example, by setting a total width and the height of the entire figure, no manual adjustments is required to fit the entire figure to the designated width (normally it is the textwidth, but can be other widths if figure wrapping is needed). 3. **Location of captions for sub-tables** In my field (engineering), normally the captions for tables are included above the table. This is the case for single tables from Quarto. But when I include sub-tables, the sub-captions and the main captions were treated like figures and were included below the tables. Is this designed by purpose or a glitch?  4. **Output to docx** Just spotted another issue when rendering the file to a docx file. The subfigures were considered as tables and the sub-figures were numbered as individual figures. 

I am curious if there is any way to enable using for-loops within tabsets in a user-friendly manner for Quarto. There is a convoluted workaround for doing it in RMarkdown using R code and exploiting calls to the console via `cat()`, but I don't know of a means of accomplishing the same result in Python. I also don't know much about how the rendering is done at the lower-levels, so I am not sure if this is even a reasonable ask. I find the tabsets to be very useful for presentations of results and use them quite frequently for demonstrating the effect of different variables or generating figures for a large number of samples. It would be awesome to have a more efficient way to do it, and also a means with Python! Here is a.. mostly.. minimal working example of the way that I know of to accomplish what I am referring to with R, ```{r, data} iris2 <- data.table::data.table(iris) ``` ``` # Single Tabset Loop {.tabset .tabset-pills} ```{r, single_loop, results='asis'} for(variable in colnames(iris2[,-'Species'])){ cat("##", variable, '<br>', '\n\n') g <- ggplot2::ggplot(data=iris2, ggplot2::aes_string(x=variable, fill='Species')) + ggplot2::geom_density(stat="density", alpha=0.2) print(g) cat('\n', '<br><br><br>', '\n\n') } ``` ``` # Nested Tabset Loop {.tabset .tabset-pills} ```{r, nesteed_loop, results='asis'} iris2 <- data.table::data.table(iris) for (i in seq_along(levels(iris2$Species))){ species <- levels(iris2$Species)[i] cat('## ', species, '{.tabset .tabset-pills}', '\n', '<br>', '\n\n') for(variable in colnames(iris2[,-'Species'])){ cat("###", variable, '<br>', '\n\n') g <- ggplot2::ggplot(data=iris2[Species %in% species], ggplot2::aes_string(x=variable)) + ggplot2::geom_density(stat="density", alpha=0.2, fill = c("#F8766D", "#00BA38", "#619CFF")[i]) print(g) cat('\n', '<br><br><br>', '\n\n') } cat('\n', '<br><br><br>', '\n\n') } ```

I would've expected empty Shiny server contexts to be allowed: ```{r} #| context: server ``` But if the server context is empty (or contains only whitespace), you get this error: ``` Error in shiny_prerendered_app(target_file, render_args = render_args) : No server contexts or server.R available for density-contours.qmd ```

After using the new docs layout a bit, the streamlined navbar is quite nice but it can be hard to know what is inside each item. I know you can currently do a manual dropdown item in the navbar, but would it be possible to have a smart navbar dropdown that shows you the exact same menu as shown on the section's left table of contents? <img width="1353" alt="Screen Shot 2021-08-06 at 9 10 28 AM" src="https://user-images.githubusercontent.com/12160301/128540724-e44cc78c-13bd-4f51-bcb0-176dc667925b.png"> When navigating from other pages, this preview would help you find your way without having to click on each section to see its TOC. I know I *could* do this manually, but it seems quite tedious 😉 Here are some similar ideas in practice: [GatsbyJS](https://www.gatsbyjs.com/) <img width="1209" alt="Screen Shot 2021-08-06 at 9 00 19 AM" src="https://user-images.githubusercontent.com/12160301/128540847-fe11c373-10a7-40aa-8e39-2c948568f699.png"> [rstudio.com](https://www.rstudio.com/) <img width="1353" alt="Screen Shot 2021-08-06 at 9 17 05 AM" src="https://user-images.githubusercontent.com/12160301/128540895-60529332-434f-44e5-8ade-ea75fea0e98f.png"> [algolia](https://www.algolia.com) <img width="1353" alt="Screen Shot 2021-08-06 at 9 18 27 AM" src="https://user-images.githubusercontent.com/12160301/128541094-0fb39844-2b0f-4ee8-9ee2-3c41774fde10.png">

I want to be able to configure the outputs of my R code in an Rmd file, according to the settings in the project `_quarto.yml` or `_variables.yml`. I could, I guess, just find these and parse them again in my R code, but is there some more idiomatic way of reading the settings in these files?

We just ran into a notebook that uses some libraries that generate some widget-based visualizations, they render OK in nbviewer but Quarto seems to have some issues with it. We wonder if the underlying issue here is related to javascript or styling files used by these libraries. i.e. Holoviews. The notebook in question: https://git.earthdata.nasa.gov/projects/LPDUR/repos/hls-tutorial/raw/HLS_Tutorial.ipynb With nbviewer: https://nbviewer.jupyter.org/urls/git.earthdata.nasa.gov/projects/LPDUR/repos/hls-tutorial/raw/HLS_Tutorial.ipynb/%3Fat%3Drefs%252Fheads%252Fmain And this is how part of the notebook renders with Quarto. 

After having now spent quite a lot of time on this issue, the answer seems to involve a lot of heavy lifting with CSS and Java Script. Wondering about the possibility of adding the Bootstrap external link icon (https://icons.getbootstrap.com/icons/box-arrow-up-right/) and `target="_blank" rel="noopener"` (<https://web.dev/external-anchors-use-rel-noopener/>) to links in the body (not navbar, toc, or sidebars) that begin with http(s). My current solution works, but I cannot get the icon color to match the link color: ``` /* Display an icon after external links */ main > p a[href^="http://"]:after, main > p a[href^="https://"]:after { content: url(./images/box-arrow-up-right.svg); vertical-align: 0%; display: inline; text-decoration: none; padding-left: .25rem; padding-right: .35rem; } ``` <img width="1370" alt="Screen Shot 2021-08-04 at 4 49 14 PM" src="https://user-images.githubusercontent.com/12160301/128269841-786ede17-70ca-4420-8ace-31fb1d93920b.png">

Please do tell me to stop if you've had enough of these, but, on current `main` branch ... ``` quarto create-project quarto-project cd quarto-project quarto render . --to gfm ``` gives: ``` ERROR: NotFound: No such file or directory (os error 2) at deno:core/01_core.js:106:46 at unwrapOpResult (deno:core/01_core.js:126:13) at Object.opSync (deno:core/01_core.js:140:12) at Object.realPathSync (deno:runtime/js/30_fs.js:137:17) at projectPath (file:///Volumes/zorg/mb312/dev_trees/quarto-cli/src/command/render/render.ts:569:16) ... ```

Shortcodes for values in `_variables.yml` are not rendering properly in navbars:  They work in the titles of individual documents, though: ``` --- title: "RStudio Workbench {{< var version >}}" date: "`r Sys.Date()`" --- ```  Relevant `_quarto.yml`: ``` site: title: "RStudio Workbench Administration {{< var version >}}" navbar: type: light # title: "RStudio Workbench Administration {{< var version >}}" ``` The shortcode isn't evaluated properly regardless of which `title` line is being used. Here's a repo with a minimal example: https://github.com/pommevilla/navbar-shortcodes

I noticed that Quarto seems to ignore my list of documents to render: ``` # _quarto.yml project: render: - index.qmd - about.qmd ``` I also have a document `another.qmd` in the same directory, and I get the following output from `quarto render .`: ``` [1/3] index.qmd [2/3] another.qmd [3/3] about.qmd ``` I was expecting, from the [docs](https://quarto.org/docs/projects/quarto-projects.html#render-targets), that Quarto would ignore `another.qmd`. Is that not so?

Please forgive my endless questions, but Is there a way of specifying the configuration that is usually in `_quarto.yml` - on the command line? As in: ``` quarto render . --config=alternative_quarto.yml ``` Our use-case is that we want to build the site as GFM, in order to resolve things like citations and execute any `r` macros, before post-processing into notebooks, but building in GFM is forbidden for the `book` type, which is the type specified in our standard `_quarto.yml` file.

This is just to tell you about a use-case that we have, that might be of interest. ## Summary We want to separate the idea of a 'page' and a 'notebook', so that a page can *contain* one or more notebooks. ## Detail We (@stefanv and I) are writing a new edition of a book on probability and statistics. For the book format, it makes most sense to us to have chapters, and pages, that are relatively long. These chapters will have several examples. Often we want to have separate notebooks for the students to interact with for each example. This in turn means we would like to change from the standard "Download / run this page as a notebook" idiom, that works really well in JupyterBook - to an idiom where, at any point in a page, there is a 'Download / run this part of the page as a notebook'. At the moment we are doing that with an awful hack, where we use macros like `r begin_notebook('first_example')` and `r end_notebook()` to drop comments into the built `.Rmd` page, and then post-process to extract the notebooks. But this is just to wonder out loud whether there is a better way to do this - and whether this is an interesting way of thinking about the relationship of pages to notebooks.

This is related to a Workbench request, something like this: https://docs.rstudio.com/ A few other footer examples: https://rstudio.github.io/distill/website.html#site-footer https://hugo-apero-docs.netlify.app/ https://www.tidyverse.org/packages/ https://www.tidymodels.org/start/

Can I ask for advice about how to replicate some formatting we developed in Bookdown? We were using the `kableExtra` package to tweak the formatting of tables, so it differed in HTML and PDF output. Specifically, we were using a custom table output function like this: ``` ketable <- function(df, caption) { rt <- kableExtra::kable(df, caption = caption, booktabs = T) if (knitr::is_latex_output()) { return(kableExtra::kable_styling(rt, latex_options = 'scale_down')) } # See https://cran.r-project.org/web/packages/kableExtra/vignettes/awesome_table_in_html.html # Update table-style.html to match kableExtra::kable_paper(rt, lightable_options = c("striped", "hover"), latex_options = c("striped", "scale_down"), full_width = F, position = "center" ) } ``` with some custom CSS to make our HTML tables look nice. Is there an equivalent way to tweak table formatting in Quarto? How can we make it differ according to output format (HTML or PDF)?

When porting a project from Bookdown, I found the following behavior surprising. We are embedding code chunks in list items, for example: ``` #. First item: ```{python} print("First item code") ``` ``` When using `rmarkdown::render()`, this executes and renders as I would expect with the rendered code output under the indented code block. When rendering the same Rmd file in Quarto, the code executes correctly, but the rendered doc has lots of stray Markdown in it, such as: ~~~ ::: {.cell indent=” “} ```{.python .cell-code} print("First item code") ``` ::: {.cell-output-stdout} ``` First item code ``` ::: ::: ~~~ With Qmd format, the rendering does not interpret the in-list code blocks correctly, and leaves output like the following: ~~~ First item: print(“First item code”) ``` ~~~ The repo <https://github.com/matthew-brett/quarto-test/tree/main> has a MWE of Markdown, Rmd via Quarto, Rmd via `rmarkdown`, and Qmd format, generated with `make all`. I am using the latest development version of Quarto, as of this evening.

I am test-driving quarto by converting my old diploma thesis originally done in LaTeX to it. I have found that math content in headings is sometimes not correctly displayed, e.g. https://rstub.github.io/diploma-thesis/period_efa.html: * In the main text the math content is correctly shown in both the chapter and the section heading * The math content in the section heading is also correctly shown in the on-page TOC * The math content from the chapter heading and the title is shown in raw form in the main TOC and on the chapter navigation lins in the previous/next chapter Anything I can do to fix this? In addition, I have the impression that the math content in the chapter heading is breaking recognition of the chapter label. At least I don't see what is wrong with https://github.com/rstub/diploma-thesis/blob/main/period_efa.qmd#L7 and the label works once I remove the math content. However, references to section 2.1 with the same math content do work w/o problems. Any idea what is going on here? BTW, is there a way to retain the intermediate markdown files (after knitr/jupyter but before pandoc)? That would help me debugging stuff like this.

In looking through the existing docs (cc @pommevilla @rich-iannone), I realized I never see an example of what goes inside a `.qmd`, `.Rmd`, etc. file within a chapter. I know the option to have a shared YAML, but I assumed that meant an individual item's YAML would override that. In an existing Quarto book project, we have several chapters, with this YAML at the top as an example: ``` --- title: Importing data for the first time --- ## Introduction some text here ``` When I render this, the level 2 header (`## Introduction`) is the chapter title. I would expect the `title` in the YAML to show up though instead. To get this outcome, we had to do: ``` --- title: Importing data for the first time --- # Importing data for the first time ## Introduction some text here ``` But even more surprising, if I have a bunch of text at the top then a header down below: ``` --- title: Importing data for the first time --- lines and lines of text here lines and lines of text here lines and lines of text here lines and lines of text here lines and lines of text here ## Introduction more text here ``` The same thing happens- the first `## Introduction` header becomes the chapter title, and the placeholder between `lines and lines of text here` and `more text here` is gone. This appears to happen with the first header I provide, regardless of the level header (I tested up to level 5), and regardless of placement (as in, it could be paragraphs into the chapter and still get transported up to the chapter title). 1. Are project files intended to have their own YAML, and if so, which options are valid there? 2. If I'm not supposed to put the title in the YAML, will the chapter title always be the first header (regardless of level and placement)?