yihui

library(xml2)

page_post_links = function(
  page = 1, baseurl = 'https://www.rstudio.com/rviews',
  xpath = '//h2[@class="entry-title"]/a'
) {
  html = read_html(sprintf('%s/page/%d/', baseurl, page))
  xml_attr(xml_find_all(html, xpath), 'href')
}

yihui

#!/bin/sh

# you can replace $HOME with any dir
sed -i 's@\$TEXLIVEHOME@'"$HOME"'@' texlive.profile
wget http://mirror.ctan.org/systems/texlive/tlnet/install-tl-unx.tar.gz
tar zxf install-tl-unx.tar.gz
./install-tl*/install-tl -profile texlive.profile

# texlive.tar.gz is a portable and full TeXLive package
tar zcf texlive.tar.gz -C $HOME texlive

yihui

---

title: "Show Your Webcam in R Markdown Presentations" author: "Yihui Xie" date: "2015/10/25" output: slidy_presentation: incremental: yes duration: 40 footer: "Webcam in R Markdown, Yihui Xie, 2015" includes:

yihui

---
title: HTML Dependencies
output: html_document
---

This example explains how HTML dependencies work in **htmltools**. Sometimes an HTML fragment may have additional dependencies to work correctly, such as JavaScript and/or CSS files. In R Markdown documents, you have to let **rmarkdown** know these dependencies, so they can be added to the HTML header when the document is rendered through Pandoc.

Another thing that you need to pay attention is you have to "protect" your HTML output so that Pandoc does not touch it, e.g. when you have four leading spaces, Pandoc may think this line is supposed to be a `<pre>` block whereas you only meant to indent the line for cosmetic purposes. In this case, the function `htmltools::htmlPreserve()` will be _automatically_ applied to your HTML content in R Markdown if the content is generated from `htmltools::tags` or wrapped in `htmltools::HTML()`.

Now we use a random CSS file in the **knitr** package to illustrate how dependencies work. The goal here is to generate a 

yihui

Moved to https://github.com/yihui/shiny-apps/tree/master/voice

yihui

\documentclass{article}
\begin{document}

<<setup, include=FALSE, cache=FALSE>>=
library(knitr)
library(ggplot2)
@

\title{Knitr and ggplot2}
\author{Daniel Haugstvedt}

yihui

Here is a list of package vignettes _possibly_ built with **knitr** on CRAN.

```{r fetch-pkgs, include=FALSE}
pkgs = available.packages(contrib.url('https://cran.r-project.org', 'source'))
deps = tools::package_dependencies('knitr', pkgs, which = 'all', reverse = TRUE)[['knitr']]
deps = setdiff(deps, 'R.rsp')  # packages that do not use knitr
deps = sort(c(deps, 'knitr'))
```
```{r vig-list, include=FALSE, echo=FALSE, cache=TRUE, cache.extra=deps, message=FALSE, error=FALSE}
vigs = lapply(deps, function(pkg) {

yihui

options(warn = -1)
# cannot really suppress warnings from the root level
withCallingHandlers(warning("hi"), warning = function(w) {
    print(w)
})
# the warning can still be captured:

## <simpleWarning in withCallingHandlers(warning("hi"), warning = function(w) {    print(w)}): hi>

# however, this always works

yihui

stopifnot(getRversion() > '3.0.1')

par("page")
dev.off()

dev.new()

par("page")
plot(1)
par("page")

yihui

These 498 packages are currently affected by the check of ::: on CRAN:

AER
BLCOP
BSagri
BatchExperiments
BatchJobs
BayHap
BayesFactor