I am trying to implement a syntax highlighter for markdown for my project in PySide. The current code covers the basic, with bold, italic, code blocks, and some custom tags. Below is an extract of the relevant part of the current code.
What is blocking me right now is how to implement the highlighting for titles (underlined with ===, for the main title, or --- for sub-titles). The method that is used by Qt/PySide to highlight the text is highlightBlock, which processes only one line at a time.
class MySyntaxHighlighter(QtGui.QSyntaxHighlighter):
def highlightBlock(self, text):
# do something with this line of text
self.setCurrentBlockState(0)
startIndex = 0
if self.previousBlockState() != 1:
startIndex = self.blockStartExpression.indexIn(text)
while startIndex >= 0:
endIndex = self.blockEndExpression.indexIn(
text, startIndex)
...
There is a way to recover the previousBlockState, which is useful when a block has a defined start (for instance, the ~~~ syntax at the beginning of a code-block). Unfortunately, there is nothing that defines the start of a title, except for the underlining with === or --- that take place on the next line. All the examples I found only handle cases where there is a defined start of the expression, and so that the previousBlockState gives you an information (as in the example above).
The question is then: is there a way to recover the text of the next line, inside the highlightBlock? To perform a look-ahead, in some sense.
I though about recovering the document currently being worked on, and find the current block in the document, then find the next line and make the regular expression check on this. This would however break if there is a line in the document that has the exact same wording as the title. Plus, it would become quite slow to systematically do this for all lines in the document. Thanks in advance for any suggestion.
If self.currentBlock() gives you the block being highlighted, then:
self.currentBlock().next().text()
should give you the text of the following block.
Related
I need to indent some math stuff in the \details section of my .Rd documentation to enhance its readability. I am using mathjaxr. Is there any way to indent without installing roxygen2 or similar?
The math stuff is inline, so simply setting to display using \mjdeqn won't solve this.
I seem to have a reasonable "cheating" work around for indenting the first line using mathjaxr, at least for the PDF and HTML output.
We need to do two things:
Use the mathjax/LaTeX phantom command. phantom works by making a box of the size necessary to type-set whatever its argument is, but without actually type-setting anything in the box. For my purposes, if I want to indent, say, about 2 characters wide, I would start the line with a \mjeqn{\phantom{22}}{ } and following with my actual text, possibly including actual mathy bits. If I want an indent of, say, roughly 4 characters wide, I might use \mjeqn{\phantom{2222}}{ }.
Because mathjaxr has a problem with tacking on unsolicited new lines when starting a line with mjeqn, we need to prefix the use of phantom in 1 above with an empty bit of something non-mathjaxr-ish like \emph{}.
Putting it all together, I can indent by about 2 characters using something like this:
\emph{}\mjeqn{\phantom{22}}Here beginneth mine indented line…
I need to explore whether the { } business actually indents for ASCII output, or whether I might accomplish that using or some such.
We are generating a pdf through exams2nops using the items in blocks of choice, we would like to delimitate the blocks in the PDF adding a horizontal line after the last exercise of each block. Having that in mind we added a ***, ---, <hr/> however the behavior was always the same:
I would like a single line without adding the exercise number that's next in the exam:
It is not so easy to solve this by putting the horizontal line into the exercise file. The reason is that the line is needed after the answerlist but the answerlist is not formatted in the exercise but by exams2nops.
A workaround is to tweak the definition of the {question} environment in the LaTeX template used by exams2nops. By default this is simply:
\newenvironment{question}{\item}{}
Where \item is executed at the beginning of the {question} and nothing at the end of it. Changing this by
\renewenvironment{question}{\item}{\hrulefill}
would insert a horizontal line after every question. If you just want it after selected questions you need to insert if/else statements for certain enumerated items. For example, for inserting the horizontal rule after the second item only, you can redefine:
\renewenvironment{question}{\item}{\ifnum\value{enumi}=2 {\hrulefill} \else {} \fi}
Thus, you get the enumi counter from the {enumerate} environment that you use and compare it with 2. If true, you insert the horizontal line, and otherwise you do nothing.
Adding escapes for the backslashes you can pass this re-definition to exams2nops through the header argument:
exams2nops(c("swisscapital", "switzerland", "tstat2", "deriv2"),
header = "\\renewenvironment{question}{\\item}{\\ifnum\\value{enumi}=2 {\\hrulefill} \\else {} \\fi}")
The resulting output is:
In R Markdown, to make a text bold, we just need to do:
**code**
The the word code shows in bold.
I was wondering if there is a way to create a new command, let's say:
***code***
That would make the text highlighted?
Thanks!
It is not easily possible to create new markup, but one can change the way existing markup commands are rendered. Text enclosed by three stars is interpreted as emphasized strong emphasis. So one has to change that interpretation and change it to something else. One way to do so is via pandoc Lua filters. We just have to match on pandoc's internal representation of emphasized strong text and convert it to whatever we want:
function Strong (strong)
-- if this contains only one element, and if that element
-- is emphasized text, convert it to highlighted text.
local element = #strong.content == 1 and strong.content[1]
if element and element.t == 'Emph' then
table.insert(element.content, 1, pandoc.RawInline('html', '<mark>'))
table.insert(element.content, pandoc.RawInline('html', '</mark>'))
return element.content
end
end
The above works for HTML output. One would have to define what "highlighted text" means for each targeted format.
See this and this question for other approaches to the problem, and for details of how to use the filter with R Markdown.
I have inherited a bunch of Ruby code with rdoc comments, but many of the options and attributes are multi-line, such as:
# +param+:: Here is a parameter with a really long description
# that won't fit in one line
The end result is really hard to read docs using rdoc, sdoc, or yard. Each displays the second line in a different way. Am I missing something? Is there some way to retain line wraps, but make the generated documentation come out correct?
In YARD it's possible to continue the line by simply adding another comment line without any special annotation, just indent of at least two characters.
You can find it documented here: https://www.rubydoc.info/gems/yard/file/docs/Tags.md
For example this code below:
# class description
class Test
# #param [String] one this parameter has an astonishingly
# long description
def method(one)
nil
end
end
renders as
I'm currently trying to build a chat app, using the official markdown package as well as underscore's escape function, and my template contains something like this:
<span class="message-content">
{{#markdown}}{{text}}{{/markdown}}
</span>
When I grab the text from the chat input box, I try to escape any HTML and then add in line breaks. safeText is then inserted into the database and displayed in the above template.
rawText = $("#chat-input-textbox").val();
safeText = _.escape(rawText).replace(/(?:\r\n|\r|\n)/g, '\n');
The normal stuff like headings, italics, and bold looks okay. However, there are two major problems:
Code escape issue - With the following input:
<script>alert("test")</script>
```
alert('hello');
```
This is _italics_!
Everything looks fine, except the alert('hello'); has become alert('hello'); instead. The <pre> blocks aren't rendering the escaped characters, which makes sense. But the problem is, the underscore JS escape function escapes everything.
SOLVED: Line break Issue - With the following input:
first
second
third
I get first second third being displayed with no line breaks. I understand this could be a markdown thing. Since I believe you need an empty line between paragraphs to get linebreaks in markdown. But having the above behaviour would be the most ideal, anyone know how to do this?
UPDATE Line break issue has been solved by adding an extra \n to my regex. So now I'm making sure that any line break will be represented with at least two \n characters (i.e. \n\n).
You should check the showdown docs and the wiki article they have on the topic.
The marked npm package, which is used by Telescope removes disallowed-tags. These include <script> of course. As the article I linked to above explains, there's still another problem with this:
<a href='javascript:alert("kidding! Im more the world domination kinda guy. Muhahahah")'>
click me for world peace!
</a>
Which isn't prevented by marked. I'd follow the advice of the author and use a HTML sanitation library. Like OWASP's ESAPI or Caja's html-sanitizer. Both of these project's seem outdated dough. I also found a showdown extension for it called showdown-xss-filter. So my advice is to write your own helper, and use showdown-xss-filter.