OK, thanks both of you, its looks like I need to sanitize all mentioned characters, since the reference strings will generally originate from formats other than ConTeXt, and we don't want ConTeXt to do any processing on them, aside from comparisons to resolve references.

As for Aditya's examples, the first results in a compilation error on my test file, while the second compiles without error, and gives the expected result.

On Thu, Sep 18, 2014 at 4:26 AM, Aditya Mahajan <adityam@umich.edu> wrote:
On Thu, 18 Sep 2014, Hans Hagen wrote:

On 9/18/2014 12:06 AM, Mark Szepieniec wrote:
Bump...

If it's not too much trouble, I would greatly appreciate some feedback
on this before I propose it to be merged into pandoc; even a "looks good
to me" from one of the ConTeXt gurus would be very helpful.

Thanks in advance,

Mark

On Tue, Sep 9, 2014 at 12:20 AM, Mark Szepieniec <mszepien@gmail.com
<mailto:mszepien@gmail.com>> wrote:

    I'm trying to fix a problem in pandoc (see
    https://github.com/jgm/pandoc/pull/1589) where it doesn't properly
    sanitize the reference labels in ConTeXt output, causing errors
    during compilation when a label contains '#' for example. Note that
    this sanitizing is needed in addition to the regular backslash
    escaping used for control characters: '\#' is still illegal in a
    label for example.

 (LaTeX label) = (ConTeXt reference). What Mark mean was references such as

\section[...]{...} or \startplacefigure[reference={...}].

    In the sanitizer function I'm writing, I'd like to properly escape
    all illegal characters, but I couldn't find an explicit list of
    allowed or illegal characters. Based on some testing I've conducted
    (see attached file), I've arrived at the following set:

    \#[]",{}%()|=

it depends on where these characters end up in

#  : always tricky as it denotes an argument, so escape
[] : depends if it gets fed into a macro that uses [] as delimiters
{} : only an issue when not balanced
%  : escaping needed as it's comment otherwise
() : depends on where it ends up, like []
|  : is special in context so needs escaping
\  : of course that one needs escaping

    1) Does this look like a reasonable set? Are there other characters
    or sequences that should be included, or are worth testing?

keep in mind that escapes should end up unescaped at some point

    2) I was told (see
    https://groups.google.com/forum/#!topic/pandoc-discuss/tYpXMUkmbEY)
    that if the characters " and , didn't work, it would count as a
    ConTeXt bug, is there any truth to that? Please let me know if any
    further info is needed on my part.

well, define bug ... one can say the same of < and > in xml -)

 Since I made that comment on the pandoc mailing list, let me explain.

Consider:

\section["some" reference]{Title}

Given how " behaves elsewhere in ConTeXt, a user would expect the above to be a valid input. If it is not, then it is bug (or atleast, surprising).

The same goes for

\section[some, reference]{Title}

if the result ends up in a comma separated list then , can be an issue but one can always wrap an argument in {} to hide that

    3) Does anyone see issues with this general approach? I'm relatively
    new to ConTeXt, so I might be missing either a huge problem, or an
    obviously easier way to do this.

i don't know ... i never used pandoc input

 Aditya

___________________________________________________________________________________
If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context
webpage  : http://www.pragma-ade.nl / http://tex.aanhet.net
archive  : http://foundry.supelec.fr/projects/contextrev/
wiki     : http://contextgarden.net
___________________________________________________________________________________