On 1/17/2014 1:15 PM, Marco Patzer wrote:
On 2014–01–15 Nicola wrote:
maybe this is a known issue,
It's a known fact that context has a different notion of valid file names than your operating system does. This is by design. To quote the manual:
“It is highly recommended, that all input files, i.e. the ConTEXt source and other included files such as image files, have only the letters a–z, digits and dashes in their names, that is in the names of their full paths, otherwise you can easily get into problems.”
I'm writing a ConTeXt document called modern-c++.tex (in OS X 10.7.5). The content of the file is:
\setupbibtex[database={modern-c++}, sort=author] \setuppublications[numbering=yes] \starttext \completepublications[criterium=all] \stoptext
Especially since c++ didn't work out, I expected dropping the “++” would work (“modern-c”), but it didn't. So I ran some tracing:
\enabletrackers [resolvers.readfile] \starttext \readfile{file++.ext}{}{} \stoptext
This reports:
files > readfile > not found by tree lookup: file .ext
Which means the “++” is replaced by two spaces, instead of searching for “file++.ext” or “file.ext” which is what I had expected. I didn't dig into the code to check where the spaces creep in.
Regardless if this particular issue gets fixed or not, I doubt that Hans will put much effort into general support for “esoteric” file names. So, it's best to avoid plus signs in file names.
indeed. names are parsed as url's (so + become space) so a possible fix is: function getreadfilename(scheme,path,name) local fullname if hasscheme(name) or is_qualified_path(name) then fullname = name else name = url.escape(name) -- yes or no ? fullname = ((path == "") and format("%s:///%s",scheme,name)) or format("%s:///%s/%s",scheme,path,name) end return resolvers.findtexfile(fullname) or "" -- can be more direct end but one cannot predict how this passes further on through the system also, because one can say: \readfile{file\letterpercent2B\letterpercent2B.ext}{}{} the hack in fact should be: if not string.find(name,"%%") then name = url.escape(name) -- if no % in names end which then handles both \readfile{file\letterpercent2B\letterpercent2B.ext}{}{} \readfile{file++.ext}{}{} ok. Of course, when moving from c++ to c# one gets things like \readfile{file\letterhash.ext}{}{} where the # will sometimes confuses macros later on. Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------