The source map comment causes a problem on some versions of internet explorer because of how it interacts with some of its backwards compatibility features.
This problem is usually innocuous (just a spurious error) but I have never been completely convinced that it is always innocuous on older versions of internet explorer. But since I think you (Brian) are a mac user, I think you should be able to ignore this issue? Remember that in javascript // introduces a comment (rather like NB. in J). Thanks, -- Raul On Mon, Feb 3, 2014 at 10:13 AM, Brian Schott <schott.br...@gmail.com> wrote: > Bill's last reply in this thread seems to characterize well the situation I > am dealing with in my attempt to employ html5 canvas in a jhs app. The good > news is that I understand better the use of `~root/` as a trick to load .js > files (libraries, mostly) into my app, but I seem to be up against a > serious problem. One of the libraries I want to load is named swizzle.js > and notice the result of a grep shows that swizzle.js *itself* calls a > version of swizzle, apparently. > > server:j64-701-user brian$ grep -n -d recurse "dist\/sizzle\.min\.map" * > tgsjhs/scripts/sizzle.js:3://# sourceMappingURL=dist/sizzle.min.map > > Now this is a real puzzle to me because it looks to me like that line is > just a comment in the file sizzle.js, and should not be executed, but it > is. Am I incorrect about that being just a comment, or if it is not just a > comment, then is there any practical way to workaround this problem? > > By the way, the error I am getting is because my scheme of revising the > verb jev_get_jfilesrc_ as follows. [To see the original script, open the > file j64-701/addons/ide/jhs/jfilesrc.ijs . Scanning below for the NB. > comments containing `Brian` should point you to my kludges.] > > NB. serves .htm .js etc pages from anywhere > NB. possible security issues! careful on allowing other suffixes! > NB. ~root kludge to read gnuplot js files from > /usr/share/gnuplot/gnuplot/4.4/js/ > jev_get=: 3 : 0 > smoutput 'y';y > if. y-:'favicon.ico' do. favicon 0 return. end. > NB. next line revised below it by Brian with an if. clause > NB. y=. jpath(5*'~root/'-:6{.y)}.y > if. '~root/'-:6{.y > do. y=.5}.y > else. y=. PATH,y > smoutput 'no ~root given ', y > end. > d=. fread y > NB. Firefox 8 requires a response header > if. ('.htm'-:_4{.y)+.'.html'-:_5{.y do. htmlresponse d,~fsrchead rplc > '<TYPE>';'text/html' return. end. > if. (#mimetypes) > i=. ({:"1 mimetypes) i. <@tolower@}.(}.~ i:&'.') y do. > t=. i{:: {."1 mimetypes > else. smoutput 'will not get file ',y return. end. > NB. Brian also created the following if statement for debugging > NB. prior to that only the following line existed: t gsrcf d > smoutput 't';t > smoutput 'd';d > if. d~:_1 do. > t gsrcf d > else. 'failed' > end. > ) > > > > On Sun, Jan 12, 2014 at 10:59 AM, bill lam <bbill....@gmail.com> wrote: > >> Oh, yes, jhs is on the server side, your browser is on the >> client side, if the original request is in the form >> http://localhost/ then you probably don't need the base tag, the >> uri src of the javascript can either be a http://.... or just >> bare path /.... . In the fist case ,the browser will ask the other >> website to get the js, while in the second, the browser will >> issue a http get cmd to jhs, and it is up to jhs how to handle >> the get request. eg jhs may interpret the ~root as filesystem's >> root. >> >> > Thanks, > > -- > (B=) > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm