David Bovill
Thu, 14 Sep 2000 15:24:46 -0700
Could you let us know how you configure the G4 TCP/IP so the browser and servers connect? I have a Java based server which listens and replies on Port 2001. I want to get Metacard and/or a browser to talk to this server. When the server loads it indicates that there is an open transport error, which I presume is due to a lack of an active TCP/IP connection. Perhaps if I knew exactly how to do this with MCHTTP and a browser offline, I could work from there... > From: Gary Rathbone <[EMAIL PROTECTED]> > Reply-To: [EMAIL PROTECTED] > Date: Thu, 14 Sep 2000 21:24:04 +0100 > To: [EMAIL PROTECTED] > Subject: Re: Trapping Errors & ..Mac to do server stuff... > > Thanks for this David, I'll give it a go. The reason I don't debug first and > then run through a browser may also be relevant to your other post > "Subject: anyone know how to get a Mac to do server stuff offline?" > > I'm running two browsers, an HTTP server (WebTen on port 80) and a Metacard > CGI server (on port 81) all on the same standalone Mac (G4 192Mb) . > > Pages are served from WebTen with cgi requests going to port 81 and Metacard > creating custom HTML responses on the fly. The build involves the > simultaneous creation of HTML forms (with Javascript processing) and the > corresponding MC cgi processing and responses. All quite chaotic but good > fun, and also a VERY quick build ! > > As I don't have DNS on my machine all calls are put through my (self > allocated) IP number followed by a port ref eg http://192.123.92.12:8181 > > For the actual project the processes will be split across numerous Mac's > with the only alteration being a change in the IP numbers... > > I know you said 'offlist' but the Mac and Metacard are a killer combo and > the world needs to know !!! > > > on 14/9/00 7:13 pm, David Bovill at [EMAIL PROTECTED] wrote: > >> Not quite sure why you don't debug first without running through a browser, >> but.... >> > --snip-- >>> Its all Mac based and I'm using MC as an HTTP server with additional code to >>> process forms. It works a treat... until I make a mistake like a typo... >>> >>> If I type : >>> >>> on mouseUp >>> put "thtml" ito "pageout" >>> end mouseUp >>> >>> into the button script of any stack, its accepted ie no semantic error. When >>> I click on the button however an error "Cant find handler" along with line >>> number and other error info is presented in the Execution Error dialog box. >>> Great - enough info to locate and fix the problem. >>> >>> Now if the above routine is called by a browser through a socket connection >>> to an MC stack then MC doesn't present the same Execution Error box and the >>> only clue that something is wrong is that nothing is returned to the >>> browser. Hence manually searching through the code. Any ideas ? >>> >> >> It should be possible to build in a debugger that works with the server, but >> figuring out how to do this/how the existing debugger works is not easy, and >> according to Scott is likely to be a moving target... >> >> An idea (if you have to debug online), is to put the cgi script into a >> variable and then have a go at something like >> >> repeat with lineNum = 1 to the number of lines of cgiScript >> put line lineNum of cgiScript into scriptLine >> try >> do scriptLine >> catch errorNum >> put "Error" && errorNum && "on line" && lineNum >> end try >> end repeat >> >> It'll be slow, but might work? >> > --snip-- > > > > Archives: http://www.mail-archive.com/metacard%40lists.best.com/ > Info: http://www.xworlds.com/metacard/mailinglist.htm > Please send bug reports to <[EMAIL PROTECTED]>, not this list. Archives: http://www.mail-archive.com/metacard%40lists.best.com/ Info: http://www.xworlds.com/metacard/mailinglist.htm Please send bug reports to <[EMAIL PROTECTED]>, not this list.