Actually, Apache has: AddDefaultCharset ISO-8859-1
And my content-type header meta tag in the HTML output also reads <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1" /> Header looks like this under regular CGI: HTTP/1.1.200.OK(CR)(LF) Date:.Tue,.13.Feb.2007.23:21:28.GMT(CR)(LF) Server:.Apache(CR)(LF) Set-Cookie:.publish=;.domain=.lfpress.com;.path=/cgi-bin;.expires=Wed,.14-Fe b-2007.23:21:29.GMT(CR)(LF) Connection:.close(CR)(LF) Transfer-Encoding:.chunked(CR)(LF) Content-Type:.text/html;.charset=ISO-8859-1(CR)(LF) (CR)(LF) And like this under mod_perl: HTTP/1.1.200.OK(CR)(LF) Date:.Tue,.13.Feb.2007.23:22:15.GMT(CR)(LF) Server:.Apache(CR)(LF) Set-Cookie:.publish=;.domain=.lfpress.com;.path=/cgi-bin;.expires=Wed,.14-Fe b-2007.23:22:15.GMT(CR)(LF) Connection:.close(CR)(LF) Transfer-Encoding:.chunked(CR)(LF) Content-Type:.text/html;.charset=ISO-8859-1(CR)(LF) (CR)(LF) So, it can't be header... this is getting truly bizarre... The system default charset for the linux box is ISO-8859-1. MySQL is using ISO-8859-1 as its default charset. Dunno what else to check. Here's another weird thing - the characters aren't showing up as encoded entities under either regular CGI or mod_perl - they are actual raw characters, not escaped or encoded. Under regular CGI, a caption line shows up as: Dennis.Garnhum,.former.mentor/instructor.at.the.National.Theatre.School,.say s.that.(93)people.who.have.a.burning.passion.that.can(92)t.be.stopped(94).ar e.most.likely.to.succeed.as.actors. Under mod_perl, it becomes: Dennis.Garnhum,.former.mentor/instructor.at.the.National.Theatre.School,.say s.that.(C2,93)people.who.have.a.burning.passion.that.can(C2,92)t.be.stopped( C2,94).are.most.likely.to.succeed.as.actors. BTW I am using an HTTP viewer for this: http://www.rexswain.com/httpview.html And the URLs I am using are: Regular: http://www.calgarysun.com/cgi-bin/publish.cgi?p=171767&x=articles&s=events Mod_perl: http://www.calgarysun.com/perl-bin/publish.cgi?p=171767&x=articles&s=events It looks like mod_perl is trying to insert double-byte characters where single-byte characters go. I checked the ASCII and ISO Latin-1 tables and those numbers are supposed to be empty charset entities - and yet they're not. When I check the ASCII table in my text editor (Ultra-edit), they show up as characters. Is this the Windows codepage (1251 I think) in action, extending my ASCII set? And yet, the characters show up under regular CGI in Fedora fine... for some reason mod_perl just seems to be adding a hex C2 before every non-ASCII character. Is it an escape sequence messing up or.... ? *scratches head 'till the blood comes* -----Original Message----- From: Jonathan Vanasco [mailto:[EMAIL PROTECTED] Sent: February-08-07 1:42 PM To: mod_perl List Subject: Re: Strange characters in output when filtered through mod_perl Just to clarify: On Feb 8, 2007, at 3:03 PM, Aaron Hawryluk wrote: > Our publishing system doesn't use any strange character sets - Your system is working with data in one character set, and publishing it to the web in another character set. The fix is *likely* just setting the right character set header in apache. Personally, I either do everything in UTF8 or ASCII with html entities for everything else. You could try doing: AddDefaultCharset utf-8 in httpd.conf or (i think this will work) $r->content_type("text/html; charset=utf-8"); in your handler // Jonathan Vanasco | - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - | SyndiClick.com | - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - | FindMeOn.com - The cure for Multiple Web Personality Disorder | Web Identity Management and 3D Social Networking | - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - | RoadSound.com - Tools For Bands, Stuff For Fans | Collaborative Online Management And Syndication Tools | - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
