[EMAIL PROTECTED] wrote:
martin 2003/12/15 06:24:31

Modified: modules/experimental mod_charset_lite.c
Log:
Flame bait: mod_charset_lite would decide in the mime checker phase
whether a conversion was required, and would stick with that decision,
even if a later handler (like CGI script) would modify the actual
MIME type returned in the response. That was fatal when an EBCDIC
machine was going to return, say, an image/gif (because the initial
MIME type would be text/plain, and only later on be set to image/gif).
But even for "ordinary" ASCII-based machines the conversion fails
if a CGI script decides to return UTF-16 data.
At the moment the following change is #if #system(bs2000), but IMHO
it should be enabled globally: I moved the conversion decision into
the filter itself (because it is the handler which does the final
decision which MIME time is tagged on the response, and we can only
evaluate its decision after it started the output generation).
Revision Changes Path
1.67 +70 -0 httpd-2.0/modules/experimental/mod_charset_lite.c
Index: mod_charset_lite.c
===================================================================
RCS file: /home/cvs/httpd-2.0/modules/experimental/mod_charset_lite.c,v
retrieving revision 1.66
retrieving revision 1.67
diff -u -u -r1.66 -r1.67
--- mod_charset_lite.c 7 Nov 2003 01:01:27 -0000 1.66
+++ mod_charset_lite.c 15 Dec 2003 14:24:31 -0000 1.67
@@ -307,7 +307,16 @@
"mime type is %s; no translation selected",
mime_type);
}
+#if #system(bs2000)

This syntax causes a compile failure on Windows.
Compiling...
mod_charset_lite.c
modules\experimental\mod_charset_lite.c(310) : error C2014: preprocessor command must start as first nonwhite space
modules\experimental\mod_charset_lite.c(310) : fatal error C1017: invalid integer constant expression
Error executing cl.exe.


Bill

Reply via email to