Ad13 vs. 3.5.1
This is an unfinished comparison between AOLserver 3.5.1 and AOLserver 3.3.1 ad13 oacs1. [1] It is based on Rob Mayoff's AD13 distribution notes. SourceForge patch #570072 brought the AD13 internationalization patches to 3.4.2, and would probably be a good starting point for integration. - Jamie Rasmussen ---------------------------------------- oacs1 patches Vinod's ns_uuencode patch (SourceForge #474259) * The oacs1 patch has a major bug as discussed on SourceForge. I believe my corrected patch works but it hasn't been tested. This is not crucial because OpenACS tests for the desired behavior and falls back on (slow) TCL-based encoding. uid/gid from Jon Griffin * This is SF Patch #509413. Jon confirms it has been fixed in 3.5 . BSD exec patch * Is this SF Patch #415475? It's marked fixed. Merge some of Vinod's and Mat's changes for nsxml's makefile * nsxml has changed since the oacs1 version so I'm not sure if these changes are needed. ---------------------------------------- ad13 patches Numerous patches to support character encodings including Henry Minsky's patches to translate form data to multiple charsets within a single HTTP request. Described in encoding-doc.html. These should be fleshed out here. There is an additional encoding file, 8bit.enc (ascii, single-byte, 8 bits) * Should this be added to TCL? Set system encoding to ns/server/$server/SystemEncoding (default utf-8) in modules/tcl/init.tcl. * This is a 1 line change form.tcl and http.tcl are different, charsets.tcl needs to be added Added byte range support in fastpath.c and ns_returnfile. * This is a HTTP/1.1 feature. Patrick Spence asks if this could be made configurable to prevent abuse by download accelerators. Reset ADP exception before executing an ADP, in case previous ADP called ns_adp_break, etc. * One-line changes in adp.c and tclstubs.cpp to reset the exception * There are also stack checks throwing ADP_OVERFLOW. The stack check function was in thread.c of the old threads lib, and is not in AOLserver 3.5.1 ADP_OVERFLOW is still caught in 3.5.1, but it doesn't appear to be thrown. Ns_QueryToSet will not allow parameters whose names end with
