[ic] discussion of patch for IC compatibility with NS 7

Jeff Dafoe interchange-users@icdevgroup.org
Sat Sep 28 01:07:01 2002


> I don't understand this. What is getting tromped on?  And how would
> you prevent a *real* sequence that meant what it said (%253a) from
> being decoded improperly?

    It's the MM value on [more]d search pages.  Mozilla "erroneously"
urlencodes it in the HTTP request.  If there are sequences with the percent
sign in path_info, as you mentioned, then my "fix" would instead break
things.

> In any case, we wouldn't make a change in the core like that.
> Mozilla works great for IC, and (unfortunately) it is a small
> percentage of the market. When it grows in popularity, it will
> be the non-buggy versions that do so. I remember running into that
> bug, and that was *eons* ago at something like 0.8. Anyone who
> runs a browser like that in this day and age must be used to lots
> of weird results....

    I am experiencing the issue in the official release version of Netscape
7, which is Mozilla 1.0 .  If it was an obscure prelease version I wouldn't
be concerned at all.  It's not really even a browser bug if you look at RFC
1738, which says that "all unsafe characters must be encoded" but any octet
"may be encoded".


Jeff