Re: [xslt] [OT] mime-types... (was Mac OS X Binaries Available)

[S Woodside <sbwoodside yahoo com>]

Jesse, you're right on, this is an IE problem. I'm involved in Mozilla 
development and there have been many flame wars about this. The 
one-minute summary is that the specs are very clear that when a 
mime-type is provided, the browser must follow it. In this case, the 
mime-type is unknown and so Apache is configured to send text/plain for 
.gz file. That forces mozilla to display it as text, without doing any 
sniffing or whatever.

There are four possible fixes:
- yell at mozilla to ignore the standards (they won't listen and I will 
agree with them)
- yell at apache for defaulting to text/plain in the standard config 
file (I've already done this but more can't hurt)
- change the default to application/octet-stream ( a much better 
default imho )
- get a newer mime-types table for apache.

simon

On Wednesday, May 21, 2003, at 08:32  AM, Jesse Pelton wrote:

>> -----Original Message-----
>> From: Steve Ball [mailto:Steve.Ball@zveno.com]
>> Sent: Wednesday, May 21, 2003 2:03 AM
>> To: xslt@gnome.org
>> Subject: Re: [xslt] Mac OS X Binaries Available
>>
>>> One note: the HTTP download of the .dmg, is not sent with
>> the proper
>>> mime-type so it fails in Mozilla. Either add the correct
>> mime-type to
>>> apache configuration, or potentially to the link, or make
>> it a .bin file
>>> (that's what apple often does).
>>
>> It works for me ;-)
>
> That doesn't mean it's right.  Simon says the mime-type header is 
> wrong; if
> he's right, why not fix it?  (I'm not saying he is right, mind you, but
> you're in a position to know.)  You've gone to some trouble to provide 
> a
> worthwhile service, and it seems a pity (and most un-Mac-like) to stop 
> short
> of making it as easy to use as possible, especially if the fix is easy.
>
> <rant>
> The idea, so prevalent on the 'net, that it's OK for things to be 
> wrong if
> they work with the author's tools is pernicious.  If things are done
> correctly, they'll work with any well-implemented tool, whether or not 
> the
> author has seen them.  If they're not done correctly, it's anybody's 
> guess
> what will happen.  Look at the mess the hacks have made of the Web, 
> largely
> thanks to the permissiveness of browsers like IE.
> </rant>