httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alexei Kosut <>
Subject Re: Apache's byteranges support (fwd)
Date Wed, 12 Feb 1997 03:56:55 GMT
Here's Lam's response to my response to his response (i.e., the third
round of email between us). Good to know that Adobe's also taking an
interest in this, given that they were mostly responsible for
byteranges, and their product uses them the heaviest.

BTW, I've also received a partial response from the Microsoft IE
people: the person I emailed has forwarded my mail to the appropriate
person WRT the Mac stuff, but also told me "you will have to sniff the
UA. It exists to handle situations such as this"

Sure... IMHO, using the User-Agent header for feature identification
is like determining whether or not a car has 4-wheel drive by looking at
the license plate frames. (feel free to quote me on that)


Anyone have any good suggestions?

Alexei Kosut <>      The Apache HTTP Server

---------- Forwarded message ----------
Date: Tue, 11 Feb 1997 19:28:04 -0800
From: Anthony Lam <alam@Adobe.COM>
To: Alexei Kosut <>
Subject: Re: Apache's byteranges support

It's nice to know you're willing to remove the "" from the header.  Thanks.
Yes, Mac IE is doing something wierd for byteserving.  We're currently
working closely with the Microsoft's Mac engineers and I'll ask them to
change that to follow the http 1.1 spec.
The behavior I described are for IE Win32.

Looking forward to hear from your decision about IE compatibility.  I
personally don't feel right about changing it either.  It's all up to your



At 06:01 PM 2/11/97 -0800, you wrote:
>On Tue, 11 Feb 1997, Anthony Lam wrote:
>> Thanks for your time looking into this.
>Thanks for getting in touch with us, and responding.
>> Let me stated what's the current browsers crop is doing and then you can
>> decide what you should change or not.
>> If you test out Netscape 3.0 successfully with 1.2b6, then seems like
>> Netscape can handle the "" in the content type header.
>> Unfortunately Internet Explorer 3.0/3.01 does not.  They're following the
>> latest HTTP 1.1 spec
>> Content-type: multipart/byteranges; boundary=THIS_STRING_SEPARATES
>>    Content-type: application/pdf
>>    Content-range: bytes 500-999/8000
>> and THIS_STRING_SEPARATES in the content-type header does not include "".
>> In fact,  all the other web servers we tested against, including IIS 2.0,
>> Netscape Enterprise and FastTrack, they all don't include "" in the header.
>> Since Internet Explorer is becoming very popular, I strongly suggest
>> removing the "" from your content-type header.
>Well, it should be pointed out that section 19.2 simply defines
>multipart/byteranges in relation to the MIME specification, and the
>example is just that - an example.
>However, if Internet Explorer really does not like the quotes, Apache
>can remove them. Our boundary strings are alphanumeric only, so they
>don't need to be quoted, although (as I said), the MIME spec
>reccomends it.
>> Netscape 3.0 is sending Request-Range and expecting multipart/x-byteranges.
>> They're using the old pairs.
>> Internet Explorer 3.0/3.01 and Acrobat is sending Range: and not
>> but they're expecting multipart/x-byteranges instead.  So somehow they're
>> out of sync.  This is a tough call.  I know you want to comply with the
>> latest spec.
>> But that's mean all the current IE 3.0 users cannot view PDF files on Apache
>> servers.
>Hmm. I downloaded IE 3.0 for the Mac (I don't personally have access
>ot a Windows machine, so cannot test that), and that was not my
>experience. Specifically, IE did not send Range:, or any other header,
>but instead tried to use the very old version of the spec; attaching
>the range request to the URL, with a semicolon seperator. No server
>that I'm aware of correctly processes this anymore. When it gets the
>404 back from the server, it retries without the range, but then
>somehow never passes the response onto the plugin. So that's very
>odd. I will likely contact Microsoft about this.
>However, assuming your description applies to the Windows version of
>IE 3.0 (which I personally have no way of testing), I'm not sure what
>can be done. As I said before, the Apache Group generally feels strongly
>about adhering to the HTTP specifications, and any change of this
>nature may not be in keeping with that policy.
>I'll have to consult our standards cop about this.
>> Please let me know what your decision are and is this something you can put
>> in your b7 release?
>1.2b7 will certainly be compatible with Navigator, as I described, and
>will likely have the quotes removed from the Content-Type string. As
>for IE compatibility, I can't yet say, but I'll keep in touch.
>Alexei Kosut <>      The Apache HTTP Server

View raw message