[Mod_gzip] "mod_gzip_send_vary=Yes" disables caching on IE? (1.3.26.1a)

Fri, 6 Dec 2002 21:58:38 EST

--part1_1c1.2a9f476.2b22bdde_boundary
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit

In a message dated 12/6/2002 7:25:29 PM Central Standard Time, 
robertc@squid-cache.org writes:

> > What versions of SQUID are you talking about.
> > 
> > Be specific and please indicate release dates.
> 
> 2.5 pre versions from way back (sorry don't remember when the merge
> occured).
> 2.5 stable 1 (released sept 25th 2002).
> and all newer versions.
> 
> Rob
> 

Thanks Rob.

I really didn't know that SQUID 2.5 has finally been released
and there is now a version that won't just throw any response
with a "Vary:" header into the electronic wastebasket.

Contratulations. I mean it. I know that SQUID 2.5 is almost 2 
years overdue and it was hard work on everyone's part.

However... this still doesn't mean a whole lot in relation
to the current thread.

It will be YEARS before this version of SQUID which doesn't
just throw away "Vary:" responses will be in widespread
use and if the history of open-source sites upgrading their
software remains as it has for years now then there are
some that will probably NEVER upgrade.

I came across someone still using Apache 1.1 the other day
and it was no small site.

I don't think the answer to the issues raised by this thread
can ever possibly be "Just make sure you have the version
of SQUID that was just released 2 months ago".

There are some that won't even touch a new release of a
major open-source product until it's been GA for a least
a year, anway. They know that one of the largest black
holes that remains today in most all open-source 
projects is the TESTING. Often times the GA release
is still really a BETA because there just hasn't been
that much real TESTING. It's just the way it is with
open-source stuff. Apache is the worst. They rarely
test ANYTHING before they 'release' and they just
rely on the bug reports to see what they missed.

The 'workaround' for the problems mentioned is simply
going to be making people aware of what a mess this
all really is and giving THEM the ability to decide 
whether they want to use "Vary:" or not and weigh
the pros/cons.

I would say that it's shaping up like this...

1. If you cannot afford the chance that someone using some
ancient and/or broken browser is going to receive an encoding
that they can't decompress... and you can afford an enormous
amount of useless hits on you Server... then by all means
use "Vary:".

2. If you cannot afford the huge increase in traffic because
browsers and (most) Proxy caches are going to throw ALL "Vary:"
responses on the floor and you really don't care if someone
might be forced to join the modern world and upgrade their
browser.... then do NOT use "Vary:"

I always lean towards number 2.

I have no patience anymore for people that are using 
User-Agents that are more than 6 years old.

Later...
Kevin

--part1_1c1.2a9f476.2b22bdde_boundary
Content-Type: text/html; charset="US-ASCII"
Content-Transfer-Encoding: 7bit

<HTML><FONT FACE=arial,helvetica><FONT  SIZE=2><BR>
In a message dated 12/6/2002 7:25:29 PM Central Standard Time, robertc@squid-cache.org writes:<BR>
<BR>
<BR>
<BLOCKQUOTE TYPE=CITE style="BORDER-LEFT: #0000ff 2px solid; MARGIN-LEFT: 5px; MARGIN-RIGHT: 0px; PADDING-LEFT: 5px">&gt; What versions of SQUID are you talking about.<BR>
&gt; <BR>
&gt; Be specific and please indicate release dates.<BR>
<BR>
2.5 pre versions from way back (sorry don't remember when the merge<BR>
occured).<BR>
2.5 stable 1 (released sept 25th 2002).<BR>
and all newer versions.<BR>
<BR>
Rob<BR>
</BLOCKQUOTE><BR>
<BR>
Thanks Rob.<BR>
<BR>
I really didn't know that SQUID 2.5 has finally been released<BR>
and there is now a version that won't just throw any response<BR>
with a "Vary:" header into the electronic wastebasket.<BR>
<BR>
Contratulations. I mean it. I know that SQUID 2.5 is almost 2 <BR>
years overdue and it was hard work on everyone's part.<BR>
<BR>
However... this still doesn't mean a whole lot in relation<BR>
to the current thread.<BR>
<BR>
It will be YEARS before this version of SQUID which doesn't<BR>
just throw away "Vary:" responses will be in widespread<BR>
use and if the history of open-source sites upgrading their<BR>
software remains as it has for years now then there are<BR>
some that will probably NEVER upgrade.<BR>
<BR>
I came across someone still using Apache 1.1 the other day<BR>
and it was no small site.<BR>
<BR>
I don't think the answer to the issues raised by this thread<BR>
can ever possibly be "Just make sure you have the version<BR>
of SQUID that was just released 2 months ago".<BR>
<BR>
There are some that won't even touch a new release of a<BR>
major open-source product until it's been GA for a least<BR>
a year, anway. They know that one of the largest black<BR>
holes that remains today in most all open-source <BR>
projects is the TESTING. Often times the GA release<BR>
is still really a BETA because there just hasn't been<BR>
that much real TESTING. It's just the way it is with<BR>
open-source stuff. Apache is the worst. They rarely<BR>
test ANYTHING before they 'release' and they just<BR>
rely on the bug reports to see what they missed.<BR>
<BR>
The 'workaround' for the problems mentioned is simply<BR>
going to be making people aware of what a mess this<BR>
all really is and giving THEM the ability to decide <BR>
whether they want to use "Vary:" or not and weigh<BR>
the pros/cons.<BR>
<BR>
I would say that it's shaping up like this...<BR>
<BR>
1. If you cannot afford the chance that someone using some<BR>
ancient and/or broken browser is going to receive an encoding<BR>
that they can't decompress... and you can afford an enormous<BR>
amount of useless hits on you Server... then by all means<BR>
use "Vary:".<BR>
<BR>
2. If you cannot afford the huge increase in traffic because<BR>
browsers and (most) Proxy caches are going to throw ALL "Vary:"<BR>
responses on the floor and you really don't care if someone<BR>
might be forced to join the modern world and upgrade their<BR>
browser.... then do NOT use "Vary:"<BR>
<BR>
I always lean towards number 2.<BR>
<BR>
I have no patience anymore for people that are using <BR>
User-Agents that are more than 6 years old.<BR>
<BR>
Later...<BR>
Kevin<BR>
<BR>
<BR>
<BR>
<BR>
</FONT></HTML>
--part1_1c1.2a9f476.2b22bdde_boundary--