I had this same problem (described in kbid=817177), but
it disappeared after I install update KB824145 ....
Does anyone know if that cumulative update actually fixes the problem?
I like empirical proof. If you have site which you know has not been
changed since you tested before installing the patch, which was not
working then and is working now it seems safe to assume that the thing
that changed, the version of mshtml.dll, is the cause of the new behaviour.
FWIW I have just tried retesting the sites which I assumed
demonstrated the problem and found that they now work properly.
Unfortunately, I have no knowledge of whether those sites have changed
since I first tested them (e.g. to avoid the problem).
They look the same but none of the .jpg have any of the headers
which I had been using as indicators of the problem.
Would the fix have changed the stored content of the .jpg?
The other reason for assuming that the new version may contain
the hotfix is this statement from the hotfix:
<quote>
The English version of this fix has the file attributes (or later)
that are listed in the following table.
....
Date Time Version Size File name
--------------------------------------------------------
31-Jul-2003 19:43 6.0.2800.1240 2,798,592 Mshtml.dll
</quote>
On two counts the version that the cumulative patch installs
satisfies that "or later"
6.0.2800.1276 2,799,104 10-16-2003 Mshtml.dll
It has a greater version and a later modification date.
All we have to assume is that the source for the hotfix
(a non-security matter) became integrated with the source
of the security patch (and its functionality not regressed).
As I said, seeing controlled tests working properly seems
more significant to me.
HTH
Robert Aldwinckle