Re: Lack of support for "If-Modified-Since"

mike mulligan (mulligan@northernlight.com)
Fri, 14 Feb 1997 10:10:43 -0500


Something else that may be going on here is the poor adherence to the date
formats that HTTP mandates:

Sun, 06 Nov 1994 08:49:37 GMT ; RFC 822, updated by RFC 1123
Sunday, 06-Nov-94 08:49:37 GMT ; RFC 850, obsoleted by RFC 1036
Sun Nov 6 08:49:37 1994 ; ANSI C's asctime() format

If you look at what comes back from real servers, you will find a number of
sloppy discrpencies: dashes around the month in the 1st and 3rd format, no
dashes in the 2nd format, single digit days or hours or minutes or seconds,
non-GMT time zones, non-English day and month names, and more.

Eventually you can write enough code to cope with most of all of this, but I
worry that when you later take the time_t and format it and send it back as
a Modified-Since, does the server code do the right thing? Is the server
expecting the standard format, or is it expecting a format consistent with
the Last-Modified it sent earlier?

- mike mulligan

>At 03:26 AM 2/14/97 +0000, Rob Hartill <robh@imdb.com> wrote:

>>On Thu, 13 Feb 1997, John W. James wrote:
>> Turns out that several of the servers (I know Netscape does this) require
>> that you feed back in the identical date that's on the document on the
>> other side for you to get back a 304. In other words, if (ifModifiedSince
>> == file.mtime()) return 304 else return 200

>I wonder if such short-cut programming is what's used to keep benchmark
>figures respectable. I certainly get the feeling that if the big boys
>were forced to follow the specs then their servers would grind to a
>complete halt.

_________________________________________________
This messages was sent by the robots mailing list. To unsubscribe, send mail
to robots-request@webcrawler.com with the word "unsubscribe" in the body.
For more info see http://info.webcrawler.com/mak/projects/robots/robots.html