Thanks, Ben, that's a great help. We'll add the 'fixed' style to the table.
We've been gradually coming to similar conclusions for the server side and are looking for hooks in the interface to start sending the response earlier. Luckily this app runs mostly inside firewalls so we may able to skip the proxy subtleties for now. I really don't want to take apart the HTTPS interface just for this! :-) Jim ----- Original Message Follows ----- From: "Ben Tilly" <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] Cc: "Christopher Schmidt" <[EMAIL PROTECTED]>, [email protected] Subject: Spam:*, Re: [Boston.pm] Force browser rendering of a partial dataset? Date: Tue, 15 Jul 2008 10:52:45 -0700 > On Tue, Jul 15, 2008 at 5:03 AM, [EMAIL PROTECTED] > > <[EMAIL PROTECTED]> wrote: Christopher, > > Thanks for the reply. At one point a some years ago I > > believe if you defined the column widths of the table in > > the top row of cells browsers would render it as > > received since they didn't need to wait to receive that > > last row to calculate column widths in such a case. Do > > modern browsers no longer do this? > > To make this work you need to set the table-layout style > attribute to "fixed". Then browsers know they can start > rendering once the first row has been received. To be > safe put a width on the table and on all of the cells in > the first row. This works in IE 5 on, and is in the CSS > standards so I'd expect it to work reasonably well. > > On Tue, Jul 15, 2008 at 9:01 AM, Christopher Schmidt > <[EMAIL PROTECTED]> wrote: > > For the record, the problem you're trying to solve is > > probably something like 'force rendering of partially > > complete HTML' or something along those lines. (The > > problem is not strongly related to HTTP or anything > else.) > > I'm sorry but this is entirely wrong if the problem is the > time to generate his document. > > In most web environments the document does not start being > delivered to the client until the document is fully > present on the server. This is needed so that the > recommended content size header can be given to the > client. That means that if his table takes a long time to > generate, then the client won't start seeing anything > until the table is generated. > > To avoid this he would need to do a *lot* more work. > First he has to convince the web server to let him > generate all of the headers. How to do this varies widely > by server and by environment, for instance for CGI scripts > search for "no parse headers". If there are any proxies > in the way (eg reverse proxies in http accelerator mode - > see http://tinyurl.com/6xx3ct) he will need to bypass > them. And then he has to generate those headers. (CGI.pm > has some support for this.) > > In the process of doing this he will indeed have to get > down and dirty with the details of HTTP. > > Cheers, > Ben _______________________________________________ Boston-pm mailing list [email protected] http://mail.pm.org/mailman/listinfo/boston-pm
