Hi all, I first used 'dillo' many years go on a 486 running Mandrake Linux around 2004. Since then, forgot all about it. Until now - it is brilliant. Anyway, to the point. I run my own web server, and while debugging CGI perl scripts etc. dillo somehow caches the results from the output produced. Usually shift+[refresh] drops the cache and gets the new page, but the only way I get get dillo to do this is exit, and restart (no big deal with a 0.02 second start-up time :) ) Ideas? Thanks, Nick -- Gosh that takes me back... or is it forward? That's the trouble with time travel, you never can tell." -- Doctor Who "Androids of Tara"
Hi Nick, On Sun, Apr 03, 2016 at 08:31:14PM +0100, Nick Warne wrote:
Hi all,
I first used 'dillo' many years go on a 486 running Mandrake Linux around 2004. Since then, forgot all about it. Until now - it is brilliant.
Good to hear you like it, thanks.
Anyway, to the point. I run my own web server, and while debugging CGI perl scripts etc. dillo somehow caches the results from the output produced.
Usually shift+[refresh] drops the cache and gets the new page, but the only way I get get dillo to do this is exit, and restart (no big deal with a 0.02 second start-up time :) )
Ideas?
Strange, I just tested this trivial cgi script and a simple reload does the trick (button press or Ctrl-r). Does it work for you? -- Cheers Jorge.- -- #!/bin/sh # # Simple bash CGI script. # Return current time and date # # # HTTP doc type (2 newlines needed). echo "Content-type: text/html" echo # HTML content echo " <html><body> <PRE> `/bin/date` </PRE> </body></html> "
Hi Jorge, On 04/04/16 13:00, Jorge Arellano Cid wrote:
Usually shift+[refresh] drops the cache and gets the new page, but the only way I get get dillo to do this is exit, and restart (no big deal with a 0.02 second start-up time :) )
Ideas?
Strange, I just tested this trivial cgi script and a simple reload does the trick (button press or Ctrl-r). Does it work for you?
OK, basically on the html page with the form you enter some letters/words... hit submit (or 'process' in my case) and the perl CGI produces the results and the time taken. Now, if if you use the same words on the index page, dillo doesn't run the CGI, but produces the previous cached page - sure, a reload works, but that doesn't seem logical. Firefox doesn't do this, it runs the CGI each time, which is what I would expect on something that produces dynamic content. The page is here is you want to have a look: https://linicks.net/AnAdIc/ Nick P.S. ssl/TLS works well on my site! -- Gosh that takes me back... or is it forward? That's the trouble with time travel, you never can tell." -- Doctor Who "Androids of Tara"
On Mon, Apr 04, 2016 at 05:16:54PM +0100, Nick Warne wrote:
Hi Jorge,
On 04/04/16 13:00, Jorge Arellano Cid wrote:
Usually shift+[refresh] drops the cache and gets the new page, but the only way I get get dillo to do this is exit, and restart (no big deal with a 0.02 second start-up time :) )
Ideas?
Strange, I just tested this trivial cgi script and a simple reload does the trick (button press or Ctrl-r). Does it work for you?
OK, basically on the html page with the form you enter some letters/words... hit submit (or 'process' in my case) and the perl CGI produces the results and the time taken.
Now, if if you use the same words on the index page, dillo doesn't run the CGI, but produces the previous cached page - sure, a reload works, but that doesn't seem logical.
Logic depends upon the considered facts.
Firefox doesn't do this, it runs the CGI each time, which is what I would expect on something that produces dynamic content.
Cache policy is not trivial (e.g. check idempotency in HTTP). For instance, if you decide to try to reload whenever it may seem useful, you start having problems when users go offline (i.e. intermitent internet connections). BTW, that's how we got "offline mode"; Good&Simple right? Hmmm, no. Unfortunately it was nearly impossible to teach the proverbial grandma when to go offline and when online. Power users started to complain they were not able to go back and forward fast on already visited pages, and that pages dissappeared. Why, the browser issued a reload every time, cleaning the cache, and guess what, websites used the opportunity to refresh advertisement. At some point there were cache directives served in HTTP, but they got abused by advertising again (i.e. always reload means more money). And some sites decided to serve their pages with always reload directives, even some decided to issue automatic reloads after a certain a mount of time (more clicks/visits on the stats, more money for payed advertisingt, etc...). One of the reasons why people love to have many open tabs is that since the reload mess inception (and some other stuff as custom navigation), it became hard to tell whether you'll be able to see the same page again by going back and forward. So tabs are a solution to the new problem. As you may see from this short sample, a cache policy must consider lots of cases. -- Cheers Jorge.-
Hi Jorge, On 11/04/16 16:23, Jorge Arellano Cid wrote:
Now, if if you use the same words on the index page, dillo doesn't run the CGI, but produces the previous cached page - sure, a reload works, but that doesn't seem logical.
Logic depends upon the considered facts.
Firefox doesn't do this, it runs the CGI each time, which is what I would expect on something that produces dynamic content.
Cache policy is not trivial (e.g. check idempotency in HTTP).
For instance, if you decide to try to reload whenever it may seem useful, you start having problems when users go offline (i.e. intermitent internet connections). BTW, that's how we got "offline mode"; Good&Simple right? Hmmm, no. Unfortunately it was nearly impossible to teach the proverbial grandma when to go offline and when online. Power users started to complain they were not able to go back and forward fast on already visited pages, and that pages dissappeared. Why, the browser issued a reload every time, cleaning the cache, and guess what, websites used the opportunity to refresh advertisement.
At some point there were cache directives served in HTTP, but they got abused by advertising again (i.e. always reload means more money). And some sites decided to serve their pages with always reload directives, even some decided to issue automatic reloads after a certain a mount of time (more clicks/visits on the stats, more money for payed advertisingt, etc...).
One of the reasons why people love to have many open tabs is that since the reload mess inception (and some other stuff as custom navigation), it became hard to tell whether you'll be able to see the same page again by going back and forward. So tabs are a solution to the new problem.
As you may see from this short sample, a cache policy must consider lots of cases.
First, I wasn't complaining, just curious (as 10CC sang, I don't like Dillo, I love it!) - and as I was debuggng a perl script, in normal use this wouldn't be an issue as the same input will always produce the same output, so a cached page is always correct. Many thanks for a very detailed explanation of the way this issue got into a mess - I am sure you are the best person to steer Dillo away from it. Regards, Nick -- Gosh that takes me back... or is it forward? That's the trouble with time travel, you never can tell." -- Doctor Who "Androids of Tara"
I am not sure its the same problem, but I have problems with the CSS files. When some page loads CSS files and I press "Refresh" (or Ctrl+R), Dillo reloads only the content, but not the linked CSS files. This way, when I try to design some page to be compatible with Dillo, I have to restart it on every test. Other browsers sometimes have the same problem, but they have a way to refresh with cache clearing - Ctrl+F5 in Firefox. Dillo seems to not have such mode. On Sun, 3 Apr 2016 20:31:14 +0100 Nick Warne <nick at linicks.net> wrote:
Hi all,
I first used 'dillo' many years go on a 486 running Mandrake Linux around 2004. Since then, forgot all about it. Until now - it is brilliant.
Anyway, to the point. I run my own web server, and while debugging CGI perl scripts etc. dillo somehow caches the results from the output produced.
Usually shift+[refresh] drops the cache and gets the new page, but the only way I get get dillo to do this is exit, and restart (no big deal with a 0.02 second start-up time :) )
Ideas?
Thanks,
Nick -- Gosh that takes me back... or is it forward? That's the trouble with time travel, you never can tell." -- Doctor Who "Androids of Tara"
_______________________________________________ Dillo-dev mailing list Dillo-dev at dillo.org http://lists.dillo.org/cgi-bin/mailman/listinfo/dillo-dev
-- http://fresh.flatassembler.net http://asm32.info John Found <johnfound at asm32.info>
On Mon, Apr 04, 2016 at 08:09:22PM +0300, John Found wrote:
I am not sure its the same problem, but I have problems with the CSS files. When some page loads CSS files and I press "Refresh" (or Ctrl+R), Dillo reloads only the content, but not the linked CSS files. This way, when I try to design some page to be compatible with Dillo, I have to restart it on every test.
Other browsers sometimes have the same problem, but they have a way to refresh with cache clearing - Ctrl+F5 in Firefox. Dillo seems to not have such mode.
Yes dillo does not have such mode yet. It would be useful to have it though. -- Cheers Jorge.-
participants (3)
-
jcid＠dillo.org -
johnfound＠asm32.info -
nick＠linicks.net