Suggestions for the comic script

[Noxious Ninja]

I have a couple of suggestions for the comic script. I.e., index.php?do_command=show_strip

One of them fixes an annoyance with Firefox and may save some bandwidth, the other two are a bit of polish.

1) Enable proper client-side caching by making the script support HTTP conditional GET.

First off, you must send a Last-Modified header with each image. This can be done with something like

Code: Select all

$lastmodstring = gmdate ('D, d M Y H:i:s \G\M\T', filemtime ($imagefile));
header ('Last-Modified: ' . $lastmodstring);

Next, send an Etag header. This can be pretty much anything, so long as it is unique to each file. I personally use an MD5 hash.

Code: Select all

$etagvalue = '"' . md5_file ($imagefile) . '"';
header ('Etag: ' . $etagvalue);

Now, when the browser requests the image again, it will send an If-Modified-Since header and an If-None-Match header. You can check this against the last-mod date and etag, respectively, and, if they match, send back an HTTP 304 header and cease output.

Code: Select all

$headers = getallheaders();

if (isset ($headers["If-Modified-Since"])) {
   if (isset ($headers["If-None-Match"]) {
      if ($headers["If-Modified-Since"] == $lastmodstring) {
         if ($headers["If-None-Match"] == $etagvalue) {
            header ('HTTP/1.1 304 Not Modified');
            ob_end_clean ();
            exit ();
         }
      }
   }
}

Benefit: The comic will not have to be retransfered every single time it is requested. For example, if I check the site, and a new comic hasn't been put up yet, I will get the cached version. Or, if a strip refers to something from a few days ago, and I go back to that strip to check it out, it won't have to reload.

Note: IE seems to have some funky aggresive caching, and doesn't rely on this, but it makes Firefox (and me) happy.

For more info on HTTP Conditional GET, see here.


2) Send a filename for the file. Something like strip_id_YYYY-MM-DD

Code: Select all

header ('Content-Disposition:inline; filename="1310_2004-10-29.png"');

Benefit: When I right-click on an image to save it, I don't have to spend a moment figuring out what the proper name is. Plus, the filename contains the date the comic was published, so I can get a quick idea of how far apart things were.


3) Send a file length

Code: Select all

$image = file_get_contents ('$imagefile');
header ('Content-Length: ' . strlen ($image));

Benefit: Um, not much, actually. It's just easy to do, and makes the script a bit more complete. Oh, and it gives you good "leet haxxor" karma.


Barring any typos, the above should be correct. I hope you find it useful.

Cheers, and thanks for the great comic.

[Mae Dean]

I will forward this on to Ukyo... perhaps he'll see fit to implement these changes. We'll see. ^^

[Focus]

Whew.... I was afraid was going to be another "Greg should wear lampshades on his feet because I think it would be a funny comic" type of suggestion. I was thinking of a different type of script.

[Noxious Ninja]

[quote="Greg Dean";p="416188"]I will forward this on to Ukyo... perhaps he'll see fit to implement these changes. We'll see. ^^[/quote]

Cool.

BTW, I fixed a couple of typos in the above.

[billf]

[quote="Focus";p="416201"]Whew.... I was afraid was going to be another "Greg should wear lampshades on his feet because I think it would be a funny comic" type of suggestion. I was thinking of a different type of script. [/quote]

I was thinking the same thing.

[Killer-Rabbit]

Same. The words are too similar. I leave it to you billf to come up with new names that no one can mix up.

[dmpotter]

s/script/web site/

That's my suggestion. Well, it's in PHP, so I guess

Code: Select all

$thread_title = preg_replace("/script/", "web site", $thread_title);

is better.

[CyberEd]

/me was thinking in the same lines as these other people...
welcome to the forum btw

[BtEO]

While that would do fine dmpotter, might I suggest practising a slightly more polished form of regexps in future.
Script should probably have been bounded by \b to make sure "scripting" doesn't become "web siteing".
Also this one probably should be a case-insensitive search, so an "i" mode modifer would be needed, as would a "g" replace all occurances modifier in the first syntax (preg_replace() does that as default though).
To sum up

Code: Select all

s/\bscript\b/web site/ig

and

Code: Select all

$thread_title = preg_replace("/\bscript\b/i", "web site", $thread_title);

You should always match as tightly as possible even when you "know" you don't really need it, that way there's less risk of getting screwed later by making foolish assumptions.
And to think, at first I hated regexps.

[WCH]

Heh, yeah... I saw the name of the thread and immediately thought "Hoo boy, here we go... someone's getting flamed..."

[Mr.Shroom]

Wow. A thread actually worth Greg's time. Whodda thunk it?

[Rembrandt Q. Einstein]

not only that, but a suggestion for the comic thread that's actually worth Greg's time. You're going places newbie.

[Noxious Ninja]

Well?

[OniNeko]

Purple Tallest: You got your sandwhich.
Me: You got your necro.

[Deacon]

Comic script improvements can help, certainly, but does anyone know why RL always loads so slowly? It's like traffic is being crammed through a highly restrictive paketeer or something...