[Cosmo-dev] Gzipping dojo.js

[Adam Christian]

Interesting, I did not know it worked that way. Sounds like the  
server side fix is the best option for everyone anywhoo.


On Feb 9, 2007, at 11:48 AM, Matthew Eernisse wrote:

> Since this is the initial page-level script tag for including the  
> JavaScript, I'd much prefer to do it with a server-side tag.
>
> Using document.write is pretty much the 'hack of last resort,' and  
> something which in theory is not even usable in XHTML:
>
> http://www.w3.org/MarkUp/2004/xhtml-faq#docwrite
>
> Of course it does actually work in all our supported browsers, but  
> I'd really prefer not to use it if we can avoid it -- particularly  
> for something like the base script tag.
>
> And as a BTW -- I'm not totally sure, but it looks like in Adam's  
> example that IE will still be grabbing the gzipped file.
>
> <script type="text/javascript" src="foo.js"></script>
>
> If you have both foo.js and foo.js.gz in the directory, I think any  
> browser that can handle gzipped content (including IE6, because it  
> thinks it can support it) will try to grab foo.js.gz first if it can.
>
> The reason it works may be that Adam has an up-to-date version of  
> IE6 where these bugs are fixed.
>
> To get it to work the way we want, we'd need three files --  
> dojo.js, dojo.js.gz -- and then a filepath specifically for IE, so  
> it doesn't try to grab the gz file --  
> dojo_dont_grab_the_compressed_file.js or some such.
>
>
> M.
>
> Jared Rhine wrote:
>> Matthew Eernisse wrote:
>>> Bottom line is that we need to be able to ensure that IE6 always  
>>> gets the unzipped file -- regardless of what type of content it  
>>> purports to accept.
>> A reasonable caveat thanks.  It seems entirely feasible to  
>> implement needed behaviors based on user-agent; at least it would  
>> with Apache.
>> Feel free to test against http://lab.osaf.us/cosmo/, where  
>> mod_deflate is enabled with no special treatment of browsers (so  
>> it should break on IE6).  We can confirm both broken behaviors and  
>> server-side exclusions of specific browsers when I get to it.
>> Just saw Adam's suggestion, which seems workable as well.  I had  
>> originally envisioned having browsers just request dojo.js and  
>> have the server negotiate resources, instead of browsers  
>> requesting different versions.
>> -- Jared
>> _______________________________________________
>> cosmo-dev mailing list
>> cosmo-dev at lists.osafoundation.org
>> http://lists.osafoundation.org/mailman/listinfo/cosmo-dev
>
> _______________________________________________
> cosmo-dev mailing list
> cosmo-dev at lists.osafoundation.org
> http://lists.osafoundation.org/mailman/listinfo/cosmo-dev