Comments for radiant's radiant  

What what!! Nice work

the old lines 3 & 4 are setup to allow running features from an extension directory. Although I need to check these changes against that.

I've made a typo here: "outputs" should be "output"

oh to add to the weirdness; it only has to do with class. switching 'class': 'tab' to 'id': 'tab' works fine in both ie7/8.

it's even weirder. in quirks mode (i.e. ie7 "standards" mode) the original version works fine but in ie8 standards mode it fails. simply adding the class after creating the element works in both standards and quirks modes. i don't know enough about either but my guess is that prototype or lowpro rely on the quirky ie7 dom.

So IE8 needs the element to exist before the class name can be added?

Yes. I realized I actually cut that part out, so I'm working on adding it back it and I'll be fixing this syntax too

Shouldn't that be :after => "Pages"? No need to use symbols here.

Oh! Good catch.

I think L86 is suppose to be sh "#{sudo}gem uninstall"...

Awesome, hopefully we can get an implementation of region sets which has a similar feel to the new tabs API

awesome.. love the new look

bug:
in line #13 +cache
also remove from repo "radiant\vendor\rack-cache\lib\rack\cache"

the cache config file is now stored in tmp

See these related commits.
http://github.com/radiant/radiant/commit/981c1d689a64a32b6b76b4c9dd520c3089dca4de
http://github.com/radiant/radiant/commit/f36033c78e39a257769338206c6af14ae4ae18ba

If the database is faster than file_store and file_store is set, then we could do this conditionally

you might want to add mkdir tmp to the radiant:update task otherwise you'll get a:

Initializing configuration...FAILED
rake aborted!
No such file or directory - /private/tmp/config_ext/tmp/radiant_config_cache.txt

when upgrading an existing site.

why only ensure that the entity cache exists and why force cache/entity to exist even if tmp/cache/entity is set in the environment?

mikehale's point is the reason I added this. I posted the feature in March with no responses. http://groups.google.com/group/radiantcms-dev/browse_thread/thread/867762de7c033db9

I'll try to put together some benchmarks

It may be slower to use Rack::Cache on the disk, but the ability to plug in memcached or some other caching backend is pretty powerful.

Is this really faster than the database? I think we should benchmark this.

Nope, this was all me -- I merged prematurely & reverted before your metal commits.