Changes between Initial Version and Version 1 of Ticket #15616, comment 14


Ignore:
Timestamp:
Dec 20, 2012, 5:30:47 PM (5 years ago)
Author:
Adam Peller
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #15616, comment 14

    initial v1  
    1 I agree with everything @neonstalwart said.
    2 
    3 I'm not sure what this ticket is saying is broke. Bill pointed out some stuff here: http://bugs.dojotoolkit.org/ticket/14587#comment:13, is that the issue?
    4 
    5 The dojo/i18n and dojo/text modules implement a super set of the API of the equivalent RequireJS modules. The extra functionality is only to support legacy stuff.
    6 
    7 @peller:
    8 As far as the loader goes, the only times proprietary machinery is required in the loader are when modules depend on functions not available in standard AMD loaders--hopefully only to support some legacy requirement. dojo/ready is the prime example. Another, extremely painful example are the requirements of the i18n module, which must load code built with prior builders, the current builder, unbuilt and then load with any of async amd, legacy sync, and legacy async...sometimes changing load mode on the fly.
    9 
    10 fwiw, I tried to simplify the latter situation by not including the preload machinery in 1.7 (imo, it is of questionable value with a properly structured, built, AMD project). This idea was rejected by others and we now have the very-complicated i18n module.
    11 
    12 
    13 @jburke
    14 I'm pretty sure AMD is not interested in implementing necessary machinery to determine when all modules are loaded. imo, this machinery is not needed for AMD. Do you agree?
    15 
    16 
    17 
    18 
     1harrisreynolds, if you're able to reproduce the problem on trunk, please open a new ticket.  (Just noticed that Bill's fixes went into trunk, not 1.8)  An example would be especially helpful.