Now, you'd get a runtime error. Before you'd get a hard compiler error because the tags would be missing from the schema...

On 10/21/10 7:32 AM, P T Withington wrote:
Are you saying you removed the top-level switch clause that Henry mentioned in 
his comments?  If so, what exactly happens if you include these components in 
DHTML now?

On 2010-10-21, at 04:09, Max Carlson wrote:

I removed the switch altogether.  It seemed like the right thing at the time - 
glad you agree!

We can add the warnings back in for DHTML later...


On 10/20/10 10:59 AM, P T Withington wrote:
I think we need to remove that<switch>   altogether, and write the 
implementation instead in a style like this:

   <class name="rtmp">
      <handler name="oninit">
        if ($as3) {
          ;
         } else if ($debug) {
           Debug.error("RTMP service is not available in this runtime");
         }
      </handler>

     etc.

Another approach would be to write the component like drawview, by writing an 
interface that specifies the schema and then having a totally script 
implementation of the class:

   if ($as3)
     class $lzc$class_rtmp { ... }
   } else if ($debug) {
     Debug.error ("RTMP service is not available in this runtime");
   }

Clearly this latter would be easier if we had better support for writing tag 
classes in script that did not require wizard knowledge about the boilerplate 
and naming conventions.

On 2010-10-07, at 23:24, Henry Minsky wrote:

approved.

That gets rid of the<switch>    statement for Flash runtimes, but we still
have to come up with a workaround for the remaining<switch>   clause which is
at the start of most of the av components which nulls out the code for
DHTML.

  <switch>
    <when runtime="dhtml">
      <!-- no support for RTMP in DHTML -->
    </when>
    <otherwise>

I can't think of anything really clever right now. Maybe we need a flavor of
<switch>   which compiles all of it's clauses into script blocks
conditionalized as  "if ($something) ... " , as opposed to the way it works
now where the parser evaluates the<switch>   conditions at parse time.



On Thu, Oct 7, 2010 at 9:45 PM, Max Carlson<[email protected]>   wrote:

Change maxcarlson-20101007-pI7 by [email protected] on 2010-10-07
19:42:47 MDT
   in /Users/maxcarlson/openlaszlo/trunk2
   for http://svn.openlaszlo.org/openlaszlo/trunk

Summary: Use<passthrough when="">   instead of<switch>   in components

Bugs Fixed: lPP-4932 - Use<passthrough when="">   instead of<switch>   in
components

Technical Reviewer: hminsky
QA Reviewer: mdemmon

Details: I'm going to check this in, hoping it makes LZOs that include
components.

Move passthroughs outside switch statements.

Tests: Visual inspection.

Files:
M       lps/components/incubator/uploader/fileupload.lzx
M       lps/components/extensions/printer.lzx
M       lps/components/extensions/av/microphone.lzx
M       lps/components/extensions/av/camera.lzx
M       lps/components/extensions/av/mediastream.lzx
M       lps/components/extensions/av/mediadevice.lzx
M       lps/components/extensions/av/videoview.lzx
M       lps/components/extensions/av/rtmpconnection.lzx
M       lps/components/utils/diagnostic/inspector/inspector.lzx
M       lps/components/utils/diagnostic/memorymonitor.lzx
M       lps/components/av/videoscreen.lzx

Changeset:
http://svn.openlaszlo.org/openlaszlo/patches/maxcarlson-20101007-pI7.tar




--
Henry Minsky
Software Architect
[email protected]


Reply via email to