Re: Somebody broke smokecheck xpathQuery tests

It's me.

I added a subview to TestCase, apparently a lot of tests depend on  
TestCase not having any subviews...

On 2008-09-30, at 18:41EDT, Henry Minsky wrote:

> Now that you mention it, test/lfc/data/alldata.lzx has these  
> errorsTests:
> 325 Failures: 10 Errors: 0
> TestFailure: test1 failed: Equals:  expected 33 got undefined
> TestFailure: test1 failed: Equals:  expected 66 got undefined
> TestFailure: test1 failed: Equals:  expected 55 got undefined
> TestFailure: test1 failed: Equals:  expected 33 got undefined
> TestFailure: test1 failed: Equals:  expected 22 got undefined
> TestFailure: test1 failed: Equals:  expected 3 got undefined
> TestFailure: test1 failed: Equals:  expected 6 got undefined
> TestFailure: test1 failed: Equals:  expected 5 got undefined
> TestFailure: test1 failed: Equals:  expected 3 got undefined
> TestFailure: test1 failed: Equals:  expected 2 got undefined
>
> I need to figure out which test suite that is.
>
> On Tue, Sep 30, 2008 at 6:12 PM, P T Withington <ptw <at> pobox.com> wrote:
>
>> ERROR @regression.lzl#738: Equals:  expected 4 got 5
>> INFO @regression.lzl#883: serialize: p is null in datapointer   
>> name: bad
>> INFO @regression.lzl#903: setNodeAttribute: p is null in  
>> datapointer  name:

For Review: Change 20080930-ptw-D Summary: Integrate Profiler fixes from Pagan-Deities to Trunk

[Let me know if you would prefer I just check this in for you to test.]

Change 20080930-ptw-D by ptw <at> dueling-banjos.local on 2008-09-30  
21:20:18 EDT
     in /Users/ptw/OpenLaszlo/honey-dew-2
     for http://svn.openlaszlo.org/openlaszlo/trunk

Summary: Integrate Profiler fixes from Pagan-Deities to Trunk

Bugs Fixed:
LPP-6960  Merge profiler fixes from Anubis (pagan) to trunk

Technical Reviewer: hminsky (pending)
QA Reviewer: mamye (pending)

Details:
     Integrated 10756,10774,10836,10841,10893,10935 from pagan-deities

     lzunit.lzx: Make the formatter delegate an attribute, not a
     subview as many tests assume TestCase has no subviews.

Tests:
     smokecheck, profiled test/styles/metasuite in swf8 and dhtml.

Files:
  M     .
M      WEB-INF/lps/lfc/profiler/LzProfile.lzs
M      WEB-INF/lps/lfc/services/LzInstantiator.lzs
M      WEB-INF/lps/lfc/LaszloLibrary.lzs
M      WEB-INF/lps/lfc/compiler/Class.lzs

Re: For Review: Change 20080930-ptw-D Summary: Integrate Profiler fixes from Pagan-Deities to Trunk

Re: For Review: Change 20080930-bargull-Peb Summary: services updates (part 3)

For Review: Change 20081001-lou-L Summary: Explorer: fix bad links and link text

Change 20081001-lou-L by lou <at> loumac.local on 2008-10-01 07:42:59 AST
     in /Users/lou/src/svn/openlaszlo/trunk
     for http://svn.openlaszlo.org/openlaszlo/trunk

Summary: Explorer: fix bad links and link text

Bugs Fixed: LPP-7076

Technical Reviewer: (pending)
QA Reviewer: (pending)
Doc Reviewer: (pending)

Details: see JIRA description

Tests: visual verify: test links

Files:
M      laszlo-explorer/coverpages/server_admin_cover.html
M      laszlo-explorer/coverpages/form_cover.html
M      laszlo-explorer/coverpages/general_cover.html

Changeset: http://svn.openlaszlo.org/openlaszlo/patches/20081001-lou-L.tar

For Review: Change 20081001-bargull-Wwh Summary: swf9 compatible array type check

Change 20081001-bargull-Wwh by bargull <at> dell--p4--2-53 on 2008-10-01 15:02:28
in /home/Admin/src/svn/openlaszlo/trunk
for http://svn.openlaszlo.org/openlaszlo/trunk

Summary: swf9 compatible array type check

New Features:

Bugs Fixed: LPP-7099

Technical Reviewer: max
QA Reviewer: (pending)
Doc Reviewer: (pending)

Documentation:

Release Notes:

Details:
The type check for Arrays needs to be "item is Array" instead of 
"item['length']".

Tests:

Files:
M lps/components/base/datalistselector.lzx
M lps/components/base/listselector.lzx

Changeset: 
http://svn.openlaszlo.org/openlaszlo/patches/20081001-bargull-Wwh.tar

Re: For Review: Change 20080928-bargull-iHa Summary: resize replication in swf9

Your great interest in the selection management makes you to the next 
reviewer for LPP-7094 to LPP-7098. 

After those bugs are resolved, I'll poke a bit a the 
datapath/replication-manager/dataselection-manager trio for the 
selection problem.

On 9/29/2008 10:44 PM, P T Withington wrote:
> Got it!
> 
> Thanks for the explanation.
> 
> Maybe there needs to be a more general mechanism for caching random view 
> properties on datanodes to solve the selection problem?
> 
> On 2008-09-29, at 16:27EDT, André Bargull wrote:
> 
>> No, we're talking about LzResizeReplicationManagers which are tuned 
>> LzLazyReplicationManagers. So, we've got only 10 views, but maybe a 
>> few hundred data-elements which are dynamically bound to a view 
>> (depending on the clipsize). Clones of LzLazyReplicationManagers need 
>> to have the same replication-size (either width or height - depends on 
>> the replication-axis), but clones of LzResizeReplicationManagers can 
>> have different replication-sizes. And when we rebound the data to a 
>> view, we need to restore the original size - which was stored in the 
>> data-element (can't be done on the view because we don't have a 1-1 
>> relation - lazy replication!).
>>
>> Concerning the selection-note:
>> the selection is also stored in the data-elements (the 
>> "sel"-property), and just as you'll notice: just one "sel"-property 
>> for multiple replications won't work that good...
>>
>>
>> On 9/29/2008 10:07 PM, P T Withington wrote:
>>> Now I'm even more curious!  Why is the size stored on the data node?  
>>> I'm guessing this is some optimization to maybe not adjust 
>>> replication if things don't actually change size, but isn't the right 
>>> place to make that optimization in the view itself by not sending a 
>>> size event if it doesn't change size?
>>> On 2008-09-29, at 15:55EDT, André Bargull wrote:
>>>> It's not about clones (LzNodes), but about LzDataNodes which can be 
>>>> replicated in different replication managers (selection wouldn't 
>>>> work properly, but that's a different story  ).
>>>>
>>>>
>>>> On 9/29/2008 8:27 PM, P T Withington wrote:
>>>>> I'm curious why there need to be multiple size properties on the 
>>>>> clones.  Can a clone actually be controlled by more than one 
>>>>> replicator?
>>>>> On 2008-09-28, at 14:53EDT, André Bargull wrote:
>>>>>> Change 20080928-bargull-iHa by bargull <at> dell--p4--2-53 on 
>>>>>> 2008-09-28 20:04:17
>>>>>> in /home/Admin/src/svn/openlaszlo/trunk
>>>>>> for http://svn.openlaszlo.org/openlaszlo/trunk
>>>>>>
>>>>>> Summary: resize replication in swf9
>>>>>>
>>>>>> New Features:
>>>>>>
>>>>>> Bugs Fixed: LPP-7082
>>>>>>
>>>>>> Technical Reviewer: hminsky
>>>>>> QA Reviewer: (pending)
>>>>>> Doc Reviewer: (pending)
>>>>>>
>>>>>> Documentation:
>>>>>>
>>>>>> Release Notes:
>>>>>>
>>>>>> Details:
>>>>>> LzResizeReplicationManager tried to add properties to a 
>>>>>> non-dynamic class (LzDataNodeMixin), which failed obviously.
>>>>>> Therefore I've added "__LZgetSize/setSize" to LzDataNodeMixin to 
>>>>>> track the view size for the replication manager.
>>>>>>
>>>>>>
>>>>>> Tests:
>>>>>> attached at bugreport
>>>>>>
>>>>>> Files:
>>>>>> M WEB-INF/lps/lfc/data/LzDataNode.lzs
>>>>>> M WEB-INF/lps/lfc/data/LzResizeReplicationManager.lzs
>>>>>>
>>>>>> Changeset: 
>>>>>> http://svn.openlaszlo.org/openlaszlo/patches/20080928-bargull-iHa.tar
>>>>>>
>>>>>>
>>>>
>>
> 
> 

For Review: Change 20081001-hqm-w Summary: remove obsolete section on canvas default font in dguide

Change 20081001-hqm-w by hqm <at> badtzmaru.home on 2008-10-01 11:48:12 EDT
     in /Users/hqm/openlaszlo/trunk
     for http://svn.openlaszlo.org/openlaszlo/trunk

Summary: remove obsolete section on canvas default font in dguide

New Features:

Bugs Fixed:

Technical Reviewer: liorio
QA Reviewer: (pending)
Doc Reviewer: (pending)

Documentation:

Release Notes:

Details:

There seems to already be an up to date section about setting the
default canvas font, (Chapter 20.1.3) using the <canvas font="xxx">
attribute. The later section 20.4.3 on setting the default font is
obsolete and no longer works as described, so I removed it.

Tests:

Files:
M      docs/src/developers/fonts.dbk
D      docs/src/developers/programs/fonts-$20.lzx
D      docs/src/developers/programs/fonts-$21.lzx

Changeset: http://svn.openlaszlo.org/openlaszlo/patches/20081001-hqm-w.tar

Re: For Review: Change 20081001-hqm-w Summary: remove obsolete section on canvas default font in dguide

For Review: Change 20081001-bargull-c2C Summary: selection management fixes

Change 20081001-bargull-c2C by bargull <at> dell--p4--2-53 on 2008-10-01 16:02:23
in /home/Admin/src/svn/openlaszlo/trunk
for http://svn.openlaszlo.org/openlaszlo/trunk

Summary: selection management fixes

New Features:

Bugs Fixed: LPP-7094, LPP-7095, LPP-7096, LPP-7097, LPP-7098

Technical Reviewer: ptw
QA Reviewer: (pending)
Doc Reviewer: (pending)

Documentation:

Release Notes:

Details:
LPP-7094: clearSelection() changed in Lz(Data)SelectionManager:
- store the previous selection and then create fresh variables to hold 
new selections
- this avoids a possible non-terminating state (see testcase at bugreport)
LPP-7095: fixed by LPP-7094 bugfix
LPP-7096: select() changed in LzSelectionManager:
- removed "if (s!=o)", which was meant as a speed-up (?), but turned out 
to be buggy and which was never true for LzDataSelectionManager,
because s is a LzView, whereas o is a LzDataNode.
LPP-7097: unselect() changed in LzDataSelectionManager:
- only call "setSelected(false)" when actually found a valid selection
LPP-7098: makeSelected() changed in LzDataSelectionManager:
- set first "singleClone" and then call "setSelected(true)"
- applied same fix in unselect() and __LZsetSelected()

Added indentation for LPP-2623.
Added typing for LPP-6049.
Used LzNode#__LZUID instead of LzNode#getUID() now that there is no 
longer the requirement to call the getter.

Tests:
a testcase for every bug is attached at the bugreports

Files:
M WEB-INF/lps/lfc/helpers/LzDataSelectionManager.lzs
M WEB-INF/lps/lfc/helpers/LzSelectionManager.lzs

Changeset: 
http://svn.openlaszlo.org/openlaszlo/patches/20081001-bargull-c2C.tar