Recent changes to bugs

VideoTransformationTest requests focus in illegal way

Anonymous — Wed, 30 Mar 2005 08:49:03 -0000

Above container requests focus when it's not visible.
Focus cannot be granted for such component (regardless
it's set to visible after request), so key listener
gets no AWT events (colour keys do not work).

Optimization in AnimationTest

Donat DEMANET — Mon, 01 Mar 2004 13:24:15 -0000

In MHP 1.0.2 and 1.0.3, the default compositing rule is
SRC_OVER as stated in the description of class
DVBGraphics in Annex U of the MHP specification. As
pointed out in the same section, this rule does not
give the highest performance.

In AnimationTest, the background is first drawn with
the default rule and then the rule is explicitly set to
SRC_OVER to paint the butterfly. This means that
SRC_OVER is always used. Performance would be better if
the rule was set to SRC to paint the background and to
SRC_OVER only to paint the butterfly on top of it.

All mandatory key events except VK_Teletext are tested

Anonymous — Fri, 30 Jan 2004 15:15:26 -0000

The key event test, tests all MHP mandatory keys,
except VK_TELETEXT

Assumes default encoding Latin1

Anonymous — Fri, 30 Jan 2004 14:49:36 -0000

MHPTester does not run on MHP conform implementations

MHP 1.0.2 and MHP1.0.3 define in 11.2.11:
"Where the speciﬁcation of the Java APIs refers to the
default character encoding of the platform, the default
for MHP shall be "UTF8" as deﬁned in 7.1.5, "Monomedia
format for text" on page 52. The encoding "latin1" shall
also be supported as deﬁned in ISO 8859-1 [20]."

However, on MHP implementations with default character
encoding UTF8, MHPTester does not start, but throws
the following exception:

sun.io.MalformedInputException
at sun.io.ByteToCharUTF8.convert()
at java.io.InputStreamReader.convertInto()
at java.io.InputStreamReader.fill()
at java.io.InputStreamReader.read()
at java.io.InputStreamReader.read()
at nanoxml.XMLElement.readChar()
at nanoxml.XMLElement.scanPCData()
at nanoxml.XMLElement.scanElement()
at nanoxml.XMLElement.scanElement()
at nanoxml.XMLElement.scanElement()
at nanoxml.XMLElement.parseFromReader()
at nanoxml.XMLElement.parseFromReader()
at com.mindonmove.mhptester.Util.getParsedXML()
at
com.mindonmove.mhptester.Config.parseAppConfiguration
()
at com.mindonmove.mhptester.Config.<init>()
at com.mindonmove.mhptester.Controller.start()
at com.mindonmove.mhptester.Controller.start()
at
com.mindonmove.mhptester.MHPTesterXlet.startXlet()

Proposed solution: make sure all input files are encoded
in UTF8 format.
Alternatively, explicitly set text encoding to Latin1 for
InputStreamReaders, Strings, etc.
However, in the latter case, I strongly advice to add an
explicit test to check the default character enconding.

signed and unsigned xlets may not communicate through ixc

Donat DEMANET — Wed, 25 Jun 2003 14:05:43 -0000

Ixc tests try to communicate between a signed and an
unsigned xlet. Signed xlets may only communicate with
signed xlets. Unsigned xlets may only communicate with
unsigned xlets. See 12.6.2.14 in MHP BlueBook A057 rev2.

signed and unsigned xlets may not communicate through ixc

Donat DEMANET — Mon, 23 Jun 2003 12:01:23 -0000

getInterface(addr) can return NULL

Donat DEMANET — Mon, 23 Jun 2003 09:50:26 -0000

In the return channel tests, to get an RCInterface, the
following call is used:
RCInterface rcinterface =
manager.getInterface(addr);
In Bluebook a057rev2, it is stated that NULL is
returned if the interface to be used is not yet known.

The tests do not take this possible return value into
account.

In this case, one possible solution would be to use the
first element of the array returned by getInterfaces().

Font Test Tabulation

Anonymous — Thu, 24 Apr 2003 10:16:49 -0000

Just to let you know,

In the font test, there is a part asking if the text
presented to the user is readable.
The text to be presented uses tabs, but tabulation
space shown in the example ( I believe you use a JPEG
picture ) and the one presented by the platform are not
the same.
Default tabulation is 56 points ( what a point is depends
on pixel aspect ratio - e.g. for a square pixel graphics
device it would be 1 pixel =1 point) and this value is
used if you construct HTextLayoutManager() - with no
arguments.

So consequently the pictures on the left and right will
not really look the same.
The question user has to answer is "Is the same text
readable on both sides?". I guess the point is that user
can read the text and not compare the spacing?

FontTest problems

Anonymous — Wed, 02 Apr 2003 14:40:58 -0000

The test in question is the FontTest.

The code goes like this ( method init() ):

/* Examine the name by which the local font goes by */
this.toolkit = Toolkit.getDefaultToolkit();
this.fonts = this.toolkit.getFontList();
this.fontTestOk[0] = false;
for( int i = 0; i < this.fonts.length; i++ ){
if( this.fonts[i].compareTo(
this.reqLogicalFontName ) == 0 )
this.fontTestOk[0] = true;
}

if( this.fontTestOk[0] ) {

Font fnt = new Font( this.reqLogicalFontName,
Font.PLAIN, 26 );
if( this.reqFontFamilyName.compareTo(
fnt.getFamily() ) == 0 ){
this.fontTestOk[0] = false;
}
if( this.reqFontFaceName.compareTo(
fnt.getName() + " "
+ fnt.getStyle() ) == 0 ){
this.fontTestOk[0] = false;
}
}

So the test checks if the logical name (SansSerif) , the
family name
(Tiresias) and the font face name (Tiresias PLAIN)
correspond to the values
in the local font.

At the end of the code this.FontTestOK[0] will be
checked and if it is true
the loading font test will pass.

So:
when comparing the family name and the font face
name, in the case they
are the same (compareTo = 0) the test sets
this.fontTestOk[0] to
false?!?!?!?
Shouldn't it then be compareTo != 0

- In the checking of the font face name ("Tiresias
PLAIN") the test
constructs this:
fnt.getName() + " " + fnt.getStyle().

BUT: getName -> returns the logical name of the font
getStyle -> returns an integer value representing
the stile of the
font ( PLAIN = 0)

So after constructing the string will look like: "SansSerif
0" which is
completely opposite to expected "Tiresias PLAIN"

pedja@sonycom.com

oidString should be aidString

Donat DEMANET — Thu, 20 Mar 2003 16:52:47 -0000

in ApplaunchTest.java, at line 172, oidString should be
changed to aidString, e.g.:

ids[0] = (String)context.getXletProperty(
IACTestXletValues.oidString );
ids[1] = (String)context.getXletProperty(
IACTestXletValues.oidString );

should be:

ids[0] = (String)context.getXletProperty(
IACTestXletValues.oidString );
ids[1] = (String)context.getXletProperty(
IACTestXletValues.aidString );