[android-developers] Disabling IME word prediction

2011-02-22 Thread LinusOnline
I'm looking for a way to disable word prediction from application
code, but haven't had any luck so far. Found this thread on it:

http://www.anddev.org/post94060.html

Can anyone here shed any light on this issue? In particular, are IMEs
required to adhere to the flags I give it via setInputType(int type)?
(I'm guessing not.) Also, can I expect the standard Android IME to
adhere to them? The standard Android Japanese IME?

-- 
You received this message because you are subscribed to the Google
Groups "Android Developers" group.
To post to this group, send email to android-developers@googlegroups.com
To unsubscribe from this group, send email to
android-developers+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-developers?hl=en


[android-developers] Cancel IMEs word prediction from application code?

2011-02-16 Thread LinusOnline
I have an EditText in my application where it never makes sense to use
word prediction. Most IMEs have a setting to enable or disable word
prediction, but is there any way I, as an application programmer, can
have control over the mechanism? Can I disable word prediction
manually/temporarily, independently of what exact IME is being used?

Would appreciate any help with this. The word prediction is
interfering with my own functionality.

-- 
You received this message because you are subscribed to the Google
Groups "Android Developers" group.
To post to this group, send email to android-developers@googlegroups.com
To unsubscribe from this group, send email to
android-developers+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-developers?hl=en


[android-developers] Possible to temporarily disable IMEs word prediction?

2011-02-14 Thread LinusOnline
I have a text box (EditText) in my application, where it never makes
sense to use word prediction (in fact, it interferes with my own
functionality, in different ways depending on the IME). Is there a way
for me as an application programmer to tell the IME not to use word
prediction, even though its word prediction setting is enabled?

Appreciate any help with this.

-- 
You received this message because you are subscribed to the Google
Groups "Android Developers" group.
To post to this group, send email to android-developers@googlegroups.com
To unsubscribe from this group, send email to
android-developers+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-developers?hl=en


[android-developers] Re: Strange behavior in EditText box on Samsung devices

2011-02-03 Thread LinusOnline
I looked up AutoCompleteTextView, and sadly it's not what I need. The
product I'm working on already has its own custom-made drop down box
of completions, which is working as expected. What I need is to add
the rest of the completion, selected, directly after the text the user
has entered into the EditText box. This works as expected on all other
devices we've tested.

/ Linus

On 3 Feb, 15:11, LinusOnline  wrote:
> THanks for the tip, Kostya! No I had not considered that. Will look
> into it.
>
> On 31 Jan, 14:13, Kostya Vasilyev  wrote:
>
>
>
> > Linus,
>
> > Have you considered using AutoCompleteTextView instead of an EditText?
>
> > It provides a ready-made infrastructure for autocomplete, and probably
> > works on all devices.
>
> > -- Kostya
>
> > 31.01.2011 16:04, LinusOnline пишет:
>
> > > I'm struggling with some strange bahavior in EditText fields on the
> > >SamsungGalaxy Tab andSamsungGalaxy S. Has anyone else experienced
> > > similar behavior to this?
>
> > > Situation:
> > > I'm using an EditText as an input window for URLs. As the user enters
> > > a character in it, I want to autocomplete it, i.e., look up the string
> > > in a list of URLs, find the first match (if any), and change the
> > > contents of the EditText to that URL, while keeping a selection around
> > > the suggested part.
>
> > > So, the sequence is:
> > > 1. User enters character (f.i. the 'c' in 'www.c').
> > > 2. onTextChanged is called, notifies my own thread, and returns
> > > 3. My own thread then posts a Runnable in which I do the following:
> > > 4. Using setText, I change the text to, i.e.,www.cnn.com.
> > > 5. Using setSelection, I select the suggested part (i.e., 'nn.com')
>
> > > Problem OnSamsungGalaxy S (and possibly otherSamsungdevices),
> > > using theSamsungIME:
> > > It seems like every time a character is entered in the URL field, the
> > > text it is set twice; once immediately, and then "magically" again a
> > > few ms later. In the above case, the text gets reset to 'www.c'
> > > *after* I've autocompleted it. Delaying the posting of the Runnable
> > > which changes the text for a few (~10) ms (using postDelayed) works
> > > around this.
>
> > > Second problem, occurring so far on the Galaxy Tab and Galaxy S, and
> > > only in landscape mode, when theIMEis fullscreen, but with *any*
> > >IME:
> > > Even with the above workaround, the selection on the suggestion part
> > > of the URL is cleared (while the actual text remains). This time, it
> > > appears that the effect of the setText call in step 4 is repeated
> > > (again, "magically") after about 100 ms. Using another call to
> > > postDelayed, I can delay the call to setSelection a further 100 ms,
> > > which again works around the problem.
>
> > > Both workarounds are very ugly, timing-dependent solutions, and I'd
> > > rather find a better one.
>
> > > Usually, the magic text-set is accompanied by some nondescript lines
> > > of logcat error output.
>
> > > On the Galaxy Tab, usingSamsungIME, I get this:
> > > 01-31 21:25:49.500: ERROR/AxT9IME(20639): WordSymbInit: 1
> > > 01-31 21:25:49.508: ERROR/KeyLedTest(20639): KeyLedTest++
> > > 01-31 21:25:49.508: ERROR/KeyBoardLed(19113): UpdateState../sys/class/
> > > sec/keyboard/keyboard_led state false
> > > 01-31 21:25:49.512: ERROR/KeyBoardLed(19113): UpdateState../sys/class/
> > > sec/keyboard/keyboard_led state false finished
> > > 01-31 21:25:49.512: ERROR/KeyLedTest(20639): service.UpdateState0
> > > 01-31 21:25:49.512: ERROR/InputMethodService(20639): Keyboard State :
> > > 0
>
> > > With a differentIMEthanSamsung's, I only get the last one of these
> > > lines.
>
> > > OnSamsungGalaxy S, I only get an error when usingSamsung'sIME.
> > > What I get is this:
>
> > > 01-31 12:22:14.622: ERROR/AxT9IME(2583): WordSymbInit: 1
>
> > > Again, has anyone else experienced similar behavior to this?
>
> > --
> > Kostya Vasilyev -- WiFi Manager + pretty widget 
> > --http://kmansoft.wordpress.com

-- 
You received this message because you are subscribed to the Google
Groups "Android Developers" group.
To post to this group, send email to android-developers@googlegroups.com
To unsubscribe from this group, send email to
android-developers+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-developers?hl=en


[android-developers] Re: Strange behavior in EditText box on Samsung devices

2011-02-03 Thread LinusOnline
THanks for the tip, Kostya! No I had not considered that. Will look
into it.

On 31 Jan, 14:13, Kostya Vasilyev  wrote:
> Linus,
>
> Have you considered using AutoCompleteTextView instead of an EditText?
>
> It provides a ready-made infrastructure for autocomplete, and probably
> works on all devices.
>
> -- Kostya
>
> 31.01.2011 16:04, LinusOnline пишет:
>
>
>
>
>
> > I'm struggling with some strange bahavior in EditText fields on the
> >SamsungGalaxy Tab andSamsungGalaxy S. Has anyone else experienced
> > similar behavior to this?
>
> > Situation:
> > I'm using an EditText as an input window for URLs. As the user enters
> > a character in it, I want to autocomplete it, i.e., look up the string
> > in a list of URLs, find the first match (if any), and change the
> > contents of the EditText to that URL, while keeping a selection around
> > the suggested part.
>
> > So, the sequence is:
> > 1. User enters character (f.i. the 'c' in 'www.c').
> > 2. onTextChanged is called, notifies my own thread, and returns
> > 3. My own thread then posts a Runnable in which I do the following:
> > 4. Using setText, I change the text to, i.e.,www.cnn.com.
> > 5. Using setSelection, I select the suggested part (i.e., 'nn.com')
>
> > Problem OnSamsungGalaxy S (and possibly otherSamsungdevices),
> > using theSamsungIME:
> > It seems like every time a character is entered in the URL field, the
> > text it is set twice; once immediately, and then "magically" again a
> > few ms later. In the above case, the text gets reset to 'www.c'
> > *after* I've autocompleted it. Delaying the posting of the Runnable
> > which changes the text for a few (~10) ms (using postDelayed) works
> > around this.
>
> > Second problem, occurring so far on the Galaxy Tab and Galaxy S, and
> > only in landscape mode, when theIMEis fullscreen, but with *any*
> >IME:
> > Even with the above workaround, the selection on the suggestion part
> > of the URL is cleared (while the actual text remains). This time, it
> > appears that the effect of the setText call in step 4 is repeated
> > (again, "magically") after about 100 ms. Using another call to
> > postDelayed, I can delay the call to setSelection a further 100 ms,
> > which again works around the problem.
>
> > Both workarounds are very ugly, timing-dependent solutions, and I'd
> > rather find a better one.
>
> > Usually, the magic text-set is accompanied by some nondescript lines
> > of logcat error output.
>
> > On the Galaxy Tab, usingSamsungIME, I get this:
> > 01-31 21:25:49.500: ERROR/AxT9IME(20639): WordSymbInit: 1
> > 01-31 21:25:49.508: ERROR/KeyLedTest(20639): KeyLedTest++
> > 01-31 21:25:49.508: ERROR/KeyBoardLed(19113): UpdateState../sys/class/
> > sec/keyboard/keyboard_led state false
> > 01-31 21:25:49.512: ERROR/KeyBoardLed(19113): UpdateState../sys/class/
> > sec/keyboard/keyboard_led state false finished
> > 01-31 21:25:49.512: ERROR/KeyLedTest(20639): service.UpdateState0
> > 01-31 21:25:49.512: ERROR/InputMethodService(20639): Keyboard State :
> > 0
>
> > With a differentIMEthanSamsung's, I only get the last one of these
> > lines.
>
> > OnSamsungGalaxy S, I only get an error when usingSamsung'sIME.
> > What I get is this:
>
> > 01-31 12:22:14.622: ERROR/AxT9IME(2583): WordSymbInit: 1
>
> > Again, has anyone else experienced similar behavior to this?
>
> --
> Kostya Vasilyev -- WiFi Manager + pretty widget 
> --http://kmansoft.wordpress.com

-- 
You received this message because you are subscribed to the Google
Groups "Android Developers" group.
To post to this group, send email to android-developers@googlegroups.com
To unsubscribe from this group, send email to
android-developers+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-developers?hl=en


[android-developers] Strange behavior in EditText box on Samsung devices

2011-01-31 Thread LinusOnline
I'm struggling with some strange bahavior in EditText fields on the
Samsung Galaxy Tab and Samsung Galaxy S. Has anyone else experienced
similar behavior to this?

Situation:
I'm using an EditText as an input window for URLs. As the user enters
a character in it, I want to autocomplete it, i.e., look up the string
in a list of URLs, find the first match (if any), and change the
contents of the EditText to that URL, while keeping a selection around
the suggested part.

So, the sequence is:
1. User enters character (f.i. the 'c' in 'www.c').
2. onTextChanged is called, notifies my own thread, and returns
3. My own thread then posts a Runnable in which I do the following:
4. Using setText, I change the text to, i.e., www.cnn.com.
5. Using setSelection, I select the suggested part (i.e., 'nn.com')

Problem On Samsung Galaxy S (and possibly other Samsung devices),
using the Samsung IME:
It seems like every time a character is entered in the URL field, the
text it is set twice; once immediately, and then "magically" again a
few ms later. In the above case, the text gets reset to 'www.c'
*after* I've autocompleted it. Delaying the posting of the Runnable
which changes the text for a few (~10) ms (using postDelayed) works
around this.

Second problem, occurring so far on the Galaxy Tab and Galaxy S, and
only in landscape mode, when the IME is fullscreen, but with *any*
IME:
Even with the above workaround, the selection on the suggestion part
of the URL is cleared (while the actual text remains). This time, it
appears that the effect of the setText call in step 4 is repeated
(again, "magically") after about 100 ms. Using another call to
postDelayed, I can delay the call to setSelection a further 100 ms,
which again works around the problem.

Both workarounds are very ugly, timing-dependent solutions, and I'd
rather find a better one.

Usually, the magic text-set is accompanied by some nondescript lines
of logcat error output.

On the Galaxy Tab, using Samsung IME, I get this:
01-31 21:25:49.500: ERROR/AxT9IME(20639): WordSymbInit: 1
01-31 21:25:49.508: ERROR/KeyLedTest(20639): KeyLedTest++
01-31 21:25:49.508: ERROR/KeyBoardLed(19113): UpdateState../sys/class/
sec/keyboard/keyboard_led state false
01-31 21:25:49.512: ERROR/KeyBoardLed(19113): UpdateState../sys/class/
sec/keyboard/keyboard_led state false finished
01-31 21:25:49.512: ERROR/KeyLedTest(20639): service.UpdateState0
01-31 21:25:49.512: ERROR/InputMethodService(20639): Keyboard State :
0

With a different IME than Samsung's, I only get the last one of these
lines.

On Samsung Galaxy S, I only get an error when using Samsung's IME.
What I get is this:

01-31 12:22:14.622: ERROR/AxT9IME(2583): WordSymbInit: 1

Again, has anyone else experienced similar behavior to this?

-- 
You received this message because you are subscribed to the Google
Groups "Android Developers" group.
To post to this group, send email to android-developers@googlegroups.com
To unsubscribe from this group, send email to
android-developers+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-developers?hl=en


[android-developers] Cryptic ANR trace file from Android Market

2010-09-30 Thread LinusOnline
I'm analyzing some ANR traces that were reported by users of my
application via Android Market. I have some questions about ANR traces
in general:

How reliable are the traces? Most the traces I have make sense, but
many don't. There is one which doesn't mention any of my code at all,
like it's been created after my application was closed.

Other traces are missing one or more of the permanent threads of my
program. When, and under what circumstances, could a single thread
within a process be killed by the system? Or could my whole
application have been killed, while part of the trace went through
just before?

-- 
You received this message because you are subscribed to the Google
Groups "Android Developers" group.
To post to this group, send email to android-developers@googlegroups.com
To unsubscribe from this group, send email to
android-developers+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-developers?hl=en


[android-developers] VFY: arbitrarily rejecting large method

2010-03-22 Thread LinusOnline
After building my project with Java debug information, I get the
following "warning" from dalvikvm when launching my application:

VFY: arbitrarily rejecting large method

It seems the Dalvik VM's verifier is rejecting one of my methods. The
method is named in the warning, and it is indeed very large,
consisting of a long switch statement. I'd like to avoid breaking it
up if I can.

Is there a way to get the verifier to accept this method, including
debug information?

-- 
You received this message because you are subscribed to the Google
Groups "Android Developers" group.
To post to this group, send email to android-developers@googlegroups.com
To unsubscribe from this group, send email to
android-developers+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-developers?hl=en

To unsubscribe from this group, send email to 
android-developers+unsubscribegooglegroups.com or reply to this email with the 
words "REMOVE ME" as the subject.