Holger Freyther [Sun, 7 Dec 2008 00:50:38 +0000 (01:50 +0100)]
[keyboard] Add a manual layout switch button to the keyboard
As requested by Will add a switch button to the keyboard. This
was not yet tested on a device, overlapping of the "Board" and
the switch button is not handled. The removal of the clip rect
might have performance implications but we will have to figure
that out.
Holger Freyther [Sun, 7 Dec 2008 00:49:01 +0000 (01:49 +0100)]
Make gcc 4.3 with fortify happy...
value appears twice in the header...
Jeremy Chang [Tue, 2 Dec 2008 08:53:16 +0000 (16:53 +0800)]
[dialer] Use show() instead of showMaximized() when launching "select
contact" dialog.
showMaximized seems to mess up the geom of the "select contact"
dialog.
This is for fixing #2142. http://docs.openmoko.org/trac/ticket/2142
I-Fan, Chen [Fri, 28 Nov 2008 09:12:37 +0000 (17:12 +0800)]
remove unnecessary setting of default value.
Marco Trevisan [Thu, 27 Nov 2008 17:05:20 +0000 (01:05 +0800)]
Add the CallVolume setting to configuration (set by default at 4).
Marco Trevisan [Thu, 27 Nov 2008 17:04:40 +0000 (01:04 +0800)]
Update the callscreen volume slider according to the new
QPhoneProfile call volume feature.
Marco Trevisan [Thu, 27 Nov 2008 17:03:48 +0000 (01:03 +0800)]
The qtopia phone profile supported only the Volume setting
that is used for the ringtone volume.
To change the modem call volume has been needed a new setting
used only for the modem volume during calls.
The configuration files:
~/.config/Trolltech/PhoneProfile.conf
/opt/Qtopia/etc/default/Trolltech/PhoneProfile.conf
should now contain a new 'CallVolume' value for each profile.
Guillaume [Wed, 26 Nov 2008 08:42:22 +0000 (16:42 +0800)]
[phonebook] Show a warning when we modify a contact name when importing to sim
I-Fan, Chen [Sun, 23 Nov 2008 03:28:53 +0000 (11:28 +0800)]
Call Volume: Set the default Call Volume to 4 For Neo A7 need louder voice.
I-Fan, Chen [Sat, 22 Nov 2008 14:23:32 +0000 (22:23 +0800)]
Volume Control Slide Bar.
Make users can control the call volume via a slide bar.
According to http://doc.trolltech.com/qtextended4.4/qphoneprofile.html#setVolume the
range of volume is 0~5.
In ficgta01 phonevendor plugin we use AT+CLVL controls the TI calypso modem volume.
You can controls the default volume by setting the QPhoneProfile::setVolume.
erin [Tue, 18 Nov 2008 07:11:37 +0000 (15:11 +0800)]
[gsm/fitgta01]
1. change AT+CFUN=0 back in Ficgta01ModemService, since it would send AT+CFUN=1 in CellModemManager.
Therefore, we would remain AT+CFUN=0 and it can avoid plane flight mode problem. But, I remove
AT+COPS=0 in Ficgta01ModemService, coz it always returns ERROR and no reason to register network
if AT+CFUN=0.
2. change m_smsIsReady && m_phoneBookIsReady in cstatNotification, it would send 'sim ready' and
'phonebook ready' if both SMS & PhoneBook ready.
3. remove 'shouldSendSimReady' related code, it means we would send 'sim ready' if we get +CPIN:READY.
it can allow CellModemManager to handle network registration functions earliery. I verified it with
3 different SIM cards and it can dial out a call in 35 seconds. Both PhoneBook and SMS are working
normally too. I will keep tracking this change if it has any side effect later.
erin [Tue, 18 Nov 2008 03:37:42 +0000 (11:37 +0800)]
enable QtopiaServer log for debugging
erin [Wed, 5 Nov 2008 10:36:27 +0000 (18:36 +0800)]
[gsm/ficgta01] send SIM ready if we get %CSTAT: PHB,1 or %CSTAT: SMS, 1,
and it can retrieve phonebook data earlier in the beginning.
erin [Wed, 5 Nov 2008 10:12:48 +0000 (18:12 +0800)]
[gsm/ficgta01] set Phone full functionality from AT+CFUN=0 to AT+CFUN=1
and it can save about 12 sec to display the current operator in the
beginning.
Marco Trevisan (Treviño) [Wed, 15 Oct 2008 07:37:14 +0000 (15:37 +0800)]
[call echo] adding Ficgta01ModemHiddenFeatures, allows to send the known hidden at commands easily
Thanks Treviño
http://docs.openmoko.org/trac/ticket/1267
I-Fan, Chen [Wed, 15 Oct 2008 02:31:36 +0000 (10:31 +0800)]
[call echo] enable the echo cancle mode whenever a call starts.
Thanks for Treviño, and Mickey's suggestion.
https://docs.openmoko.org/trac/ticket/1267
Holger Freyther [Wed, 8 Oct 2008 16:12:28 +0000 (18:12 +0200)]
[keyboard] Fix compile
Holger Freyther [Wed, 8 Oct 2008 14:16:47 +0000 (16:16 +0200)]
[keyboard] Launch the predictive keyboard again
I was too fast for the rest of the system. So Om2008.10 can almost
change the keyboard but settings and illume will not be ready to do
so.
erin [Wed, 8 Oct 2008 11:23:15 +0000 (19:23 +0800)]
apply mwester patch for bug#1024 to send AT%SLEEP=2
I-Fan, Chen [Wed, 8 Oct 2008 09:58:04 +0000 (17:58 +0800)]
[CUSD] decode the cusd string with default codec.
https://docs.openmoko.org/trac/ticket/2029
We think the cusd will encoded with the TE codec format, so decode it with
default codec.
This is happened when the default is not HEX, now we use UCS2.
This is not tested yet.
I-Fan, Chen [Wed, 8 Oct 2008 09:54:27 +0000 (17:54 +0800)]
[Data Coding Scheme] Put the table of Data Coding Scheme according to 3GPP TS 23.038 section 5
Holger Freyther [Fri, 26 Sep 2008 19:02:00 +0000 (21:02 +0200)]
[gsm/ficgta01] Use CUSD instead of ATD to send supplementary messages
Based on patches from https://docs.openmoko.org/trac/ticket/2038.
Quote and encode the number according to our default charset. This is
not tested and will be tested once I'm back in germany.
Marco Trevisan (Treviño) [Fri, 26 Sep 2008 18:00:58 +0000 (20:00 +0200)]
[gsm] Move detecting of supplementary service detection to one place
Holger Freyther [Wed, 8 Oct 2008 02:20:49 +0000 (04:20 +0200)]
[gsm] Tell everyone what charset we intend to use
AT+CSCS=UCS2 can have a influence on many commands when the
default encoding is used. Now everyone can ask the QModemService
they are associated with about the charset.
QModemSupplemntaryServices will be the first user for CUSD handling.
Holger Freyther [Tue, 7 Oct 2008 14:18:20 +0000 (16:18 +0200)]
[gsm] Fixes from trolltech for CUSD handling that are in Qtopia4.3.3
- In
4f7df94616ffac06cc3927fd1d8c2dbf6de724af CUSD parsing was fixed for
a german default encoding case, a bug was filed at trolltech
- Now include the mostly unverified fix from Trolltech that was put
into Qtopia 4.3.3.
Bug report at Trolltech:
http://trolltech.com/developer/task-tracker/index_html?method=entry&id=212600
Bug report at Openmoko:
https://docs.openmoko.org/trac/ticket/2029
I wanted to take care of this when merging Qt4.3.3 (actually did in
master-4.3.3) but it looks like we will go for Qt Extended 4.4.x directly
so let us merge this patch and test it.
erin [Tue, 7 Oct 2008 09:25:49 +0000 (17:25 +0800)]
apply mwester patch for bug#1024 GSM modem re-registering problem
Holger Freyther [Tue, 7 Oct 2008 08:03:42 +0000 (10:03 +0200)]
[gta01] Do not autoadd a random path to the environment
This can break and did break when people actually had a toolchain
in this directory.
Holger Freyther [Tue, 7 Oct 2008 07:15:17 +0000 (09:15 +0200)]
[keyboard] Provide fdo desktop files for the keyboard application
These can be used by illume and settings to select
Holger Freyther [Tue, 7 Oct 2008 07:07:05 +0000 (09:07 +0200)]
[keyboard] Build the keyboards as standalone applications
- Build the predictive and stylus keyboard as application. They can be
launched by illume or any other window manager.
- The !x11 magic inside the .pro files might not work and qws will
build the keyboards as applications as well
Holger Freyther [Tue, 7 Oct 2008 06:03:37 +0000 (08:03 +0200)]
[keyboard] Remove the Keyboard Handler, keyboards will be external apps
Holger Freyther [Mon, 6 Oct 2008 17:32:29 +0000 (19:32 +0200)]
[keyboard] Allow to switch the key on the fly using a dbus interface
- Allow to switch between predictive and non predictive stylus keyboard
Holger Freyther [Mon, 6 Oct 2008 15:48:14 +0000 (17:48 +0200)]
[keyboard] Fix caption on the keyboard for X11
Remove #ifdef block. This used to contain QWS stuff but it does not
contain it anymore. This gives us a caption/label on the alpha keys
Holger Freyther [Mon, 6 Oct 2008 15:46:59 +0000 (17:46 +0200)]
[keyboard] Properly map the escape keycode
Use Qt::Key_Escape and Qt/X11 stops complaining that it can not map
the "keycode" to keysym. This needs to be tested on a device.
Holger Freyther [Mon, 6 Oct 2008 15:43:23 +0000 (17:43 +0200)]
[inputmethod] Map a key if it is not known to X
Inspired by Ecore_X use XChangeKeyboardMapping when KeySym->KeyCode
mapping is failing. Try to find a free keycode index and then set it. The
information about the calculation can be found in the man page.
Holger Freyther [Mon, 6 Oct 2008 07:06:35 +0000 (09:06 +0200)]
[keyboard] Compile the other keyboard implementation into Qtopia
Holger Freyther [Mon, 6 Oct 2008 07:59:32 +0000 (09:59 +0200)]
[keyboard] Be able to build the normal keyboard on Qt/X11
Holger Freyther [Sun, 5 Oct 2008 15:39:02 +0000 (17:39 +0200)]
[inputmethod] pred and keyboard have keyboard.{cpp,h}. stop to clash
Currently the keyboard implementations get compiled into the server
process. With VPATH and having the same filename this will not work. Move
things around
Holger Freyther [Sun, 5 Oct 2008 14:42:49 +0000 (16:42 +0200)]
[keyboard] Add the normal keyboard to the build process
Holger Freyther [Sun, 5 Oct 2008 14:39:04 +0000 (16:39 +0200)]
[inputmethod] Move QWSInputmethod/X11 into a library to be shared with pred and normal keyboard
Be able to share Qt -> X11 code to send key events
Guillaume [Mon, 6 Oct 2008 08:15:49 +0000 (16:15 +0800)]
[phonebook] Don't trim the name when exporting to sim if there is no extension
Should fix (at least part of) bug #1987
Holger Freyther [Sun, 5 Oct 2008 12:59:26 +0000 (14:59 +0200)]
[addressbook] Change AbSimEditor to look better in fullscreen mode
Switch to QFormLayout, this adds the proper spaceritem to the
end of the dialog and should keep everything else like it is.
Holger Freyther [Sun, 5 Oct 2008 10:15:47 +0000 (12:15 +0200)]
[suspend/resume] Do not suspend while we show an alarm
Introduce ODeviceUsage which is talking with the ompowerd
to lock any kind of resource.
Switch clock and datebook over to this class. Be friendly when
we try to unlock more often than it was locked, unlock on
destruction of ODeviceUsage.
Specially in alarmview.cpp it is hard to figure out when to
lock/unlock. Everytime we start/stop mAlarmTimer I added a
lock/unlock.
(untested)
Holger Freyther [Sun, 5 Oct 2008 03:50:07 +0000 (05:50 +0200)]
[clock] Play the alarm until the user is doing something with the device
Instead of only ringing a couple of times and then giving up
try the best to wake up the user. This involves ringing until the
user clicks ok. In the worse case the battery will go empty while
it is attempted to wake up the user but at least we tried. In the
majority of cases the alarm is set because it is important.
Holger Freyther [Sun, 5 Oct 2008 03:32:49 +0000 (05:32 +0200)]
[sim editor] Have a frame so that it is under windowmanager control
http://docs.openmoko.org/trac/ticket/1984
The AbSimEditor was set to frameless, which is bypassing the window
manager. So it might stay always on top. Treat it like a normal editor
window as it is one.
Holger Freyther [Sun, 5 Oct 2008 03:28:18 +0000 (05:28 +0200)]
[date] Disable setting hour format in our build
http://docs.openmoko.org/trac/ticket/1562
The setting is not shared between Qtopia and the rest of the
system. This is confusing and the most easy fix for now is to
disable setting this format.
Holger Freyther [Sun, 21 Sep 2008 20:15:03 +0000 (22:15 +0200)]
[contact selector] Optionaly enable the searching in the contact selector as well
https://docs.openmoko.org/trac/ticket/1966
Thanks to the code duplication we only changed this in the addressbook
application and thanks to the community the code duplication was spotted
and we could change the code at one other place.
This is mostly based on the patch of Treviño, the showEvent was added to
transfer the focus to the ListView.
Holger Freyther [Sun, 21 Sep 2008 19:45:18 +0000 (21:45 +0200)]
[qtopia] Trim the text for the contact searching
From https://docs.openmoko.org/trac/ticket/1966#comment:13
The virtual keyboard might add a space. When searching a contact
Foo you might end up with " Foo" or "Foo " but they both should match
Foo. So let us trim the space in the QTextEntryProxy (reciever of the
keyboard input).
Marco Trevisan [Fri, 19 Sep 2008 16:29:19 +0000 (18:29 +0200)]
[dialer] Have a entry to clear the number
https://docs.openmoko.org/trac/ticket/1997
Marco Trevisan [Fri, 19 Sep 2008 16:16:38 +0000 (18:16 +0200)]
[style] Reduce the size of the links
https://docs.openmoko.org/trac/ticket/1994 Reduce the size of links in the
thumbstyle.
I-Fan, Chen [Thu, 11 Sep 2008 16:03:37 +0000 (00:03 +0800)]
[PhoneBook] make UCS2 as default text codec
Guillaume [Thu, 11 Sep 2008 10:44:46 +0000 (18:44 +0800)]
[dialer] Skip mute action if the profile volume is zero on incoming call.
In fact what it really does is trigger the mute action automatically if the volume is zero.
Marco Trevisan [Wed, 10 Sep 2008 14:25:24 +0000 (16:25 +0200)]
[phonenumber] Wrong matching of itialian phone numbers
Btw, the Italian code rules in Qtopia are wrong since here the country
code is "+39" and the local area code (that starts with 0) could have a
length from 2 digits to 4 digits (so it can't be decoded easily).
This is from https://docs.openmoko.org/trac/ticket/1975
Guillaume [Wed, 10 Sep 2008 02:47:37 +0000 (10:47 +0800)]
[Addressbook] Shows a warning when we try to add a group that already exists.
erin [Tue, 9 Sep 2008 06:56:00 +0000 (14:56 +0800)]
add checking total used SMS number
Holger Freyther [Sat, 6 Sep 2008 21:00:08 +0000 (23:00 +0200)]
[addressbook] Allow to show the Findbar to find contacts
- Make showing the findbar configurable and default to
the current defaults. (Not shown in thumb style mode)
- Make ContactListPane have the default focus
- Set the input method hint on the QTextEntryProxy so that
with a click to focus policy the findbar gets focus and
the keyboard is shown. So by default most of the screen
is available for the contacts and only with a click the
keyboard comes up..
Put the following into ~/.config/Trolltech/Contacts.conf
or /opt/Qtopia/etc/default/Trolltech/Contacts.conf
[default]
EnableFindBar=1
Holger Freyther [Sat, 6 Sep 2008 20:58:13 +0000 (22:58 +0200)]
[qtopia] Abuse the QTextEntryProxy to get input itself
If the text entry proxy gets input itself it should handle it as well. By
default QTextEntryProxy is setting the focus policy to NoFocus but if a
consumer of the API is changing that it will continue to work. It will get
text input the Target and itself.
Holger Freyther [Fri, 5 Sep 2008 15:15:35 +0000 (17:15 +0200)]
[addressbook] Use QTOPIA_PER_CONTACT_RINGTONE for per contact/group ringtone
In
4c51f80648b1eaa074922219d7ecb182b5f65b99 Jeremy and me missed the
ringtone in the addressbook.h this should be fixed now.
Guillaume [Fri, 5 Sep 2008 03:56:56 +0000 (11:56 +0800)]
[addressbook] Don't show the group SetRingTone action in the menu.
I did it because we have a ticket about this : #1759. Since I am not sure we really want to remove this, I just commented
the addAction call, so to bring the action back just uncomment those 3 lines.
Guillaume [Fri, 5 Sep 2008 03:20:01 +0000 (11:20 +0800)]
[addressbook] Shows a warning if we try to add a contact when an other contact has the same name.
I added a mew method : AddressbookWindow::acceptEntry, that we can modify in the future for more condition testing -like
: adding a contact with the same number than an other-
Guillaume [Thu, 4 Sep 2008 08:53:03 +0000 (16:53 +0800)]
[Messages] Changed 'Back' label to 'Query' or 'Search', to close ticket #1496.
I guess the next thing to do would be to also add a 'cancel' option.
Guillaume [Thu, 4 Sep 2008 08:38:07 +0000 (16:38 +0800)]
Added preprocessor directives to allow compiling when MEDIA_SERVER is not defined
At a few places in the code the #ifndef blocs were ommited
Holger Freyther [Fri, 29 Aug 2008 21:51:49 +0000 (23:51 +0200)]
[buildfix] Make sure the build succeeds and remove the unit tests
Looks like someone at trolltech does not want to cross compile the
unit tests.
Holger Freyther [Fri, 29 Aug 2008 21:50:49 +0000 (23:50 +0200)]
[buildfix] Attempt to get qtestlib build for the target of pdutest...
Holger Freyther [Fri, 29 Aug 2008 14:23:13 +0000 (16:23 +0200)]
[ticalypso] Enable noise reduction (command from our mailinglist)
As discussed on the community mailinglist and pointed out on IRC this
is supposed to help with the voice quality. Give it a try.
Holger Freyther [Wed, 27 Aug 2008 14:17:56 +0000 (16:17 +0200)]
[SMS] Add test case for the PDU parsing
Add the PDUs from https://docs.openmoko.org/trac/ticket/1728 and some other
to test that the text and sender can be properly extracted.
Currently we pass, so if Tick's change regressed some other PDU and we get
a bug report we can fix it and make sure the other stuff does not break
./pdutest
********* Start testing of TestPDUParsing *********
Config: Using QTest library 4.3.4, Qt 4.3.4
PASS : TestPDUParsing::initTestCase()
PASS : TestPDUParsing::testPDUParsing()
PASS : TestPDUParsing::cleanupTestCase()
Totals: 3 passed, 0 failed, 0 skipped
********* Finished testing of TestPDUParsing *********
Holger Freyther [Tue, 26 Aug 2008 20:23:13 +0000 (22:23 +0200)]
[ficgta01] Fix phonebook init regression from
9971aa8a69e31057b0f02af43b4c20333a77f13e
When not waiting for RDY, 1 anymore my phone started to show the old
issue with not querying the addressbook correctly. For some reason somewhen
early in the init sequence a %CSTAT PHB, 1 comes by and we set
m_phoneBookIsReady to true.
To fix this issue and hopefully not regress in another way we keep
track of the status. So if phonebook goes from 0, 1 to 0 again we set
m_phoneBookIsReady is to false. This means we wait for the point when both
SMS and PHB are ready.
I-Fan, Chen [Tue, 26 Aug 2008 10:31:55 +0000 (18:31 +0800)]
[SMS] decode semi-octet strings with ALPHANUMBER correct
fix bug https://docs.openmoko.org/trac/ticket/1728
Holger Freyther [Thu, 21 Aug 2008 13:14:30 +0000 (15:14 +0200)]
[ficgta01] Fix the No registration bug (#1766)
In
8380a8534e802c152c160e6699634ad9c66b1951 the SIM initialisation was
postponed until after receiving "%CSTAT RDY, 1" from the modem but
according to https://docs.openmoko.org/trac/ticket/1766 for some people
RDY is never sent.
Postpone initialisation until after SMS and PHB are ready (or if RDY
comes first use that). That should resolve #1766 and still delay the init
enough to have listing of the phonebook working.
The code has an early exit for the case that status is not 1 and for the
%CSTAT EONS, 1 it might fail but we are not interested in that notification
anyway and it is doing no harm.
This is minimally tested.
Holger Freyther [Sat, 9 Aug 2008 19:42:40 +0000 (21:42 +0200)]
[atchat] Do not add an \r to the wakeup command to avoid many issues
- If an \r is needed this has to be added in the register command
call now.
- The removal allows to send wakeup commands that will wake the modem
but will never generate a response. So missed notifications or
not ignoring commands long enough and relating the "OK" to the wrong
command can be totally avoided.
Aug 8 17:21:50 om-gta02 user.notice root: AtChat : T : "ATE0"
Aug 8 17:21:51 om-gta02 user.notice root: AtChat : T : "AT%CSTAT=1"
Aug 8 17:21:51 om-gta02 user.notice root: AtChat : F : "OK" <- OK from ATE0
Aug 8 17:21:51 om-gta02 user.notice root: AtChat : T : "AT+CSCS?"
Aug 8 17:21:51 om-gta02 user.notice root: AtChat : F : "OK" <- OK from AT%CSTAT
Aug 8 17:21:51 om-gta02 user.notice root: AtChat : T : "AT+CIMI"
Aug 8 17:21:51 om-gta02 user.notice root: AtChat : F : "+CSCS: "IRA""
Aug 8 17:21:51 om-gta02 user.notice root: AtChat : F : "OK" <- OK from AT+CSCS?...
Holger Freyther [Sat, 9 Aug 2008 18:48:53 +0000 (20:48 +0200)]
[atchat] Write the log message before we possible emit signals...
In case of %CPI I want to see the log message before the handling
of that notification.
Holger Freyther [Sat, 9 Aug 2008 18:43:53 +0000 (20:43 +0200)]
[ti calypso] Change the init sequence to use ^Z and disable echo on startup
- Read away everything from the serial before we send our first command
- Send a ^Z and wait for OK
- Disable echo, QAtChat seems to be more happy with that.
Holger Freyther [Sat, 9 Aug 2008 14:54:45 +0000 (16:54 +0200)]
[qatchat] Add a paranoia check that we do not send commands while we are waking up
Do not prime any command while our wake up is in progress. This is a paranoia
mode to be able to find the root cause for issues like the one attached...
Aug 9 02:13:00 om-gta02 user.notice root: AtChat : T : "^Z"
Aug 9 02:13:01 om-gta02 user.notice root: AtChat : T : "AT%CTZV=1"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : F : "^Z"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : F : "OK"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : T : "AT%CPI=2"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : F : "AT%CTZV=1"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : F : "OK"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : T : "AT+COPS=0"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : F : "AT%CPI=2"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : F : "OK"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : T : "AT%CSTAT=1"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : F : "AT+COPS=0"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : F : "ERROR"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : T : "AT+CSCS?"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : F : "AT%CSTAT=1"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : F : "OK"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : T : "AT+CIMI"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : F : "AT+CSCS?"
Aug 9 02:13:02 om-gta02 user.notice root: AtChat : F : "+CSCS: "IRA""
Holger Freyther [Sat, 9 Aug 2008 14:39:20 +0000 (16:39 +0200)]
[ti calypso] Move the init to the write place.
QSerialIODeviceMultiplexer::chat is just a simple implementation and
not a good a place to execute more complex commands, it does not take
echo into account so we might get a "wrong" OK...
Move that to the vendor plugin that is using a real QAtChat. The only downside
is that AT+CFUN=0 gets executed after the AT+CBC (invoked from QModemIndicator
and created in QModemService) but that is fine for our use.
Holger Freyther [Sat, 9 Aug 2008 14:34:19 +0000 (16:34 +0200)]
[ti calypso] Use 0x1a for the wakeup command.
This is from mickeyl's ogsmd and is a command that will _never_
generate a response and will wake the modem up. The last part is
actually not quite true as an "\r\n" is added which will make the
modem generate an OK but that is okay as it is ignored by the
QAtChat. But there is no semantic attached which makes it better
over ATE0 (which should be send early in the init sequence).
Holger Freyther [Tue, 5 Aug 2008 16:59:48 +0000 (18:59 +0200)]
[modem] Make sure we call setCurrentOperator at least once
Currently we might or might not call setCurrentOperator (calling
AT+COPS=0). In tryDoReady we might change to the Ready state
and if we have no registration yet call doAutoRegister. doAutoRegister
might not end up calling setCurrentOperator as the AT+CREG=? query
will say that we are already registered. Make sure that we always
call setCurrentOperator at least once as this will execute some more
queries that we will need.
So even if the TI Calypso seems to auto register, make sure that we
send AT+COPS=0 to it. This might help with the loss of network registration
and the failure to find a network again.
AtChat : T : "AT+CREG=2"
AtChat : F : "OK"
AtChat : T : "AT+COPS=3,0"
AtChat : F : "OK"
AtChat : T : "AT+CREG?"
AtChat : F : "+CREG: 2,1,"0240","B53C""
AtChat : F : "OK"
AtChat : T : "AT+CSSN=1"
void CellModemManager::registrationStateChanged() reg state 0 1
void CellModemManager::tryDoReady() 2 10 2 1 1
void CellModemManager::doStateChanged(CellModemManager::State) 2 3 QValueSpaceObject(0x342308)
State 3 3 1 <- not call setCurrentOperator because of the above result
void CellModemManager::stopAutoRegisterTimer()
State 3 3 1
void CellModemManager::stopAutoRegisterTimer()
Holger Freyther [Tue, 5 Aug 2008 15:00:32 +0000 (17:00 +0200)]
[ti calypso] Another attempt at retrying to get the addressbook
If we said the phone book is ready but the selection or CPBR? failed there
are two options. One is that it failed because the phonebook is not
available or because the SIM/calypso is still upset. Assume that we
don't ask for non existing phone book's and just restart the query
Holger Freyther [Tue, 5 Aug 2008 14:40:17 +0000 (16:40 +0200)]
[gsm] Handle failing AT+CFUN? properly (
3eb865ecd10867f1ba5040e86dfaff6dcd509c37)
This is an extension to the fix in
3eb865ecd10867f1ba5040e86dfaff6dcd509c37.
The issue was not the ERROR return to AT+CFUN=1 but the failure to report
the level status properly. Change that by emitting a filaure when AT+CFUN?
fails and handle this by scheduling a retry for the CFUN.
This should stop flooding the modem with AT+CFUN requests until a SIM PIN
was entered but still help the upset modem to slowly get started.
Holger Freyther [Tue, 5 Aug 2008 14:10:52 +0000 (16:10 +0200)]
Revert "[phonebook] AT+CPBR=? might fail on modem init, force a retry"
This seems to make things worse...
This reverts commit
c017d3b09596d4ea1aa77e960214e4f3f6e558e3.
Holger Freyther [Tue, 5 Aug 2008 13:57:44 +0000 (15:57 +0200)]
[ti calypso] Reset does not work, at least set it to CFUN=0 and send %CWUP=1
%CWUP is supposed to power up parts of the protocol stack, whatever that
means but it might be helpful.
Also turn off the modem (setting to a state with minimal functionality).
This is a defined point for startup, in the case of a restart (due a
Qtopia crash) this means that the PIN needs to be entered again but that
looks fine.
Holger Freyther [Tue, 5 Aug 2008 13:15:00 +0000 (15:15 +0200)]
[ti calypso] Experimental change for modem initialisation
Sometimes the modem really don't want to talk to us. Attempt to force a
reset on startup to see what the modem is up to and if it wants to talk
to us. The idea is that we will reset before doing any other kind of
initialisation so the rest of the stack should not be affected.
Holger Freyther [Tue, 5 Aug 2008 13:09:10 +0000 (15:09 +0200)]
[pindialog] Make "Options" go away by setting a dummy menu
Set a dummy menu on the lineedit's to make the "Options" entry
on the contextlabel go away.
Holger Freyther [Tue, 5 Aug 2008 12:37:57 +0000 (14:37 +0200)]
[pindialog] Wording changes to fit into a line.
- Make the introduction fit into one line
- add a space for the SIM PUK message
Holger Freyther [Tue, 5 Aug 2008 12:31:16 +0000 (14:31 +0200)]
[pindialog] Do not recreate the layout all the time
Allocate the PUK Label and Entry as well but do not show it. This avoids
recreating the QFormLayout all the time.
Holger Freyther [Tue, 5 Aug 2008 12:06:49 +0000 (14:06 +0200)]
[keyboard] Show the OptionsWindow only when we have something to show
Holger Freyther [Tue, 5 Aug 2008 11:38:02 +0000 (13:38 +0200)]
[pindialog] Strip the entered PIN from any whitespace
Our virtual keyboards emits a space to separate two words. If the PIN
was entered incorrectly the first time, the second entry contained a
leading whitespace. We simply remove any whitespace at the beginning
or end of the entry.
Holger Freyther [Tue, 5 Aug 2008 11:37:15 +0000 (13:37 +0200)]
[pindialog] Show the '*' (asterisk) directly on password entry
This makes our QA team happy.
Holger Freyther [Tue, 5 Aug 2008 11:28:31 +0000 (13:28 +0200)]
[pindialog] Queue the CellModemManager connection
Delay the showing of the PIN dialog a bit due the queued
signal and slot connection.
Holger Freyther [Tue, 5 Aug 2008 11:00:45 +0000 (13:00 +0200)]
[call] Remove the developer "hints" from the error message
This is finishing
ea42173147759e5ffb0008420de581cfc4805ac2 to remove the
developer hints as they didn't help at all.
Julian_chu [Mon, 4 Aug 2008 21:22:42 +0000 (23:22 +0200)]
[mediaplayer] Remove one category to show the app only once on the illume desktop
with the xdg menu xml and illume/e-wm the mediaplayer was shown twice
on the desktop. Change that.
Holger Freyther [Sun, 3 Aug 2008 20:43:50 +0000 (22:43 +0200)]
[modem] AT+COPS=3,0 might fail on init, increase the retryChat count..
Part of the fix was to call simready at a later point of time
(after %CSTAT) the other part is to retry the COPS call more often.
The only users of retryChat are so far the one in this log message. The
only one we actually use is COPS and we want to retry this more often.
45 is just a case but this means we will try at least 45 times to
connect.
src/libraries/qtopiaphonemodem/qmodemcallprovider.cpp: d->service->retryChat( "AT+CLIP=1" );
src/libraries/qtopiaphonemodem/qmodemcallprovider.cpp: d->service->retryChat( "AT+COLP=1" );
src/libraries/qtopiaphonemodem/qmodemcallprovider.cpp: d->service->retryChat( "AT+CCWA=1" );
src/libraries/qtopiaphonemodem/qmodemnetworkregistration.cpp: d->service->retryChat( "AT+COPS=3,0" );
Aug 2 08:33:16 om-gta02 user.notice root: AtChat : F : "+CME ERROR: 100"
Aug 2 08:33:17 om-gta02 user.notice root: AtChat : T : "AT+COPS=3,0"
Aug 2 08:33:17 om-gta02 user.notice root: AtChat : F : "+CME ERROR: 100"
Aug 2 08:33:17 om-gta02 user.notice root: AtChat : T : "AT+COPS=3,0"
Aug 2 08:33:17 om-gta02 user.notice root: AtChat : F : "+CME ERROR: 100"
Aug 2 08:33:18 om-gta02 user.notice root: AtChat : T : "AT+COPS=3,0"
Aug 2 08:33:18 om-gta02 user.notice root: AtChat : F : "+CME ERROR: 100"
Aug 2 08:33:18 om-gta02 user.notice root: AtChat : T : "AT+COPS=3,0"
Aug 2 08:33:18 om-gta02 user.notice root: AtChat : F : "+CME ERROR: 100"
Aug 2 08:33:19 om-gta02 user.notice root: AtChat : T : "AT+COPS=3,0"
Aug 2 08:33:19 om-gta02 user.notice root: AtChat : F : "+CME ERROR: 100"
Aug 2 08:33:19 om-gta02 user.notice root: AtChat : T : "AT+CPBS=?"
Holger Freyther [Sun, 3 Aug 2008 20:09:47 +0000 (22:09 +0200)]
[phonebook] AT+CPBR=? might fail on modem init, force a retry
Assume that the phone storage does exist and retry if CPBR=? failed. This
can happen on modem initialisation. The existing "retry" code does not work
and instead of issueing 'AT+CPBS="SM";AT+CPBR=?' in one race free command
the current code will attempt to fixup any damage with reading a wrong
phonebook because the two commands get executed separately...
Any other/better solution would require to rewrite this code.
Holger Freyther [Sun, 3 Aug 2008 12:41:51 +0000 (14:41 +0200)]
[gstreamer] Set the element/pipeline to READY instead of PAUSED to avoid some fd leaks
There is still something fishy with the refcounting of GStreamer itself or
the way it is used. If we set the element to paused the pulse element will use
two new file descriptors that are never freed.. if we set it to READY we should
still get our preroll ("speed improvement") without leaking the fd's.
Holger Freyther [Sun, 3 Aug 2008 10:42:18 +0000 (12:42 +0200)]
[ringtone] Replace the alarm.wav with junggle_btn402.mp3 of freesound.org
Repeath the sample four more times and export to wave using audacity. This
sounds better on a neo as alarm sound.
http://www.freesound.org/samplesViewSingle.php?id=26777
This work is licensed under a Creative Commons Sampling Plus 1.0 License.
Holger Freyther [Sun, 3 Aug 2008 10:34:23 +0000 (12:34 +0200)]
[modem] Allow to postpone "simready" until %CSTAT notifications arrived
Do not directly send "simready" in QModemPinManager but allow to postpone
sending this message. For the TI Calypso I want to delay it until the
modem reports that the SMS and Phonebook is ready. This should help
with the various failures I see on initialisation with my base card.
Possible breakage might happen when RDY never gets emitted. I have no
idea if that is ever going to happen but we will have to find that out
the hardway.
Aug 2 08:33:19 om-gta02 user.notice root: AtChat : T : "AT+COPS=3,0"
Aug 2 08:33:19 om-gta02 user.notice root: AtChat : F : "+CME ERROR: 100"
Aug 2 08:33:19 om-gta02 user.notice root: AtChat : T : "AT+CPBS=?"
Aug 2 08:33:19 om-gta02 user.notice root: AtChat : N : "%CSTAT: PHB, 1"
Aug 2 08:33:19 om-gta02 user.notice root: AtChat : N : "%CSTAT: RDY, 1"
Aug 2 08:33:19 om-gta02 user.notice root: AtChat : F : "+CPBS: ("EN","BD","FD","DC","LD","RC","LR","MT","AD","SM","SD","MC","LM","AF","ON","UD")"
Aug 2 08:33:19 om-gta02 user.notice root: AtChat : F : "OK"
Aug 2 08:33:19 om-gta02 user.notice root: AtChat : T : "AT+CPBS="EN""
Aug 2 08:33:19 om-gta02 user.notice root: AtChat : F : "OK"
Aug 2 08:33:19 om-gta02 user.notice root: AtChat : T : "AT+CPBR=?"
Aug 2 08:33:19 om-gta02 user.notice root: AtChat : F : "+CME ERROR: 100"
Holger Freyther [Sat, 2 Aug 2008 21:42:10 +0000 (23:42 +0200)]
[bugs] Add some more known bugs
Holger Freyther [Sat, 2 Aug 2008 21:33:52 +0000 (23:33 +0200)]
[bugs] Remove scroll indicator and mediaserver from the list
- Scrollindicator should be resolved due the major changes in the
Qt/X11 backingstore handling
- mediaserver. We have exchanged the ringtone and played a bit
with the GStreamer backend..
Holger Freyther [Sat, 2 Aug 2008 16:15:37 +0000 (18:15 +0200)]
[predictivekeyoard] Make the whole AcceptWindow optional and do not use it on X11
Holger Freyther [Sat, 2 Aug 2008 09:26:41 +0000 (11:26 +0200)]
[predictivekeyboard] Never animate the AcceptWindow on X11
We are unlikely to have ARGB visuals on the X server and we do not
know the on screen position of the text input.
With these two issues it is better to not animate the text on X11.
Holger Freyther [Fri, 1 Aug 2008 22:39:27 +0000 (00:39 +0200)]
[gstreamer] Fix the refcounting of the bus
gst_element_get_bus will ref the bus and we need to unref this
bus somewhere. Do it in the d'tor of the playbin session after
having deleted the buswatcher.
Holger Freyther [Fri, 1 Aug 2008 21:49:10 +0000 (23:49 +0200)]
[ringtone] Do not recreate the QSoundControl if it is not needed
Do not recreate the QSoundControl if the new and old filename is matching.
This is an improvements and works around a fd leak of gstreamer or the
Qtopia gstreamer media engine.