Re: Patch for better Samsung B2100 support

[Pawel Kot]

Hi,

On Sat, Jul 16, 2011 at 21:46, grunge <address@hidden> wrote:
> Some communication is attached, however it gets somehow complicated as
> two new problems arose"

Can you please show also some successful communication?

> 1) when using writephonebook to popullate empty phone memory,
>   gnokii fails to (attempt to) write the very first entry:
>
>        Initialisation completed
>        Message sent: 0x00 / 0x000d
>        41 54 2b 43 50 42 53 3d 22 4d 45 22 0d          | AT+CPBS="ME"
>        write: [AT+CPBS="ME"<cr>]
>        read : [AT+CPBS="ME"<cr><cr><lf>OK<cr><lf>]
>        Message received: 0x00 / 0x0013
>        02 41 54 2b 43 50 42 53 3d 22 4d 45 22 0d 0d 0a |  AT+CPBS="ME"
>        4f 4b 0d                                        | OK
>        Received message type 00
>        Message sent: 0x67 / 0x000a
>        41 54 2b 43 50 42 52 3d 3f 0d                   | AT+CPBR=?
>        write: [AT+CPBR=?<cr>]
>        read : address@hidden<cr><cr><lf>+CME ERROR: 23<cr><lf>]
>        Message received: 0x67 / 0x001c
>        05 00 17 2b 43 50 42 52 3d 3f 0d 0d 0a 2b 43 4d |    +CPBR=?   +CM
>        45 20 45 52 52 4f 52 3a 20 32 33 0d             | E ERROR: 23
>        Received message type 67
>        Write failed (Command failed.): memory type: ME, loc: 1, name: test1, 
> number: 1

Well, we didn't get here to the write command. This is just checking
memory status. Perhaps it fails with the empty memory only? I guess I
could add some workaround for that.

>   There are two workaround for this - either creating a single contact
>   in phone memory using the phone, or adding a dummy entry as the first
>   one in phonebook.

>From gnokii perspective I can do two things:
 - ignore the error
 - catch the error -> write a dummy entry -> read memory status ->
delete a dummy entry

>   Is there already a driver flag not to get confused by getting no
>   reasonable response on AT+CPBS="ME"?

No, I do not think so. That's probably the first phone that suffers
this problem.

> 2) getphonebook stopped working after I wiped phone memory and repopulated it 
> using
>   writephonebook. Now the phone responds with
>        \05\00\03d+CPBR=1   +CME ERROR: 100
>   when attempting to read non-empty positions.

With any non empty? Does phone display them correctly?

>> Quite frankly I'd prefer to have the flags:
>>  - encode_number_write
>>  - encode_number_read
>>  - encode_name_write
>>  - encode_name_read
>>
>> Do you mind updating the patch like that?
>
> No problem, but it will probably take some days to finish as the
> getphonebook issue seems more important now (and also separating
> encode_number will need some more changes).

No problem.

>> > With this patch, both readphoonebook and writephonebook work with
>> > Samsung B2100, although writephonebook's -o (overwrite) option seems
>> > to have no effect, the phone probably uses some internal find-free
>> > function
>>
>> Which memory type? It if was SM the fix was added just recently.
>
> Memory type is ME. Observed behavior is that repeated writephonebook calls
> with the same entry create repeated identical entries in the phone.

I see. I wonder why did Samsung so screw up with your phone.

take care,
-- 
Pawel Kot