Аноним | вход | зарегистрировать новую учетную запись | 2021-01-16 13:56 EET |
главная | моя сводка | список вопросов | журнал изменений | план развития | документация | моя учетная запись |
Простой вид [ комментарии ] | [ расширенный вид ] [ история ] [ печать ] | ||||||
Номер | Категория | Серьезность | Воспроизводимость | Создан | Изменен | ||
0000129 | [obex-data-server] General | нововведение | иногда | 2008-09-06 08:45 | 2008-10-25 22:09 | ||
Инициатор | Jo | Видимость | общая | ||||
Ответственный | skirsdeda | ||||||
Приоритет | срочный | Решение | решен | ||||
Состояние | закрыт | Версия продукта | 0.3.4 | ||||
Суть | 0000129: Application sometimes misses consecutive ODS signals. | ||||||
Подробности |
1. As an ODS client (session) ODS emits the "Connected" signal right after the "SessionCreated" signal. The application sometimes doesn't catch the "Connected" signal and hangs especially when the bonding procedure is performed just before. 2. As an ODS server (server-session) ODS emits the "TransferStarted" signal right after the "Connected" signal and the application misses the "TransferStarted" signal and hangs too. I run the ODS and bluez on a mobile phone platform. In a normal situation when the pairing is done previously, the OPP operates well. But when there are heavy IO overheads, e.g., just after the bonding procedure, the application sometimes misses a signal from ODS. I think if the ODS emits the second signal after getting a feedback from the application about the first signal this problem won't happen. The simple g_signal_emit functions seems not to assure the reception of the signals. In a tentative measure, I make the ODS sleep for a short period of time before sending the second signal and this works. Do you have a better suggestion or plan to handle this problem? |
||||||
Дополнительные сведения | |||||||
Tэги | Нет прикрепленных тэгов. | ||||||
Вложенные файлы | |||||||
|
![]() |
||||||
|
![]() |
|
(0000410) skirsdeda (администратор) 2008-09-12 19:47 |
The first scenario was fixed by changing API and emitting SessionConnected in Manager object. This is available in SVN and will be included in 0.4 release. Second scenario wasn't previously reported so we need to figure something out. |
(0000416) Jo (инициатор) 2008-09-19 09:10 |
When we try to send a file to a non-paired device via OPP, we first perform bonding procedure and start the transmission right after that. If the mobile device does lots of I/O related jobs, the device sometimes misses the TransferStarted signal. This doesn't happen if the pairing procedure has already performed before the OPP transfer trial has made. I'm really anticipating the ODS 0.3.5 or 0.4 release :) Thanks a lot! |
(0000417) heston_james (инициатор) 2008-09-20 13:22 |
Jo, Just out of interest, which mobile platform are you running this on? Cheers, Heston |
(0000425) Jo (инициатор) 2008-09-29 05:48 изменен: 2008-09-29 06:01 |
Sorry for the delayed reply. I'm running bluez + ODS on a linux(2.6.26) based mobile platform, one of LiMo candidate. It is not perfectly stabilized yet but running well :) I'm sorry but I can't say about our current project status in more detail. |
(0000426) heston_james (инициатор) 2008-09-29 11:58 |
Hey Jo, No problem about the slow reply :-) Sounds interesting, I can appreciate that you've got to keep certain things quiet. There seems to be a large amount of work on mobile based linux at the moment, its all quite exciting, it'll be interesting to see where it all goes :-) Cheers, Heston |
(0000490) skirsdeda (администратор) 2008-10-23 23:25 |
A fix is now in svn. See 0000140 for more info. |
Mantis 1.1.6[^] Copyright © 2000 - 2008 Mantis Group |