Hi Kristen,

On 11/24/2010 04:53 PM, Kristen Carlson Accardi wrote:
> ---
>  src/sms.c     |    9 +++++++++
>  src/smsutil.c |   38 ++++++++++++++++++++++++++++++++++++++
>  src/smsutil.h |    4 ++++
>  3 files changed, 51 insertions(+), 0 deletions(-)
> 
> diff --git a/src/sms.c b/src/sms.c
> index 6eab0fa..f987946 100644
> --- a/src/sms.c
> +++ b/src/sms.c
> @@ -769,6 +769,10 @@ static void tx_finished(const struct ofono_error *error, 
> int mr, void *data)
>               goto next_q;
>       }
>  
> +     sms_tx_backup_remove(sms->imsi, entry->id,
> +                             ofono_uuid_to_str(&entry->uuid),
> +                             entry->cur_pdu);
> +
>       entry->cur_pdu += 1;
>       entry->retry = 0;
>  
> @@ -813,8 +817,13 @@ next_q:
>               message_set_state(sms, &entry->uuid, ms);
>       }
>  
> +     sms_tx_backup_destroy(sms->imsi, entry->id,
> +                     ofono_uuid_to_str(&entry->uuid));
> +
>       tx_queue_entry_destroy(entry);
>  
> +     tx_counter--;
> +

I'm not sure this is going to work, e.g. start with tx_counter == 0:

Queue -> tx_counter = 1, id = 0
Queue -> tx_counter = 2, id = 1,
Queue -> tx_counter = 3, id = 2,

Sent -> tx_counter = 2, (id of 0 sent)

Queue -> tx_counter = 3, id = 2

Now you have two messages with the same id and lost all guarantees on
the order...

Perhaps simply relying on always increasing tx_counter might be enough.
 It is unlikely anyone will ever send 4 billion SMSes ;)  Renaming the
directory during the restore operation might be an option as well.

>       if (g_queue_peek_head(sms->txq)) {
>               DBG("Scheduling next");
>               sms->tx_source = g_timeout_add(0, tx_next, sms);
> diff --git a/src/smsutil.c b/src/smsutil.c
> index 4904419..7a6b70a 100644
> --- a/src/smsutil.c
> +++ b/src/smsutil.c
> @@ -3151,6 +3151,44 @@ gboolean sms_tx_store(const char *imsi, unsigned long 
> id, const char *uuid,
>       return TRUE;
>  }
>  
> +void sms_tx_backup_destroy(const char *imsi, unsigned long id, const char 
> *uuid)

The existing code tends to name this foo_backup_free rather than
backup_destroy.

> +{
> +     char *path;
> +     struct dirent **entries;
> +     int len;
> +
> +     path = g_strdup_printf(SMS_TX_BACKUP_PATH_DIR,
> +                                     imsi, id, uuid);
> +
> +     len = scandir(path, &entries, NULL, versionsort);

Rule M1 please

> +     while (len--) {
> +             struct dirent *dir = entries[len];
> +             char *file = g_strdup_printf("%s/%s", path, dir->d_name);
> +
> +             unlink(file);
> +
> +             g_free(file);
> +             g_free(entries[len]);
> +     }

Rule M1 please

> +     g_free(entries);
> +
> +     rmdir(path);
> +
> +     g_free(path);
> +}
> +
> +void sms_tx_backup_remove(const char *imsi, unsigned long id, const char 
> *uuid,
> +                             guint8 seq)
> +{
> +     char *path;
> +
> +     path = g_strdup_printf(SMS_TX_BACKUP_PATH_FILE,
> +                                     imsi, id, uuid, seq);
> +     unlink(path);
> +
> +     g_free(path);
> +}
> +
>  static inline GSList *sms_list_append(GSList *l, const struct sms *in)
>  {
>       struct sms *sms;
> diff --git a/src/smsutil.h b/src/smsutil.h
> index 5279489..2e2bae6 100644
> --- a/src/smsutil.h
> +++ b/src/smsutil.h
> @@ -518,6 +518,10 @@ void status_report_assembly_expire(struct 
> status_report_assembly *assembly,
>  
>  gboolean sms_tx_store(const char *imsi, unsigned long id, const char *uuid,
>                               struct sms *s, guint8 seq);
> +void sms_tx_backup_remove(const char *imsi, unsigned long id, const char 
> *uuid,
> +                             guint8 seq);
> +void sms_tx_backup_destroy(const char *imsi, unsigned long id,
> +                             const char *uuid);
>  
>  GSList *sms_text_prepare(const char *to, const char *utf8, guint16 ref,
>                               gboolean use_16bit,

Regards,
-Denis
_______________________________________________
ofono mailing list
[email protected]
http://lists.ofono.org/listinfo/ofono

Reply via email to