| Message ID | 20250224111056.3255513-1-quic_amisjain@quicinc.com |
|---|---|
| State | New |
| Headers | show |
| Series | [RESEND,v2] obex: Add messages_get_message() implementation for MAP plugin | expand |
Hi Paul, > -----Original Message----- > From: Paul Menzel <pmenzel@molgen.mpg.de> > Sent: Monday, February 24, 2025 7:34 PM > To: Amisha Jain (QUIC) <quic_amisjain@quicinc.com> > Cc: linux-bluetooth@vger.kernel.org; Mohammed Sameer Mulla (QUIC) > <quic_mohamull@quicinc.com>; Harish Bandi (QUIC) > <quic_hbandi@quicinc.com>; Anubhav Gupta (QUIC) > <quic_anubhavg@quicinc.com> > Subject: Re: [RESEND PATCH v2] obex: Add messages_get_message() > implementation for MAP plugin > > Dear Amisha, > > > Am 24.02.25 um 12:10 schrieb Amisha Jain: > > GET Message() operation should be supported for passing below PTS > > testcases - > > > > 1.MAP/MSE/MMB/BV-12-C > > Verify that the MSE can return an email message to the MCE. > > 2.MAP/MSE/MMB/BV-13-C > > Verify that the MSE can return a SMS message in native format to the MCE. > > a*n* SMS > > > 3.MAP/MSE/MMB/BV-14-C > > Verify that the MSE can return a SMS message with text trans-coded to > > UTF-8 > > a*n* > > > to the MCE. > > I’d add a space after the bullet points (the dot). > > > Currently get message operation is not implemented, hence above > > testcases are failing. > > Added code to send the complete bmessage in response > > Should this be *message* or is *bmessage* some terminology? > This is 'bmessage', message format used by MAP to store messages. As per Spec, bMessages are application objects used by MAP for message transport. Exchanged messages shall use the bMessage format. > > to the get request for the requested message handle. > > > > --- > > obexd/plugins/mas.c | 4 ++-- > > obexd/plugins/messages-dummy.c | 27 ++++++++++++++++++++++++++- > > 2 files changed, 28 insertions(+), 3 deletions(-) > > > > diff --git a/obexd/plugins/mas.c b/obexd/plugins/mas.c index > > 10b972d65..f63fcf6c6 100644 > > --- a/obexd/plugins/mas.c > > +++ b/obexd/plugins/mas.c > > @@ -612,11 +612,11 @@ static void *message_open(const char *name, int > oflag, mode_t mode, > > return NULL; > > } > > > > + mas->buffer = g_string_new(""); > > + > > *err = messages_get_message(mas->backend_data, name, 0, > > get_message_cb, mas); > > > > - mas->buffer = g_string_new(""); > > - > > if (*err < 0) > > return NULL; > > else > > diff --git a/obexd/plugins/messages-dummy.c > > b/obexd/plugins/messages-dummy.c index e313c6163..665face3f 100644 > > --- a/obexd/plugins/messages-dummy.c > > +++ b/obexd/plugins/messages-dummy.c > > @@ -516,7 +516,32 @@ int messages_get_message(void *session, const > char *handle, > > messages_get_message_cb callback, > > void *user_data) > > { > > - return -ENOSYS; > > + struct session *s = session; > > + FILE *fp; > > + char *path; > > + char buffer[1024]; > > + > > + DBG(" "); > > + path = g_build_filename(s->cwd_absolute, handle, NULL); > > + fp = fopen(path, "r"); > > + if (fp == NULL) { > > + DBG("fopen() failed"); > > + return -EBADR; > > + } > > + > > + /* 1024 is the maximum size of the line which is calculated to be more > > + * sufficient*/ > > I do not fully grok this sentence. Could you please rephrase? > Sure, this corresponds to max size of a line in a file, as we are reading the file line by line. Same way it is present in existing function - get_messages_listing(). static gboolean get_messages_listing(void *d) { struct message_listing_data *mld = d; /* 1024 is the maximum size of the line which is calculated to be more * sufficient*/ char buffer[1024]; GMarkupParseContext *ctxt; > > + while (fgets(buffer, 1024, fp)) { > > + if (callback) > > + callback(session, 0, 0, (const char*)buffer, user_data); > > + } > > + > > + if (callback) > > + callback(session, 0, 0, NULL, user_data); > > + > > + g_free(path); > > + fclose(fp); > > + return 0; > > } > > > > int messages_update_inbox(void *session, messages_status_cb > > callback, > > > Kind regards, > > Paul Thanks, Amisha
Hi Amisha, On Wed, Feb 26, 2025 at 2:11 AM Amisha Jain (QUIC) <quic_amisjain@quicinc.com> wrote: > > Hi Paul, > > > -----Original Message----- > > From: Paul Menzel <pmenzel@molgen.mpg.de> > > Sent: Monday, February 24, 2025 7:34 PM > > To: Amisha Jain (QUIC) <quic_amisjain@quicinc.com> > > Cc: linux-bluetooth@vger.kernel.org; Mohammed Sameer Mulla (QUIC) > > <quic_mohamull@quicinc.com>; Harish Bandi (QUIC) > > <quic_hbandi@quicinc.com>; Anubhav Gupta (QUIC) > > <quic_anubhavg@quicinc.com> > > Subject: Re: [RESEND PATCH v2] obex: Add messages_get_message() > > implementation for MAP plugin > > > > Dear Amisha, > > > > > > Am 24.02.25 um 12:10 schrieb Amisha Jain: > > > GET Message() operation should be supported for passing below PTS > > > testcases - > > > > > > 1.MAP/MSE/MMB/BV-12-C > > > Verify that the MSE can return an email message to the MCE. > > > 2.MAP/MSE/MMB/BV-13-C > > > Verify that the MSE can return a SMS message in native format to the MCE. > > > > a*n* SMS > > > > > 3.MAP/MSE/MMB/BV-14-C > > > Verify that the MSE can return a SMS message with text trans-coded to > > > UTF-8 > > > > a*n* > > > > > to the MCE. > > > > I’d add a space after the bullet points (the dot). > > > > > Currently get message operation is not implemented, hence above > > > testcases are failing. > > > Added code to send the complete bmessage in response > > > > Should this be *message* or is *bmessage* some terminology? > > > > This is 'bmessage', message format used by MAP to store messages. > As per Spec, bMessages are application objects used by MAP for message transport. > Exchanged messages shall use the bMessage format. > > > > to the get request for the requested message handle. > > > > > > --- > > > obexd/plugins/mas.c | 4 ++-- > > > obexd/plugins/messages-dummy.c | 27 ++++++++++++++++++++++++++- > > > 2 files changed, 28 insertions(+), 3 deletions(-) > > > > > > diff --git a/obexd/plugins/mas.c b/obexd/plugins/mas.c index > > > 10b972d65..f63fcf6c6 100644 > > > --- a/obexd/plugins/mas.c > > > +++ b/obexd/plugins/mas.c > > > @@ -612,11 +612,11 @@ static void *message_open(const char *name, int > > oflag, mode_t mode, > > > return NULL; > > > } > > > > > > + mas->buffer = g_string_new(""); > > > + > > > *err = messages_get_message(mas->backend_data, name, 0, > > > get_message_cb, mas); > > > > > > - mas->buffer = g_string_new(""); > > > - > > > if (*err < 0) > > > return NULL; > > > else > > > diff --git a/obexd/plugins/messages-dummy.c > > > b/obexd/plugins/messages-dummy.c index e313c6163..665face3f 100644 > > > --- a/obexd/plugins/messages-dummy.c > > > +++ b/obexd/plugins/messages-dummy.c > > > @@ -516,7 +516,32 @@ int messages_get_message(void *session, const > > char *handle, > > > messages_get_message_cb callback, > > > void *user_data) > > > { > > > - return -ENOSYS; > > > + struct session *s = session; > > > + FILE *fp; > > > + char *path; > > > + char buffer[1024]; > > > + > > > + DBG(" "); > > > + path = g_build_filename(s->cwd_absolute, handle, NULL); > > > + fp = fopen(path, "r"); > > > + if (fp == NULL) { > > > + DBG("fopen() failed"); > > > + return -EBADR; > > > + } > > > + > > > + /* 1024 is the maximum size of the line which is calculated to be more > > > + * sufficient*/ > > > > I do not fully grok this sentence. Could you please rephrase? > > > > Sure, this corresponds to max size of a line in a file, as we are reading the file line by line. > Same way it is present in existing function - get_messages_listing(). > > static gboolean get_messages_listing(void *d) > { > > struct message_listing_data *mld = d; > /* 1024 is the maximum size of the line which is calculated to be more > * sufficient*/ > char buffer[1024]; > GMarkupParseContext *ctxt; > > > > + while (fgets(buffer, 1024, fp)) { > > > + if (callback) > > > + callback(session, 0, 0, (const char*)buffer, user_data); > > > + } > > > + > > > + if (callback) > > > + callback(session, 0, 0, NULL, user_data); Btw, if the callback is NULL then perhaps we shouldn't be looping with fgets, I also thinking that perhaps the whole think about reading the messages and copying it over to a buffer perhaps could be avoided, perhaps using mmap is better here, although this is just the dummy implementation with perhaps doesn't care so much about doing things in a 'proper' way. > > > + > > > + g_free(path); > > > + fclose(fp); > > > + return 0; > > > } > > > > > > int messages_update_inbox(void *session, messages_status_cb > > > callback, > > > > > > Kind regards, > > > > Paul > > Thanks, > Amisha
diff --git a/obexd/plugins/mas.c b/obexd/plugins/mas.c index 10b972d65..f63fcf6c6 100644 --- a/obexd/plugins/mas.c +++ b/obexd/plugins/mas.c @@ -612,11 +612,11 @@ static void *message_open(const char *name, int oflag, mode_t mode, return NULL; } + mas->buffer = g_string_new(""); + *err = messages_get_message(mas->backend_data, name, 0, get_message_cb, mas); - mas->buffer = g_string_new(""); - if (*err < 0) return NULL; else diff --git a/obexd/plugins/messages-dummy.c b/obexd/plugins/messages-dummy.c index e313c6163..665face3f 100644 --- a/obexd/plugins/messages-dummy.c +++ b/obexd/plugins/messages-dummy.c @@ -516,7 +516,32 @@ int messages_get_message(void *session, const char *handle, messages_get_message_cb callback, void *user_data) { - return -ENOSYS; + struct session *s = session; + FILE *fp; + char *path; + char buffer[1024]; + + DBG(" "); + path = g_build_filename(s->cwd_absolute, handle, NULL); + fp = fopen(path, "r"); + if (fp == NULL) { + DBG("fopen() failed"); + return -EBADR; + } + + /* 1024 is the maximum size of the line which is calculated to be more + * sufficient*/ + while (fgets(buffer, 1024, fp)) { + if (callback) + callback(session, 0, 0, (const char*)buffer, user_data); + } + + if (callback) + callback(session, 0, 0, NULL, user_data); + + g_free(path); + fclose(fp); + return 0; } int messages_update_inbox(void *session, messages_status_cb callback,