Sachin Agrawal [Wed, 24 Sep 2014 21:22:47 +0000 (14:22 -0700)]
IOT-36 : Fixing memory leaks
Change-Id: I8ed2d398d79c43abf4d0131232c88f9ae8890f8f
Erich Keane [Wed, 24 Sep 2014 16:20:14 +0000 (09:20 -0700)]
Merge "Add const modifier"
Caiwen Zhang [Wed, 24 Sep 2014 20:11:26 +0000 (04:11 +0800)]
Add const modifier
const modifier is required for new avr-gcc compiler
Change-Id: I1a7b4b4059632e0c6ea83a65d7514785575416cf
Sashi Penta [Tue, 23 Sep 2014 23:38:36 +0000 (16:38 -0700)]
use of const reference where possible.
Change-Id: I4f47fdc385f58f2e52ce10bffa68ae0910e496c9
Erich Keane [Tue, 23 Sep 2014 22:49:15 +0000 (15:49 -0700)]
Merge "1. arrays in OCRepresentation 2. Fixed review comments from Jesse. 3. Refactored the code to get JSON from vector."
Erich Keane [Tue, 23 Sep 2014 22:49:04 +0000 (15:49 -0700)]
Merge "JSON Object implementation"
Sashi Penta [Tue, 23 Sep 2014 07:06:07 +0000 (00:06 -0700)]
1. arrays in OCRepresentation
2. Fixed review comments from Jesse.
3. Refactored the code to get JSON from vector.
Change-Id: I1f31f8f0c242f53f56cf50f9d9e781d8ed770f8c
Sashi Penta [Fri, 19 Sep 2014 07:06:05 +0000 (00:06 -0700)]
JSON Object implementation
1. Support for JSON object in OCRepresentation.
2. Code refactor for JSON composition in various parts of the stack into one place (OCRepresenation).
3. Similar code refactoring for JSON parsing needs to be done. I will revisit when we look into
JSON parse library alternatives.
4. Implemented correct JSON composition. Earlier we were returning strings for all the json types.
Now strings are only used for string types and JSON objects.
5. garageserver.cpp and garageclient.cpp examples to explain the user of JSON object with in
the representation.
Change-Id: I9461db811c7b5e3b827d5704cde489aa2005e291
Erich Keane [Mon, 22 Sep 2014 18:16:33 +0000 (11:16 -0700)]
Non-discoverable resources cause bad JSON (trailing comma). By putting the comma generation first, we buy ourselves some room
Change-Id: Ie270c22e0f61c657f15ecf45fcbdb96de72d495f
Yamin S Al-mousa [Fri, 19 Sep 2014 17:56:47 +0000 (10:56 -0700)]
A fix for a bug reported by Omkar and Vijay:
The check for required uri lenght should consider that it includes the uri and query.
Change-Id: I1ca3e4709de02cb4f5217aba5133b8aa2bc4ee14
Sashi Penta [Fri, 19 Sep 2014 09:15:50 +0000 (02:15 -0700)]
Makefile in ocicuc is not cleaning up all the object files.
Change-Id: Ia1d198b96b61a82c07c069b4545ad40460cf2f4c
Erich Keane [Thu, 18 Sep 2014 17:32:45 +0000 (10:32 -0700)]
Merge "Revert "Replace to_string and stoi""
Erich Keane [Thu, 18 Sep 2014 17:32:32 +0000 (10:32 -0700)]
Revert "Replace to_string and stoi"
We have decided that sprintf is not stable/future proof enough, so we are going to revert this change (and the other one) and investigate what it would take to make android-only implementations of tostring and stoi
This reverts commit
dcfe2eb3bee0296aee3cc4453f4237c784251671.
Change-Id: I82bb37e42aa5b6809bdf6a6522dac678ba74aac0
Caiwen Zhang [Thu, 18 Sep 2014 17:56:05 +0000 (01:56 +0800)]
Replace to_string and stoi
Part of C++11 features aren't supported by Android NDK, includes
to_string and stoi. Replace them with snprintf and atoi.
Change-Id: I6edda1856ab34c0e1dcc23daef0b105875a09189
Erich Keane [Thu, 18 Sep 2014 16:49:57 +0000 (09:49 -0700)]
Merge "Fixing Arduino compilation error"
Erich Keane [Thu, 18 Sep 2014 16:49:22 +0000 (09:49 -0700)]
Merge "Avoid redefine 'NDEBUG'"
Caiwen Zhang [Thu, 18 Sep 2014 17:18:50 +0000 (01:18 +0800)]
Avoid redefine 'NDEBUG'
Change-Id: I1a8e1c5b0a8b7aa0c1ab19e292485e4e20ab90a5
Erich Keane [Thu, 18 Sep 2014 16:48:45 +0000 (09:48 -0700)]
Merge "Fixed warning (control reaching end of non void function) in OCPlatform.cpp"
Sachin Agrawal [Thu, 18 Sep 2014 15:39:57 +0000 (08:39 -0700)]
Fixing Arduino compilation error
Change-Id: I030f3d66b54ab61bd866f434e3cadfee30ff4ac7
Sudarshan Prasad [Thu, 18 Sep 2014 02:18:38 +0000 (19:18 -0700)]
Fixed warning (control reaching end of non void function) in OCPlatform.cpp
Change-Id: Ibab82216537c0927794bfb57dbe9e97f52b4d028
Erich Keane [Wed, 17 Sep 2014 21:07:36 +0000 (14:07 -0700)]
Merge "Support for notifying specific observers using observation Id"
Erich Keane [Wed, 17 Sep 2014 21:01:16 +0000 (14:01 -0700)]
Added release notes for Sept 17th, 2014 release.
Change-Id: I3bf64deb84e54b2c9306c03688ff5d8be9fe4053
Vijay [Wed, 17 Sep 2014 20:48:56 +0000 (13:48 -0700)]
Support for notifying specific observers using observation Id
1. Initial code changes
2. Combine GET & observe registration calls to entity handler into a single call; support for CON messages in NotifySpecificObservers; updates sample server app
3. Added the support for immediate observe cancelation in OCCancel
4. Fixed the new interface to OCCancel in C++ layer
5. A temp solution for the case when observation sequence number wraps around.
6. Merging with Yamin's changes
7. Changed OCNotifyObservers to OCNotifyAllObservers
8. Build issue
9. Another Merge to support calling the entity handler when deregister an observer
10. Rebase & CPP Changes from Sudarshan and Sashi to support Notify specific observers.
11. Command line argument options for simpleserver (so it can choose between all observers or specific observers)
12. CPP bugfix.
13. Fixed review comments for CPP.
14. Changed COAP_DEFAULT_RESPONSE_TIMEOUT to 2 and all the sample applications to sleep accordingly.
15. Fixed review comment for new line.
16. Fixed review commend for line exceed.
17. Response to comments received on patch 11
18. Header file comments updated
Change-Id: I03aa7ab2a580899441b70bd2a76cca05575b192a
Erich Keane [Tue, 16 Sep 2014 21:31:56 +0000 (14:31 -0700)]
Removed <oic-resource>/csdk/android and <oic-resoruce>/csdk/controller plus <oic-resource>/csdk/libcoap-4.1.1/config.log and config.status
PatchSet 2: Erich Rebased for submission sake
Change-Id: I0d5a355924aca94847aa70f8adcb10c700164cdc
Signed-off-by: Charunethran PG <charunethran.panchalam.govindarajan@intel.com>
Erich Keane [Tue, 16 Sep 2014 20:58:10 +0000 (13:58 -0700)]
Merge "Partial Implementation of US1574:"
Sashi Penta [Wed, 10 Sep 2014 22:19:16 +0000 (15:19 -0700)]
Partial Implementation of US1574:
OCRepresentation is now provides simpler interface to setValue and getValue from the applications.
Applications doesn't have to go through attributeMaps in OCRepresentation.
OCRepresentation now supports serialization/deserialization for primitive attributes (numbers, booleans and strings) so application doesn't have to deal with strings.
Change-Id: I09d323fc624d87c0b2a3c9b510935a6d5907a11b
Erich Keane [Tue, 16 Sep 2014 18:02:38 +0000 (11:02 -0700)]
Merge "Removed extraneous comment. The callback process is triggered from a lower point in the stack than from here. The consideration written in comment here has already been taken care of, as per Yamin."
Erich Keane [Tue, 16 Sep 2014 18:02:15 +0000 (11:02 -0700)]
Merge "IOT-48 : Updated makefiles for compiling OCStack with WiFi libraries for Arduino 1.0.5 Patch 2 : Included suggestions provided in Review comments."
Erich Keane [Tue, 16 Sep 2014 17:57:42 +0000 (10:57 -0700)]
Merge "Fixing a few warnings that we see coming out of the libcoap bits.h file."
Joseph Morrow [Tue, 16 Sep 2014 17:56:57 +0000 (13:56 -0400)]
Removed extraneous comment. The callback process is triggered from a
lower point in the stack than from here. The consideration written in
comment here has already been taken care of, as per Yamin.
Change-Id: I982922f8c87b42385e358cfe91b24bfee5ffc17f
Caiwen Zhang [Tue, 16 Sep 2014 18:16:41 +0000 (02:16 +0800)]
Remove <ifaddrs.h>
This header file is unnecessary. It will break android build as
android ndk doesn't have this file.
Change-Id: Iacd08868d1e426fe73d6b4c31d0c67ce38df9e69
Erich Keane [Tue, 16 Sep 2014 17:50:13 +0000 (10:50 -0700)]
Removing references to OCReflect that were left behind but didnt cause build issues
Change-Id: I5480fa14d5ea4960b361d62a67563f45d1dbbfed
Sachin Agrawal [Tue, 16 Sep 2014 17:42:24 +0000 (10:42 -0700)]
IOT-48 : Updated makefiles for compiling OCStack with WiFi libraries for Arduino 1.0.5
Patch 2 : Included suggestions provided in Review comments.
Change-Id: I70190ce85efcaf3b560687bff1ccace407778e44
Erich Keane [Mon, 15 Sep 2014 21:15:43 +0000 (14:15 -0700)]
Fixing a few warnings that we see coming out of the libcoap bits.h file.
Essentially, the result of the bit-shift on a uint8_t is always upconverted to a signed int despite it being impossible for it to be anything greater than 5 bits long (8 bits, shifted 3 more). We cast to unsigned size_t to prevent the warning and to prevent unsigned vs signed comparison.
Change-Id: I47d05611e75e55240b55b800c9370dbc019aedd2
Joseph Morrow [Mon, 15 Sep 2014 21:02:17 +0000 (17:02 -0400)]
Patch 1: Sending RST in response to the receipt of a notification from a server
that is not a presence nor an observe notifcation AND there is not
a callback associated.
Patch 2: Rebase
Patch 3: Changed debug print out message.
Patch 4: Oops. I had too much message in my log message.
Change-Id: If8f949827e9accf0dc45b69a5568960578b388f5
Erich Keane [Mon, 15 Sep 2014 17:50:33 +0000 (10:50 -0700)]
Merge "Patch 1: Arduino pdu malloc fix."
Erich Keane [Mon, 15 Sep 2014 15:33:19 +0000 (08:33 -0700)]
Merge "Updated Arduino sample app to reflect the current SRAM available. \n This will help in debugging Arduino crash issues in low SRAM scenarios Patch 2 : Updated code to execute only for MEGA2560 boards."
Joseph Morrow [Fri, 12 Sep 2014 17:17:46 +0000 (13:17 -0400)]
Patch 1: Arduino pdu malloc fix.
Patch 2: Rebase
Patch 3: Changed implementation of fix to simply change size of
COAP_MAX_PDU_SIZE to 320 for Arduino only. This will stay the same
going forth, but there is an uncaught case that can only be fixed
in the next sprint(as there is not enough time)-- We must implement
blockwise transfer so that a PDU larger than 320 is properly
handled (i.e. broken up into chunks of 320 or smaller). A Rally task
has already been created for Sprint 8 for this.
Patch 4: Meant to use "else" instead of "elif".
Change-Id: I4df20d28c821f6fc2c98a6bf3672331264978be5
Sachin Agrawal [Fri, 12 Sep 2014 16:08:25 +0000 (09:08 -0700)]
Updated Arduino sample app to reflect the current SRAM available. \n This will help in debugging Arduino crash issues in low SRAM scenarios
Patch 2 : Updated code to execute only for MEGA2560 boards.
Change-Id: If130815b1f8eedd9373e68e7e144ef8713a2f2cb
Erich Keane [Thu, 11 Sep 2014 23:56:30 +0000 (16:56 -0700)]
Merge "Revert "Doxygen run for *.h files""
Erich Keane [Thu, 11 Sep 2014 23:56:19 +0000 (16:56 -0700)]
Revert "Doxygen run for *.h files"
We should never have committed generated files. When necessary, Shamit can generate these himself.
This reverts commit
fdaa9cd9e05732b5087f11dcb40c2c66cc3a73e6.
Change-Id: Ibeb9f036ddf421d0d3e0f3d509ab6d2d92477db4
Thiago Macieira [Fri, 5 Sep 2014 01:21:00 +0000 (18:21 -0700)]
Remove unused variable
It's not used anywhere
Change-Id: I834b7e8f561e85dea45846d2bfe96ad90fed6ab8
Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>
Thiago Macieira [Fri, 5 Sep 2014 01:20:13 +0000 (18:20 -0700)]
Fix mistaken use of && where & was intended
Found by Clang:
occoaphelper.c:93:59: warning: use of logical '&&' with constant operand [-Wconstant-logical-operand]
It's clearly meant to be a &, as seen by the right shift by 5 and the
constant of 31 (5 bits)
Change-Id: Ie35ad6c313ad53d766fc6fe88ffd03b462d49519
Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>
Erich Keane [Thu, 11 Sep 2014 17:57:44 +0000 (10:57 -0700)]
Fix use of uninitialized variable
If ctx == NULL, then VERIFY_NON_NULL would goto exit, which would then
OCFree the uninitialized variables.
Found by Clang:
occoap.c:274:5: warning: variable 'rcvObserveOption' is used uninitialized whenever 'if' condition is true [-Wsometimes-uninitialized]
occoap.c:44:36: note: expanded from macro 'VERIFY_NON_NULL'
#define VERIFY_NON_NULL(arg) { if (!arg) {OC_LOG_V(FATAL, TAG, "%s is NULL", #arg); goto exit;} }
^~~~
occoap.c:459:16: note: uninitialized use occurs here
--Erich: rebase on master
Change-Id: Ie829c6ac284dfd988eeeffad446d9872424e604a
Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>
Thiago Macieira [Fri, 5 Sep 2014 01:00:53 +0000 (18:00 -0700)]
Fix warnings about diff and base possibly not initialized
This is a bogus warning by GCC, but it's easy to fix by having an "else"
case with no extra conditions.
Change-Id: I0f99527065138ff96c0613a57b0462613139900c
Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>
Thiago Macieira [Fri, 5 Sep 2014 00:52:18 +0000 (17:52 -0700)]
Rewrite examples/ocicuc/Makefile
This Makefile was almost completely broken, as the comment at the top
indicated. This rewrite makes it work and also respect the top-level set
CXX variable.
Change-Id: I462b5caedfc090e6ebff72c869e82e3d57dc16c2
Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>
Thiago Macieira [Fri, 5 Sep 2014 00:27:00 +0000 (17:27 -0700)]
Fix parallel build
All the .o files must be compiled before we can create the .a library
from them. When compiling with a -j option (as low as -j4), often one or
two .o files would be missing from the resulting .a because they hadn't
finished compiling yet.
Change-Id: Ie7bc0ad489b8b0afecab3ec318f2784ee706ae12
Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>
Thiago Macieira [Fri, 5 Sep 2014 00:09:53 +0000 (17:09 -0700)]
Use CXX for the variable containing the name of the C++ compiler
CXX is standard, CCPLUS is not used anywhere else.
Change-Id: I2b98d609ade694306dd20062b8149a242ff7e91b
Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>
Thiago Macieira [Fri, 5 Sep 2014 00:05:52 +0000 (17:05 -0700)]
Building examples requires OCLib.a to be built first
Change-Id: I2864d8c2adc7373f6a7c6ed5619f671af26c5727
Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>
Thiago Macieira [Thu, 4 Sep 2014 23:58:47 +0000 (16:58 -0700)]
Update the #pragma GCC diagnostic usage
Don't force pedantic back on. Instead, push and pop the current state.
Don't pass these options to Clang and ICC, since they don't have the
-pedantic option anyway.
Change-Id: I302f7fcfe974f23423ca2ac7711a7c4a8ac39d88
Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>
Thiago Macieira [Thu, 4 Sep 2014 23:57:56 +0000 (16:57 -0700)]
Don't pass linker flags to compilation stages
Change-Id: Ie6d45dee1dd1919fb32a05c30143c4ba6ad085d8
Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>
Thiago Macieira [Thu, 4 Sep 2014 23:40:39 +0000 (16:40 -0700)]
Add -p option to mkdir commands in the Makefiles
This makes mkdir not complain if it cannot create a directory that
already exists. The result of the command is ignored, but we don't need
the diagnostic either.
Change-Id: I25b9da5246e1469bccd5827886eecadd62ad319f
Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>
Erich Keane [Thu, 11 Sep 2014 17:45:09 +0000 (10:45 -0700)]
Merge "Doxygen run for *.h files"
Erich Keane [Thu, 11 Sep 2014 17:44:23 +0000 (10:44 -0700)]
Merge "Add .gitignore file"
Erich Keane [Thu, 11 Sep 2014 16:50:10 +0000 (09:50 -0700)]
Merge "Added check for Time libraries in expected location for the Arduino Time Libraries from the TBStack makefile. If they are not there, the user will be told to put them where we expect them to be now."
Thiago Macieira [Thu, 4 Sep 2014 23:45:22 +0000 (16:45 -0700)]
Add .gitignore file
Change-Id: Ibb7a2b474bc4eb8a46654e2925fd73fddac0545a
Signed-off-by: Thiago Macieira <thiago.macieira@intel.com>
Joseph Morrow [Wed, 10 Sep 2014 21:36:18 +0000 (17:36 -0400)]
Added check for Time libraries in expected location for the Arduino
Time Libraries from the TBStack makefile. If they are not there,
the user will be told to put them where we expect them to be now.
Change-Id: I93693a4623d038e152559ff9d8260bdb91c9000f
Erich Keane [Wed, 10 Sep 2014 18:08:09 +0000 (11:08 -0700)]
Provided a mechanism for deleting the context pointer when the C Stack deletes callback. Otherwise, the callback context will be leaked.
Fixed some line endings.
Fixed some line lengths.
Fixed a series of memory leaks in C++ callbacks which assumed the C stack owned/deleted OCCallbackData
Patch 2: Removed deletion of function pointers becaues GCC says this violates the C spec.
Patch 3: Comments from Joey regarding the comment on the OCClientContextDeleter typedef
Patch 4: changed deleter to deleteCallback per sashi's suggestion
Change-Id: I0fb71896afac93356c1bd2f4916eb9f3eaa0ad13
Erich Keane [Wed, 10 Sep 2014 19:36:26 +0000 (12:36 -0700)]
Merge "IOT-31: Add copyright info to all makefiles"
Erich Keane [Wed, 10 Sep 2014 19:34:10 +0000 (12:34 -0700)]
Merge "A fix for building the Arduino TBStack without wifi libraries. This bug was apparent even when building for only ethernet, the makefile expected the wifi libraries to exist. Now the makefile only expects the wifi libraries to exist if you're building for wifi shield."
Charunethran PG [Wed, 10 Sep 2014 17:39:16 +0000 (19:39 +0200)]
Doxygen run for *.h files
Change-Id: Ia26ec0edc0f817ac872b261e9490c2319db466b4
Signed-off-by: Charunethran PG <charunethran.panchalam.govindarajan@intel.com>
Erich Keane [Wed, 10 Sep 2014 16:52:14 +0000 (09:52 -0700)]
IOT-31: Add copyright info to all makefiles
Change-Id: Idcdbf7e9a81b5f7f28e953cf31083d083c917aea
Erich Keane [Wed, 10 Sep 2014 16:35:04 +0000 (09:35 -0700)]
Merge "Just realized that these should NOT be PCF'd. If a server or client were using these as defined to be in program memory, the stack would have been storing the location the strings were stored in memory instead of their actual values when running on Arduino. PCF has no effect on Linux."
Joseph Morrow [Wed, 10 Sep 2014 16:20:47 +0000 (12:20 -0400)]
A fix for building the Arduino TBStack without wifi libraries. This bug
was apparent even when building for only ethernet, the makefile
expected the wifi libraries to exist. Now the makefile only expects
the wifi libraries to exist if you're building for wifi shield.
Patch 2: Rebase
Patch 3: Moved implementation in to just arduinomega.properties
as per Sachin's suggestion.
Change-Id: Id06387e0d10244d71ae83a00068a0ef2f615ee84
Erich Keane [Tue, 9 Sep 2014 22:36:33 +0000 (15:36 -0700)]
Cleaned up a number of line lengths. Cleaned up iterator-for-loops vs for:each syntax for clarity where easily possible. Switched to by ref where we were previously doing unnecessary pointer parameters
Went entirely thorugh InProcClientWrapper.cpp and fixed line lengths.
Change-Id: I69f706e64f67d9a822639f239575eb316675c1d7
Joseph Morrow [Tue, 9 Sep 2014 21:26:41 +0000 (17:26 -0400)]
Just realized that these should NOT be PCF'd. If a server or client
were using these as defined to be in program memory, the stack would
have been storing the location the strings were stored in memory
instead of their actual values when running on Arduino. PCF has no
effect on Linux.
Change-Id: Iada6282befa17878f87045141580c7e330c732c1
Sashi Penta [Tue, 9 Sep 2014 20:09:51 +0000 (13:09 -0700)]
Removing unwanted .h and .cpp files from C++ stack.
Change-Id: Ib130b2c6849b4b13683229fcf0811ad246dccd9a
Erich Keane [Tue, 9 Sep 2014 18:24:46 +0000 (11:24 -0700)]
Merge "With respect to Rally User Story: US1556"
Erich Keane [Tue, 9 Sep 2014 18:20:43 +0000 (11:20 -0700)]
Missed removing OCProperties.cpp earlier
Change-Id: I52eb9255cf7a75b312a8d556f2e53bcbe81e9512
Joseph Morrow [Tue, 9 Sep 2014 18:18:50 +0000 (14:18 -0400)]
With respect to Rally User Story: US1556
and Rally Task: US1596
Updated ocstack.c to notify observers of this 'stack's presence
whenever any modification or deletion to ANY of this stack's
resource(s) occurs.
Change-Id: I0bdb72f56fa36a94e3d6d6b548f87a6a75a362ca
Erich Keane [Tue, 9 Sep 2014 17:06:37 +0000 (10:06 -0700)]
Merge "Fixing a logging bug introduced in earlier WiFi related check-in. Patch 2 : Fixing the issue caught by Joey."
Erich Keane [Tue, 9 Sep 2014 16:54:53 +0000 (09:54 -0700)]
Merge "Deleting OCReflect and OCServer, which are boht dead code. Had to remove references from a few files and fix some other includes Fixed OCResource's line lengths Change-Id: Ica6e66a1ab3ecffd6c93d8f4787a3be8a9ed1128"
Erich Keane [Mon, 8 Sep 2014 22:37:28 +0000 (15:37 -0700)]
Deleting OCReflect and OCServer, which are boht dead code. Had to remove references from a few files and fix some other includes
Fixed OCResource's line lengths
Change-Id: Ica6e66a1ab3ecffd6c93d8f4787a3be8a9ed1128
Sachin Agrawal [Tue, 9 Sep 2014 16:21:42 +0000 (09:21 -0700)]
Fixing a logging bug introduced in earlier WiFi related check-in.
Patch 2 : Fixing the issue caught by Joey.
Change-Id: Ib254202e6d87fed70de2bf8d1db639b475d377f6
Sashi Penta [Mon, 8 Sep 2014 22:14:15 +0000 (15:14 -0700)]
OCStartPresence comments in header.
Every OCStartPresence sends a random nonce.
C++ Stack cleanup to use unsigned integer for nonce.
Change-Id: If83e0868bbc05fa641cd3a0758e250b59da6393c
Sashi Penta [Mon, 8 Sep 2014 22:27:21 +0000 (15:27 -0700)]
Sending NONCE and TTL in the payload for active presence:
Patch set #1:
1) The payload is formatted as NONCE#:TTL#
2) NONCE is a 32bit unisgned random int.
3) The client app is not called when there is not change in NONCE, OC_STACK_PRESENCE_NO_UPDATE was deleted
4) Code organization in occoap.c
5) Rebased on master
Change-Id: I2c4f8c4b30c3d692a38b3286001d55f3387f5a73
Erich Keane [Mon, 8 Sep 2014 21:37:35 +0000 (14:37 -0700)]
Merge "OCStartPresence now can be called when the server comes up for the first time, when it comes up from offline, or when it reenters network."
Sashi Penta [Fri, 5 Sep 2014 04:36:27 +0000 (21:36 -0700)]
OCStartPresence now can be called when the server comes up for the first time, when it comes up from offline, or when it reenters network.
Similarly, OCStopPresence can be called when the server is terminating, going offline or when it leaves network.
Change-Id: I610979c26c26c1a7c0fcd4fb95294c5dedaeb4e4
Erich Keane [Fri, 5 Sep 2014 23:43:55 +0000 (16:43 -0700)]
Gave typedefs to every use of std::function in the C++ stack. Only existing use of std::function< without a typedef is a comment in OCClientTestApp.cpp
Patch 2: Sudarshan's review fixes
Change-Id: I8c62347c1091a2225d5db1b0bda5221a3346586e
Erich Keane [Mon, 8 Sep 2014 19:50:07 +0000 (12:50 -0700)]
Merge "Get/Put overload implementation in OCResource. Allows for specifiying interface and type."
Erich Keane [Mon, 8 Sep 2014 18:28:30 +0000 (11:28 -0700)]
Merge "Patch 1 : Adding Arduino WiFi support. This requires updated Arduino WiFi Shield firmware Patch 2 : Removed ocserver_wifi app and incorporated wifi changes in the single ocserver application. Updated location of Time library in arduinomega and arduinodue properties file. Added some more comments in sample app. Added license info in ocserver which was missing."
Erich Keane [Mon, 8 Sep 2014 18:03:27 +0000 (11:03 -0700)]
Merge "Fix for arduinomega.properties. It was pointing at an invalid location for avrdude with arduino 1.0.5."
Erich Keane [Mon, 8 Sep 2014 17:57:13 +0000 (10:57 -0700)]
Merge "Stack/Observer: Free OCEntityHandlerRequest object after sent notification"
Erich Keane [Thu, 4 Sep 2014 18:23:16 +0000 (11:23 -0700)]
Get/Put overload implementation in OCResource. Allows for specifiying interface and type.
Patch 2: fixing the emplace for a crappy old compiler
Patch 3: Sashi's comments
Patch 4: Sudarshan's comments
Change-Id: Ic1e9c07424229325443c532bcb03c67c5bfce6f1
Sachin Agrawal [Mon, 8 Sep 2014 16:17:28 +0000 (09:17 -0700)]
Patch 1 : Adding Arduino WiFi support. This requires updated Arduino WiFi Shield firmware
Patch 2 : Removed ocserver_wifi app and incorporated wifi changes in the single ocserver application.
Updated location of Time library in arduinomega and arduinodue properties file.
Added some more comments in sample app.
Added license info in ocserver which was missing.
Change-Id: Ie852a61ff3b0b5707f79899427668698792d1fcc
Joseph Morrow [Fri, 5 Sep 2014 16:39:15 +0000 (12:39 -0400)]
Fix for arduinomega.properties. It was pointing at an invalid
location for avrdude with arduino 1.0.5.
Change-Id: I15878e29a2daa341886d0b8e0fa2010f21771f32
Erich Keane [Fri, 5 Sep 2014 17:35:33 +0000 (10:35 -0700)]
Updated the name of the API to unregisterResource. And this API returns
whatever result it got from the stack. NOTE: Only this API has been changed.
Added C++ API for unregistering a resource
Patch 3: Rebase
Patch 4: added test code, fixed ocicuc bug
Change-Id: Ida95886fcbbaceab349bc127ae4b4e44ff823828
Erich Keane [Fri, 5 Sep 2014 16:18:02 +0000 (09:18 -0700)]
Merge "OCLib is now called src. Now liboc.a is generated instead OCLib.a ."
Joseph Morrow [Fri, 5 Sep 2014 15:17:52 +0000 (11:17 -0400)]
Using the compiler that comes with Arduino 1.0.5, it
complains that the token in the async struct cannot
have a size of zero.
I fixed it, since it's the end of the struct, this is okay.
Patch 2: Updated with respect to Sachin's comment review.
Patch 3: When building in debug mode, the compiler that
comes with Arduino 1.0.5 complains about the tokens defined
in pdu.h as well. In this patch as well, also removed errors
found when building in debug. Rebased on Sachins changeset.
Change-Id: I568481562af2a614624608851def5bf5aa68287a
Sashi Penta [Fri, 5 Sep 2014 04:10:47 +0000 (21:10 -0700)]
OCLib is now called src. Now liboc.a is generated instead OCLib.a .
Change-Id: Id9aa7652987a0915842fa6441efa118809a0b01b
Sachin Agrawal [Fri, 5 Sep 2014 01:27:02 +0000 (18:27 -0700)]
Fixing broken TB stack due to compilation errors for Arduino platforms.
Change-Id: Id211b8f03ddb07dd8c50148ece513bd2548b447f
Wu Jiangbo [Wed, 3 Sep 2014 11:08:23 +0000 (19:08 +0800)]
OCResourceRespnse: JSON value should be wrapped in double quotes
C++ stack packages the JSON payload and send it as payload within CoAP.
And the JSON format said, "A string is a sequence of zero or more
Unicode characters, wrapped in double quotes, using backslash escapes."
Change-Id: I9641df115945dead2dcd41a08576047dcdd01d98
Sashi Penta [Wed, 3 Sep 2014 21:01:11 +0000 (14:01 -0700)]
Minor formatting fix for the release notes.
Change-Id: Ieee8839edbaef07ae5ab18efb81162c99d70d440
Wu Jiangbo [Wed, 3 Sep 2014 07:09:02 +0000 (15:09 +0800)]
Stack/Observer: Free OCEntityHandlerRequest object after sent notification
Change-Id: I539778ed8ad65c45635ff64f0b71a5ade1743894
Sashi Penta [Wed, 3 Sep 2014 04:10:52 +0000 (21:10 -0700)]
Fixed a minor bug in C stack and minor cleanup from Presence Notification.
Release notes from Sudarhan, with minor comments from Sashi.
OCPlatform.h with better comments from Sudarshan.
Change-Id: Ie1950b6d3486e3a2314ac7428350f61bba20351f
Sashi Penta [Wed, 3 Sep 2014 03:17:01 +0000 (20:17 -0700)]
Includes C++ Presence Notification changes from Erich, Jesse, Sudarshan and Sashi.
Change-Id: I113cd2dc3051bace0d0111289184299e5fd21f87
Erich Keane [Wed, 3 Sep 2014 02:50:26 +0000 (19:50 -0700)]
Merge "Patch 1 (Yamin): This is where all work is merged from Sashi, Joey and myself"
Yamin S Al-mousa [Wed, 3 Sep 2014 02:16:08 +0000 (19:16 -0700)]
Patch 1 (Yamin): This is where all work is merged from
Sashi, Joey and myself
Patch 2 (Sashi): Looks like some debugging printf work. Should
be removed. Sashi?
Patch 3 (Joey): Applying changes as per review by Chanchala in parent
changeset Change-Id: I2625063e4a6bdfa53e3954310a060ea3c0ee5d3c
Patch 4 (Joey & Yamin): Squash all of Active Discovery and rebase
on to master.
Patch 5 (Joey): Cleaned up warnings and refactored merge conflicts
that caused build server to fail.
Patch 6 (Yamin): More debugging and restructuring... more work needed specifically for the TTL logic
Patch 7 and 8 (Yamin): Fixed TTL logic and presence cancelation, more clean up to come...
Patch 9: Added the random TTL logic and cleaned up.
Change-Id: I4bbd72e066ed4b50f4d2be52367844918a29823a
Jesse Williamson [Fri, 29 Aug 2014 23:53:30 +0000 (16:53 -0700)]
Checkpoint: monoprocess bug fixed-- need to improve CLI handling.
Change-Id: I73ad330df3716cda2f5907a62dc09478bc03fe5e
Erich Keane [Fri, 29 Aug 2014 16:27:09 +0000 (09:27 -0700)]
Merge "Updated Makefile to support different build targets. Removed delegating ctors (work-around non-conforming compiler). Configuration requires explicit ctor. Restructured client/server program; review fixes; hopefully extricated from the dangerous world of b0rked commits."