From mach at zhaw.ch Fri Feb 7 11:39:01 2014 From: mach at zhaw.ch (Christof Marti) Date: Fri, 7 Feb 2014 11:39:01 +0100 Subject: [Miwi-middleware] Meeting canceled Message-ID: Hi As discussed last week we skip todays meeting. Please check the minutes and my email from this wednesday regarding open action points: e.g. - Add/schedule items for Sprint 3.3.2 (February) in the tracker (until Monday noon). - Fix errors on guides (see review dashboard from miguel: https://docs.google.com/spreadsheet/ccc?key=0AqLSWp0KXaaDdGZ0ekFhREgyMFdoT1owZ1BYdFlnbGc&usp=docslist_api#gid=2 @Jaime: In the last meeting you mentioned to release RPC over REST this week. Any progress? Best regards and have a good weekend. Christof ---- InIT Cloud Computing Lab - ICCLab http://cloudcomp.ch Institut of Applied Information Technology - InIT Zurich University of Applied Sciences - ZHAW School of Engineering Phone: +41 58 934 70 63 Skype: christof-marti From rubinstein at cs.uni-saarland.de Thu Feb 13 12:54:33 2014 From: rubinstein at cs.uni-saarland.de (Dmitri Rubinstein) Date: Thu, 13 Feb 2014 12:54:33 +0100 Subject: [Miwi-middleware] KIARA: Version 0.9.3 Message-ID: <52FCB279.3000409@cs.uni-saarland.de> Hi. I finished and pushed to ZHAW Gitorious Repos a new version of KIARA. This is a minor feature and bugfix release. * New features/bugfixes: 1. Automatic allocation with malloc and initialization of C-structs. In previous release all structs required explicit allocators and deallocators. Here example from the documentation: KIARA_DECL_STRUCT_WITH_API(LocationList, KIARA_STRUCT_ARRAY_MEMBER(Location_ptr, locations, KIARA_INT, num_locations) , KIARA_USER_API(AllocateType, LocationList_Allocate) KIARA_USER_API(DeallocateType, LocationList_Deallocate) ) LocationList_Allocate allocated list with malloc and initialized it. LocationList_Deallocate deallocated the list. Starting with the current release both can be omitted: KIARA_DECL_STRUCT(LocationList, KIARA_STRUCT_ARRAY_MEMBER(Location_ptr, locations, KIARA_INT, num_locations) ) KIARA compiler will generate default allocator and deallocator which will call malloc/free. Then compiler will generate constructor and destructor which will initialize all members of the structure. Note that currently only array members declared with KIARA_STRUCT_ARRAY_MEMBER are supported. All other members are not initialized. 2. Semantic annotation of types. In order to generate correct code we need to distinguish between different types. However, in C/C++ primitive types like char* can mean different things: either pointer to raw data ot pointer to null-terminated string. It is not possible to distinguish between the two uses based on the type. In the current version the type system was extended with annotated types. Annotated type is an alias for another type and annotates it with additional semantic. Currently this feature is mostly used internally. But public API now provides two char* types: KIARA_CHAR_PTR and KIARA_RAW_CHAR_PTR KIARA_CHAR_PTR is internally an alias to the representation of char* with additional information that it is a pointer to the null terminated string. KIARA_RAW_CHAR_PTR is just a pointer to char or int8_t and has no additional semantic. The difference is the generated code when type is allocated/deallocated and serialized/deserialized. 3. Fixed bug when registering and printing of recursive types like: typedef struct IntList { struct IntList *next; int data; } IntList; There are no changes in installation instructions. Best, Dmitri From Philipp.Slusallek at dfki.de Fri Feb 21 07:42:08 2014 From: Philipp.Slusallek at dfki.de (Philipp Slusallek) Date: Fri, 21 Feb 2014 07:42:08 +0100 Subject: [Miwi-middleware] Call today Message-ID: <5306F540.2040104@dfki.de> Hi, We have Audi visiting DFKI and IVCI today but I would have had tome to sneek out for our meeting. However, I just learned that my co-director who was organizing the event caught a bad flu yesterday. As a result I will have to take over for him today and will not be able to make it, sorry. Anyway, Dmitri has some nice new results -- but he can talk about that himself in the call. Best, Philipp -- ------------------------------------------------------------------------- Deutsches Forschungszentrum f?r K?nstliche Intelligenz (DFKI) GmbH Trippstadter Strasse 122, D-67663 Kaiserslautern Gesch?ftsf?hrung: Prof. Dr. Dr. h.c. mult. Wolfgang Wahlster (Vorsitzender) Dr. Walter Olthoff Vorsitzender des Aufsichtsrats: Prof. Dr. h.c. Hans A. Aukes Sitz der Gesellschaft: Kaiserslautern (HRB 2313) USt-Id.Nr.: DE 148646973, Steuernummer: 19/673/0060/3 --------------------------------------------------------------------------- -------------- next part -------------- A non-text attachment was scrubbed... Name: slusallek.vcf Type: text/x-vcard Size: 441 bytes Desc: not available URL: From rubinstein at cs.uni-saarland.de Fri Feb 21 10:37:03 2014 From: rubinstein at cs.uni-saarland.de (Dmitri Rubinstein) Date: Fri, 21 Feb 2014 10:37:03 +0100 Subject: [Miwi-middleware] KIARA: Version 0.10.0 Message-ID: <53071E3F.8080809@cs.uni-saarland.de> Hi. I finished and pushed to ZHAW Gitorious Repos a new version of KIARA. This is a major feature release. * New features: 1. A new network stack was implemented with TCP-based transport additionally to HTTP-based. Now it is possible to create connections that communicate with less overhead. The API was not changed much: Old version: // Server waiting for the incoming negotiation connections (over HTTP!) // on specified port and HTTP path is "/service" : KIARA_Server *server; server = kiaraNewServer(ctx, "0.0.0.0", atoi(port), "/service"); // The specified service is provided via HTTP on path "/rpc/aostest" kiaraAddService(server, "/rpc/aostest", protocol, service); In the new version path is an URL (similar to ZMQ address). When it is relative it refers to the HTTP negotiation server created with kiaraNewServer. When it is absolute KIARA starts a new HTTP server: // New HTTP Server on port 9090 kiaraAddService(server, "http://0.0.0.0:9090/rpc/aostest", protocol, service); TCP works in the same way: kiaraAddService(server, "tcp://0.0.0.0:2020", protocol, service); Internally each transport has a priority, and TCP has a higher priority then HTTP. Thus, when service is available over HTTP and TCP, TCP will be used. Currently protocol is not taken into account. The implemented TCP transport is simple. First comes 32-bit little endian number which specifies a size of the message, then message itself. Errors are represented by empty messages (size == 0). There are no changes in installation instructions. Best, Dmitri From mach at zhaw.ch Fri Feb 21 13:09:49 2014 From: mach at zhaw.ch (Christof Marti) Date: Fri, 21 Feb 2014 13:09:49 +0100 Subject: [Miwi-middleware] Agenda-minutes Message-ID: <288C0417-4BC3-4370-9CBA-DBA4ADDCA137@zhaw.ch> Hi Just got back from my examination and are still a bit groggy. But I would like to have at least a short meeting to catch up and sync. Minutes are here: https://docs.google.com/document/d/11D24v0F3ZYRN_3c8uXzEJBRB6Q-rFbeLtL4i60V6ces/edit?usp=sharing Hangout is here: https://docs.google.com/document/d/11D24v0F3ZYRN_3c8uXzEJBRB6Q-rFbeLtL4i60V6ces/edit?usp=sharing Cheers Christof From rubinstein at cs.uni-saarland.de Thu Feb 27 18:34:06 2014 From: rubinstein at cs.uni-saarland.de (Dmitri Rubinstein) Date: Thu, 27 Feb 2014 18:34:06 +0100 Subject: [Miwi-middleware] KIARA: Version 0.10.1 Message-ID: <530F770E.5050003@cs.uni-saarland.de> Hi. I finished and pushed to ZHAW Gitorious Repos a new version of KIARA. This is a minor bugfix release. * Bugfixes: 1. Fixed memory leak when service method in the IDL was registered twice. 2. Fixed memory leak in MCJIT support code. 3. Improved testing with valgrind by auto-flusing stdout. There are no changes in installation instructions. Best, Dmitri From mach at zhaw.ch Fri Feb 28 11:50:19 2014 From: mach at zhaw.ch (Christof Marti) Date: Fri, 28 Feb 2014 11:50:19 +0100 Subject: [Miwi-middleware] Cancel todays KIARA Meeting Message-ID: Hi I?m sick at home and have to cancel todays KIARA meeting from my side. All the urgent important points where discussed last week. Main action points at the moment are: - Sprint 3.3.2 ends today. Please update your tracker entries - close done sprint3.3.2 entries - reschedule if not finished (should be exception) - prepare new entries for next Sprint 3.3.3. - fix those entries with issues according to Manuels analysis (if not yet done) - Continue integration work Upcoming activities - New version of Open Specification and OpenAPI Specification is due mid March for Middleware. I will send a separate email regarding organizing the work. Best regards Christof ---- InIT Cloud Computing Lab Institut of Applied Information Technology - InIT Zurich University of Applied Sciences - ZHAW School of Engineering Phone: +41 58 934 70 63 Skype: christof-marti From rubinstein at cs.uni-saarland.de Fri Feb 28 22:59:33 2014 From: rubinstein at cs.uni-saarland.de (Dmitri Rubinstein) Date: Fri, 28 Feb 2014 22:59:33 +0100 Subject: [Miwi-middleware] Dmitri is on vacation Message-ID: <531106C5.10301@cs.uni-saarland.de> Hi, I am on vacation next week. Best, Dmitri