Activity
From 07/31/2022 to 08/29/2022
08/29/2022
- 02:53 PM Feature #6650: improve performance of Scopeable notification in TM.processScopeNotifications
- I can not find a good way to limit BufferManager/ChangeBroker. I have changes which sync them (so they are always no...
08/23/2022
- 01:56 AM Bug #6693: StringFormat fixes - finish up and merge branch 4880a|b|c
- Although the main fix for a @<complex_gui_application>@ is done (see #4880), a lot of improvements have already been ...
- 01:52 AM Bug #6693 (New): StringFormat fixes - finish up and merge branch 4880a|b|c
08/22/2022
- 04:11 PM Bug #6640: parsing bug for initial ",0000" for decimal variables or fields
- > Also, we need to check if this must be configured in p2j.cfg.xml or not (i.e. if it must consider -numsep at the 4G...
08/16/2022
- 03:25 PM Feature #6650: improve performance of Scopeable notification in TM.processScopeNotifications
- Constantin Asofiei wrote:
> * @DatabaseTriggerManager@ - the problem here is that the context-local instance is refe... - 02:55 PM Feature #6650: improve performance of Scopeable notification in TM.processScopeNotifications
- Ovidiu and Eric: Please review the persistence related Scopeables in #6650-2 and see if you can offer some guidance.
- 02:54 PM Feature #6650: improve performance of Scopeable notification in TM.processScopeNotifications
- > For some cases (like UnnamedStreams and SharedVariableManager), the decision is easy - if the context-local WorkAre...
- 09:01 AM Feature #6650: improve performance of Scopeable notification in TM.processScopeNotifications
- The complex part of the task is to determine where and when to register a Scopeable with the next (or current) block....
- 07:07 AM Feature #4065: server-side processing of client platform dependencies
- Yes, this is a flaw but it is better than the alternative.
Can we detect when the same lvalue is re-assigned a new... - 06:57 AM Feature #4065: server-side processing of client platform dependencies
- Greg Shah wrote:
> We don't actually allocate any additional memory, but there will be an extra entry in our registr... - 06:56 AM Feature #4065: server-side processing of client platform dependencies
- > FWD will keep it allocated in the context's address space, as it 'allocated' this address space when the RETURN or ...
- 06:47 AM Feature #4065: server-side processing of client platform dependencies
- Greg Shah wrote:
> > My point is if the memory is not deallocated on FWD server, this will lead to a memory leak in ... - 06:45 AM Feature #4065: server-side processing of client platform dependencies
- > I understand, but I have at least a case in #6277 where the RETURN parameter is not being deallocated on the FWD se...
- 06:28 AM Feature #4065: server-side processing of client platform dependencies
- Constantin Asofiei wrote:
> ... in this case, the library call is https://www.openssl.org/docs/man1.1.1/man3/EVP_sha... - 06:26 AM Feature #4065: server-side processing of client platform dependencies
- Greg Shah wrote:
> In this case, we can augment the RETURN or OUTPUT parameter with additional size information. Ac... - 06:21 AM Feature #4065: server-side processing of client platform dependencies
- > The only solution I see is to disable the address space check for RETURN and OUTPUT parameters, in the FWD runtime....
- 05:46 AM Feature #4065: server-side processing of client platform dependencies
- Greg Shah wrote:
> Perhaps we should require that the 4GL code is modified to tell us the size of the returned point... - 05:39 AM Feature #4065: server-side processing of client platform dependencies
- > But the deregistration of this pointer from the FWD context's address space may not happen, if the memptr is not de...
- 05:06 AM Feature #4065: server-side processing of client platform dependencies
- Another approach would be to allow memptr to be assigned addresses returned by the native API calls, and not register...
- 05:05 AM Feature #4065: server-side processing of client platform dependencies
- There is an issue with server-side memptr which affects #6277. The problem is related to memptr used at @define retu...
08/02/2022
- 09:57 AM Feature #6416: explore and implement ideas to make OO 4GL more efficient
- Constantin Asofiei wrote:
> A part of the problem is the instance initialization in 4GL; the constructors are calle...
08/01/2022
- 12:58 PM Feature #6650: improve performance of Scopeable notification in TM.processScopeNotifications
- I've tried a simple looping block executed (via Java FOR statement) 100k times and the times are with/without executi...
- 12:54 PM Feature #6650 (WIP): improve performance of Scopeable notification in TM.processScopeNotifications
- 10:44 AM Feature #6416: explore and implement ideas to make OO 4GL more efficient
- Greg Shah wrote:
> What specific features of the constructor processing are impacted if we used direct construction ... - 10:35 AM Feature #6416: explore and implement ideas to make OO 4GL more efficient
- What specific features of the constructor processing are impacted if we used direct construction (did not properly du...
- 10:19 AM Feature #6416: explore and implement ideas to make OO 4GL more efficient
- A part of the problem is the instance initialization in 4GL; the constructors are called *after* the initialization ...
- 10:25 AM Feature #6649: improve the performance or SourceNameMapper runtime
- @SourceNameMapper@ APIs like @getExternalProgram@ or @getInternalEntry@ get called each time something dynamic is exe...
- 10:21 AM Feature #6649 (Test): improve the performance or SourceNameMapper runtime
- 10:22 AM Support #4061: base language performance work
- * #4149 - BlockManager and TransactionManager handle TransactionType.NONE incorrectly
* #6649 - SourceNameMapper per... - 07:38 AM Feature #1970: improve ControlFlowOps method resolution by reflection
- In regard to my idea to use a call-site specific "token" (e.g. @private static final Object _CALL_SITE_TOKEN_xyz = ne...
07/31/2022
- 05:58 AM Feature #4350: method overload when they differ by a temp-table, dataset, buffer or object or ext...
- Runtime for constructor overload with @object@ arguments is not yet complete. Having two constructors like:...
Also available in: Atom