Activity
From 11/08/2023 to 12/07/2023
12/07/2023
- 08:53 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- Alexandru Lungu wrote:
> Regarding your last comment: it depends upon the delegate right? If the delegate is not len... - 08:40 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- Eduard, I started converting a really large application with 7045b (on top of 7156b). You have 3 days left to do the ...
- 08:34 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- I think is not needed, as @QueryWrapper@ is used in conversion of @open query@ statement. And in this case QQE is not...
- 08:04 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- Alexandru, I haven't looked at the full code, but is the @QueryWrapper.handleQueryOffEnd@ still needed?
- 07:08 AM Feature #6417 (WIP): cleanup some known items in appserver support
12/06/2023
- 06:18 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- Eduard, please focus on reaching a "state-of-the-art" solution for 7045b (also referring to #7045-84). Also, triple c...
- 02:00 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- Alexandru Lungu wrote:
> I've run 7045b with the performance tests and got 8.168s. Comparing to the baseline of 7.95...
12/05/2023
- 11:01 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- I am looking now in the changes of 7045b:
* Eduard, mind that you have some hard-tabs in your last commit (@Adaptive... - 08:58 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Hynek Cihlar wrote:
> The logic gets more and more complex so it would be nice to be able to know how does it perfor... - 05:33 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Hynek Cihlar wrote:
> With the LinkedBlockingQueue on the limit, adding more tasks to the executor will cause Reject... - 05:14 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Dănuț Filimon wrote:
> Hynek Cihlar wrote:
> Initially the cache entry was a @Map@, but a potential leak was notice... - 03:15 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Hynek Cihlar wrote:
> In @createSourceCacheEntry@ the object submitted to the @mappingExecutor@ is a @Thread@, but i... - 05:51 AM Bug #8008: read-json on a 1252 longchar should convert to utf-8 automatically before hydrating te...
- In this Note, I summarize the parts from #7944 which were out of scope there.
The @convert_json_1252_fcp_memptr.p@ f...
12/04/2023
- 10:38 AM Feature #7045 (Internal Test): re-implement "normal" (non-abend) stack unwinding to avoid throwin...
- I've run 7045b with the performance tests and got 8.168s. Comparing to the baseline of 7.958s, this is +2.5% slower. ...
- 05:54 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- Committed on 7045b, revision 14835.
I tested @uast/query_off_end/qoe_exception@ , did not see any regressions. In ... - 03:47 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- Eduard, please address the review and commit a final solution to 7045b (including javadoc / history entries / etc.) a...
- 03:07 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- Eric Faulhaber wrote:
> I have noted the discussion above that @QOEE@ is not yet fully eliminated. I see this, for i... - 03:25 AM Bug #7460: implement Progress.Lang.Class:getMethods
- 6410b rev 14828 was rebased from trunk rev 14856 - new rev 14860
- 03:19 AM Bug #7460: implement Progress.Lang.Class:getMethods
- Constantin Asofiei wrote:
> Yes, I'll need to rebase. Are all changes in 6410b?
I have some local changes but no... - 03:00 AM Bug #7460: implement Progress.Lang.Class:getMethods
- Yes, I'll need to rebase. Are all changes in 6410b?
- 02:50 AM Bug #7460: implement Progress.Lang.Class:getMethods
- Constantin Asofiei wrote:
> Created task branch 7460a from trunk rev 14850.
>
> 7460a rev 14851 has a regression ... - 03:24 AM Feature #6410: implement additional built-in OO classes/interfaces
- 6410b rev 14828 was rebased from trunk rev 14856 - new rev 14860
12/02/2023
- 03:36 PM Feature #6649: improve the performance or SourceNameMapper runtime
- Code review 6649b revisions 14840..14847.
In @createSourceCacheEntry@ the object submitted to the @mappingExecutor...
12/01/2023
- 06:00 AM Bug #8084 (WIP): refactor the 'execute' method for legacy classes
- Created task branch 8084a from trunk rev 14853
11/30/2023
- 01:53 PM Bug #7460: implement Progress.Lang.Class:getMethods
- Branch 7460a was merged into trunk as rev. 14851 and archived.
- 01:38 PM Bug #7460 (Merge Pending): implement Progress.Lang.Class:getMethods
- Code Review Task Branch 7460a Revisions 14850 and 14851
No objections. - 12:52 PM Bug #7460: implement Progress.Lang.Class:getMethods
- Created task branch 7460a from trunk rev 14850.
7460a rev 14851 has a regression fix: Do not use 'TypeFactory.obje...
11/29/2023
- 08:59 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- I am addressing some regression on running POC. But I think, I am close to find the problem causing this regression. ...
- 08:35 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- Eduard, please make an update on your status of running the POC with your changes and addressing Eric's #7045-77.
11/28/2023
- 09:20 AM Feature #6649: improve the performance or SourceNameMapper runtime
- I ran a large POC with *7156b/rev.14846* and the changes from *6649b/rev.14835 - 14847* and obtained good results (th...
- 05:50 AM Feature #6649 (Review): improve the performance or SourceNameMapper runtime
- *Committed 6649b/rev.14847* which solves the problem mentioned in #6649-144. By not keeping the thread idle, the corr...
- 03:49 AM Bug #8084: refactor the 'execute' method for legacy classes
- Oh right, this makes sense - getting rid of the @BlockManager@ API. This way we avoid doing @ProcedureManager@ work.
- 03:31 AM Bug #8084: refactor the 'execute' method for legacy classes
- What I meant is to emit all the code (like buffer scopes) directly in the Java method, and do not enclose it in @Bloc...
- 03:29 AM Bug #8084: refactor the 'execute' method for legacy classes
- Contantin, just a question here: I thought we open scopes in @externalProcedure@ due to the fact that we are aware *o...
11/27/2023
- 02:28 PM Bug #8084: refactor the 'execute' method for legacy classes
- Currently, each legacy class has a synthetic 'execute' instance method which includes initialization like instance-le...
- 02:23 PM Bug #8084 (Test): refactor the 'execute' method for legacy classes
11/24/2023
- 06:11 AM Feature #6649 (WIP): improve the performance or SourceNameMapper runtime
- 06:05 AM Feature #6649: improve the performance or SourceNameMapper runtime
- After investigating, I found out that the @mappingExecutor@ creates an @AssociatedThread@ with the current context, m...
- 02:11 AM Feature #6649: improve the performance or SourceNameMapper runtime
- There is an issue with how the cache entries are built. The @AssociatedThread@ doesn't use the correct context when b...
- 04:11 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- Code review 7045b/14825-14834:
There are file header entries problems (i.e., duplicate entries in some files, miss...
11/22/2023
- 03:02 PM Feature #6819: refactor FWD proxy implementation to use ReflectASM instead of Java Method reflection
- > > What specific use cases would be using the modified code?
> RecordBuffer, TemporarBuffer, DataSet, ControlFlowOp... - 08:56 AM Feature #6819: refactor FWD proxy implementation to use ReflectASM instead of Java Method reflection
- Greg Shah wrote:
> Please post the test harness code or whatever it is that you are using to create the failing case... - 08:22 AM Feature #6819: refactor FWD proxy implementation to use ReflectASM instead of Java Method reflection
- Please post the test harness code or whatever it is that you are using to create the failing case ("where I created 1...
- 04:11 AM Feature #6649: improve the performance or SourceNameMapper runtime
- I tested with a rebased 6649b from 7156b (7156b at rev. 14846, branched from trunk rev. 14822 + all 8 commits from 66...
11/21/2023
- 02:43 PM Feature #6819: refactor FWD proxy implementation to use ReflectASM instead of Java Method reflection
- Currently there are three experimental branches:
* 6819c - uses singleton, byte-code generated, BiFunction inner cla... - 01:56 PM Feature #6819: refactor FWD proxy implementation to use ReflectASM instead of Java Method reflection
- Please post details of the results from your testing described in #6819-61 (6 months ago).
I understand you've bee... - 02:40 PM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- Eric: Please don't forget to do a code review ASAP.
- 07:52 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- Eduard, I didn't had the chance to convert yet the POC. Can you attempt it? I have already queued some other performa...
- 11:24 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Alex, when you test the large app, do you use 7156b with the changes from 6649b on top? Will the latest 7156b work OK?
- 09:35 AM Feature #6649: improve the performance or SourceNameMapper runtime
- *6649b* was rebased to *trunk/rev.14834* and the current revision is *14845*.
*Committed 6649b/rev.14846*:
* Used... - 07:29 AM Bug #7656: investigate performance of context-local helpers (e.g. TransactionManager.TransactionH...
- Alexandru, in @devsrv01:/tmp/AccessModifier.7z@ there is the project I use to automatically 'fix' the access modifie...
11/20/2023
- 06:03 AM Feature #6649: improve the performance or SourceNameMapper runtime
- On a large POC, the total improvement of the changes in 6649b was of -0.3%.
- 04:00 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Hynek Cihlar wrote:
> Do you mean 70 individual propath entries or 70 paths in a single propath?
One single long ... - 02:34 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Alexandru Lungu wrote:
> I've done some tests, and there were no obvious problems. I am moving this to the profiler ...
11/18/2023
- 07:33 PM Feature #6431: implement support to allow CPINTERNAL set to UTF-8 and CP936
- Marian Edu wrote:
> Hi Theodoros, just pushed the i18n tests on testcases (rev#1514), we're still working one some c...
11/17/2023
- 08:47 AM Feature #6431: implement support to allow CPINTERNAL set to UTF-8 and CP936
- Theodoros Theodorou wrote:
> Hello Marian. Greg asked me to coordinate with you that we do not do the same effort tw... - 07:41 AM Feature #6431: implement support to allow CPINTERNAL set to UTF-8 and CP936
- Hello Marian. Greg asked me to coordinate with you that we do not do the same effort twice. Please have a look at @te...
- 07:11 AM Feature #6431: implement support to allow CPINTERNAL set to UTF-8 and CP936
- Yes, it needs to be fixed. Such conversions should yield identical results. It seems likely that the cause is relat...
- 08:23 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Alexandru Lungu wrote:
> Mind that the search-path cache is not synchronized. Please add @synchronize@ to that acces... - 08:13 AM Feature #6649: improve the performance or SourceNameMapper runtime
- *Review of 6649b/rev.14808*
Mind that the search-path cache is not synchronized. Please add @synchronize@ to that ... - 08:03 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Alexandru Lungu wrote:
> Some extra ideas:
>
> * Also, what about @convertedNames@? Can't it possible leak as we ... - 07:33 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Alexandru Lungu wrote:
> Danut please fix #6649-131 - I am planning to start performance testing with 6649a as soon ... - 07:08 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Some extra ideas:
* Also, what about @convertedNames@? Can't it possible leak as we are constantly adding to it th... - 06:26 AM Feature #6649: improve the performance or SourceNameMapper runtime
- I've done some tests, and there were no obvious problems. I am moving this to the profiler asap.
However, I want t... - 04:04 AM Feature #6649: improve the performance or SourceNameMapper runtime
- I converted a large application and generated a @name_map.xml@ with ~39k (~7k of OO classes) class-mappings and only ...
11/16/2023
- 06:03 PM Feature #6431: implement support to allow CPINTERNAL set to UTF-8 and CP936
- Hello, I wrote a small test (chr_and_asc.p) which tests the conversion of a character from one code page to another (...
- 12:40 AM Feature #6649: improve the performance or SourceNameMapper runtime
- *Committed 6649b/rev.14807*. Fixed a NPE when there are no path-mapping nodes in name_map.xml.
11/15/2023
- 08:12 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- Understood. For such extensive changes to our control flow, it seems necessary.
- 08:09 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- Greg, I can do that for POC, but it will take me a day to have it converted. Also mind that I am already reconverting...
- 07:31 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- If you think it is ready, let's get ChUI regression tests, ETF and the big POC tested.
- 03:40 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- Converted and setup a large customer application using 7045b branch, and used it for a bit. Did not noticed any major...
11/13/2023
- 05:16 AM Feature #6649 (Internal Test): improve the performance or SourceNameMapper runtime
- 05:09 AM Feature #6649: improve the performance or SourceNameMapper runtime
- I also tried to find out the cause of the incremental version where nodes are duplicated but after running a @./gradl...
- 05:06 AM Feature #6649 (WIP): improve the performance or SourceNameMapper runtime
- *Committed 6649b/rev.14805 and 6649b/rev.14806*:
* Made @sourceCache@ configurable through @CacheManager@. I previou...
11/12/2023
- 04:29 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Code review 6649b revision 14804.
@initializationPaths.size() == 0@ in @createSourceCacheEntry@ should be protecte...
11/10/2023
- 09:56 AM Feature #6819: refactor FWD proxy implementation to use ReflectASM instead of Java Method reflection
- That code resolves the @MethodAccess@ target for a @java.lang.reflect.Method@. @Method@ instances are not singleton,...
- 09:50 AM Feature #6819: refactor FWD proxy implementation to use ReflectASM instead of Java Method reflection
- Constantin, I reached quite randomly in @Utils.invoke@ and saw the double map access on @METHOD_DATA@ and @methodAcce...
- 09:02 AM Feature #6649 (Review): improve the performance or SourceNameMapper runtime
- *Committed 6649b/rev.14804*.
* After solving the path-mapping node duplication problem, I reverted the changes from ... - 03:23 AM Feature #6410: implement additional built-in OO classes/interfaces
- Marian, please remind me of your issues - I plan to look at them over the next week. Also, point me to any test repr...
11/09/2023
- 09:31 AM Feature #6649: improve the performance or SourceNameMapper runtime
- > The number look good on the paper. The question is how will it perform in the actual customer scenario.
I can ta... - 09:26 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Dănuț Filimon wrote:
> I retested the performance of the cache after saving the @fastConvertName@ results of program... - 06:17 AM Feature #6649: improve the performance or SourceNameMapper runtime
- I retested the performance of the cache after saving the @fastConvertName@ results of program names that are not foun...
- 02:23 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Dănuț Filimon wrote:
> Got it, I went and deleted @checkSourceCache@ and integrated it's functionality in @fastConve... - 01:01 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Got it, I went and deleted @checkSourceCache@ and integrated it's functionality in @fastConvertName@, this was the be...
- 03:59 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- I rebased 7045b to latest trunk. It is now at rev. 14833.
Eduard, please double-check: @RecordBuffer@ (especially @s...
11/08/2023
- 10:50 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Dănuț Filimon wrote:
> The problem is that I can't differentiate between which case is executed. I am still thinking... - 08:15 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Hynek Cihlar wrote:
> The cache should eliminate 100% of the heavy execution of @fastConvertName@ for all the encoun... - 08:06 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Dănuț Filimon wrote:
> * The number of cache hits (when cache returned a value) was *21288* from the total *35971* @... - 07:22 AM Feature #6649 (WIP): improve the performance or SourceNameMapper runtime
- 07:21 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Hynek Cihlar wrote:
> Code review 6649b revisions 14800..14803.
>
> Saving the RUN entries should be saved in a s... - 04:03 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Hynek Cihlar wrote:
> Does it mean that the non-program files don't require the additional @fastConvertName@ logic (... - 03:37 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Dănuț Filimon wrote:
> Hynek Cihlar wrote:
> > Dănuț Filimon wrote:
> > > In my opinion, adding images to @path-ma... - 03:15 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Code review 6649b revisions 14800..14803.
Saving the RUN entries should be saved in a set rather than saving direc... - 02:35 AM Feature #6649: improve the performance or SourceNameMapper runtime
- Hynek Cihlar wrote:
> Dănuț Filimon wrote:
> > In my opinion, adding images to @path-mappings@ will just slow down ... - 09:55 AM Feature #6410: implement additional built-in OO classes/interfaces
- 6410b rev 14794 was rebased from trunk rev 14824 - new rev 14828.
- 09:52 AM Feature #6410: implement additional built-in OO classes/interfaces
- Starting rebase now.
- 08:55 AM Feature #6410: implement additional built-in OO classes/interfaces
- Constantin Asofiei wrote:
> OK, let me know when it's done.
All done what I've could get 'proper' enough... rev #14... - 07:21 AM Feature #6410: implement additional built-in OO classes/interfaces
- Marian Edu wrote:
> Constantin Asofiei wrote:
>
> > If you have nothing else to commit, I'll rebase.
> Give me 1... - 07:18 AM Feature #6410: implement additional built-in OO classes/interfaces
- Constantin Asofiei wrote:
> If you have nothing else to commit, I'll rebase.
Give me 10-15 mins please to commit ... - 07:14 AM Feature #6410: implement additional built-in OO classes/interfaces
- I see these commits on xfer/6410b:...
- 07:08 AM Feature #6410: implement additional built-in OO classes/interfaces
- Constantin Asofiei wrote:
> Marian, see another patch - please integrate it also in 6410b.
>
Had to manually appl... - 09:25 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception
- Eduard, mind setting up some large customer applications and do some smoke tests. I will take the changes and do on s...
- 03:30 AM Feature #6819: refactor FWD proxy implementation to use ReflectASM instead of Java Method reflection
- For a while, I've been looking through videos, blog posts and other info I could get my hands on, related to lambdas....
Also available in: Atom