FWD » Core Development » Base Language

08/22/2023

08:27 AM Bug #7722: improve performance of equals/not equals when used with character/longchar

*Review of 7722a/rev.14704*
* Use Override annotation for @equals@ in @Text@. Make it return false if the paramete... Alexandru Lungu

06:58 AM Bug #7722: improve performance of equals/not equals when used with character/longchar

*Committed 7722a/rev.14704*. Added Text.equals() method which uses @regionMatches@. The method is called in @CompareO... Dănuț Filimon

05:22 AM Bug #7722: improve performance of equals/not equals when used with character/longchar

Created 7722a. Alexandru Lungu

03:27 AM Bug #7722: improve performance of equals/not equals when used with character/longchar

Constantin Asofiei wrote:
> This only works with intern'ed strings, right?
Right. In FWD-H2 this short-circuit ha... Alexandru Lungu

07:28 AM Bug #5538: fix readByte() for DirStream, ProcessStream, and others

How has it been tested? Greg Shah

08/21/2023

03:13 PM Bug #5538: fix readByte() for DirStream, ProcessStream, and others

Can I get into the queue for merging this branch? Theodoros Theodorou

11:10 AM Bug #7722: improve performance of equals/not equals when used with character/longchar

Alexandru Lungu wrote:
> This makes sense. We have some recent implementation in FWD-H2 that was targeting this exac... Constantin Asofiei

08:46 AM Bug #7722: improve performance of equals/not equals when used with character/longchar

This makes sense. We have some recent implementation in FWD-H2 that was targeting this exact issue. We can even compa... Alexandru Lungu

06:20 AM Bug #7722: improve performance of equals/not equals when used with character/longchar

In a large app, there are ~350k calls of @Text.compareTo@ which originate from equals/not-equals operators (via @Comp... Constantin Asofiei

06:14 AM Bug #7722 (WIP): improve performance of equals/not equals when used with character/longchar

Constantin Asofiei

08/18/2023

05:38 PM Feature #6431: implement support to allow CPINTERNAL set to UTF-8 and CP936

I think this task will drive me crazy.... Theodoros Theodorou

08/17/2023

04:49 PM Feature #6431: implement support to allow CPINTERNAL set to UTF-8 and CP936

... Theodoros Theodorou

07:36 AM Feature #6431: implement support to allow CPINTERNAL set to UTF-8 and CP936

Look at statements like @PUT@, @EXPORT@, @IMPORT@, @INPUT FROM@, @OUTPUT TO@. I/O in different character sets is the... Greg Shah

07:32 AM Feature #6431: implement support to allow CPINTERNAL set to UTF-8 and CP936

Can you please help me by giving me a 4GL code to do I/O in different charsets? (UTF-8 / CP936) Theodoros Theodorou

08/16/2023

09:00 AM Bug #5538: fix readByte() for DirStream, ProcessStream, and others

The branch 5538a has been rebased to trunk with rev no 14697. Theodoros Theodorou

06:20 AM Bug #5538: fix readByte() for DirStream, ProcessStream, and others

Since the indeterminate order of input is an existing issue, we will ignore it here. Greg Shah

06:19 AM Bug #5538: fix readByte() for DirStream, ProcessStream, and others

Anything using @INPUT THROUGH@, @INPUT-OUTPUT THROUGH@ or @OUTPUT THROUGH@ creates a child process and uses that as a... Greg Shah

06:18 AM Bug #5538: fix readByte() for DirStream, ProcessStream, and others

... Greg Shah

08/15/2023

05:16 PM Bug #5538: fix readByte() for DirStream, ProcessStream, and others

It looks like the developer who implemented the @ProcessStream@ was aware of the issue. I found in the javadoc (@Proc... Theodoros Theodorou

08/14/2023

05:41 PM Bug #5538: fix readByte() for DirStream, ProcessStream, and others

In these situations we do our testcases and run them against 4GL to see the actual behaviour. Along the years we enco... Ovidiu Maxiniuc

04:13 PM Bug #5538: fix readByte() for DirStream, ProcessStream, and others

Ovidiu Maxiniuc wrote:
> My point is that, instead of @WhileBlack@ the result would be @WBlackhite@. If I am right, ... Theodoros Theodorou

01:01 PM Bug #5538: fix readByte() for DirStream, ProcessStream, and others

I had some thought about the implementation of this issue over the weekend. The current implementation scans both @ou... Ovidiu Maxiniuc

11:20 AM Bug #5538: fix readByte() for DirStream, ProcessStream, and others

As with most of what we do, it makes most sense to implement 4GL code that can test this. Pure Java unit tests don't... Greg Shah

10:09 AM Bug #5538: fix readByte() for DirStream, ProcessStream, and others

Ovidiu Maxiniuc wrote:
> Review of 5538a/14681.
>
> I agree with the code. Just two notes about code:
> * please... Theodoros Theodorou

03:12 PM Feature #6425: SESSION handle features

Greg Shah wrote:
> Theodoros: Please list those things that are complete and those things (other than the items in #... Theodoros Theodorou

06:26 AM Support #4032: block/transaction processing context-local and other performance improvements

Branch 4032a was merged to trunk rev 14689 and archived.
I've created 4032b from trunk rev 14689 where there will ... Constantin Asofiei

06:12 AM Support #4032: block/transaction processing context-local and other performance improvements

Please merge 4032a to trunk. Greg Shah

08/11/2023

10:56 AM Bug #5538 (Review): fix readByte() for DirStream, ProcessStream, and others

Greg Shah

10:56 AM Bug #5538: fix readByte() for DirStream, ProcessStream, and others

Once these items are addressed, what testing is needed before this can be merged? Greg Shah

10:55 AM Feature #6422: implement auditing support including AUDIT-POLICY and AUDIT-CONTROL

Eric: This needs your code review. Greg Shah

10:55 AM Feature #6420: implement the AUDIT-ENABLED() built-in function

Eric: This needs your code review. Greg Shah

10:53 AM Feature #6425: SESSION handle features

Some changes for this task were written in branch 4938b and were merged into trunk as revision 14688. Runtime suppor... Greg Shah

10:52 AM Feature #4391 (Closed): enhance existing SESSION:STARTUP-PARAMETERS support to honor -H -db -S -l...

The changes for this task were written in branch 4938b and were merged into trunk as revision 14688. Runtime support... Greg Shah

10:51 AM Feature #6497 (Closed): add equivalent support for the -T command line option

The changes for this task were written in branch 4938b and were merged into trunk as revision 14688. Greg Shah

08/07/2023

07:01 PM Bug #5538 (Test): fix readByte() for DirStream, ProcessStream, and others

Review of 5538a/14681.
I agree with the code. Just two notes about code:
* please move comments from lines 462 an... Ovidiu Maxiniuc

05:33 PM Feature #6425 (WIP): SESSION handle features

Theodoros Theodorou

08/04/2023

05:48 PM Bug #5538 (Review): fix readByte() for DirStream, ProcessStream, and others

The implementations of the abstract @Stream.readByte()@ are the following:
ClipboardStream:... Theodoros Theodorou

08:30 PM Bug #5538 (WIP): fix readByte() for DirStream, ProcessStream, and others

Theodoros Theodorou

02:09 PM Bug #7656: investigate performance of context-local helpers (e.g. TransactionManager.TransactionH...

Agreed! It is a really good find. Greg Shah

01:13 PM Bug #7656: investigate performance of context-local helpers (e.g. TransactionManager.TransactionH...

Greg, this looks promising.
The @access@ synthetic methods are a pitfall of using inner classes and assuming a pri... Constantin Asofiei

10:42 AM Bug #7656: investigate performance of context-local helpers (e.g. TransactionManager.TransactionH...

I've done the test. The results are quite good. I tested with 7156b/rev. 14663 and your changes are close to -0.5%. Alexandru Lungu

08/03/2023

11:57 AM Bug #7656: investigate performance of context-local helpers (e.g. TransactionManager.TransactionH...

I've tested with ControlFlowOps, ProcedureManager and TransactionManager. I don't see a definite improvement, the ti... Constantin Asofiei

08:23 AM Bug #7656: investigate performance of context-local helpers (e.g. TransactionManager.TransactionH...

Greg Shah wrote:
> Is it worth it to manually try the changes to TM so we know the benefit?
Yes, I think so. There... Constantin Asofiei

08:19 AM Bug #7656: investigate performance of context-local helpers (e.g. TransactionManager.TransactionH...

Is it worth it to manually try the changes to TM so we know the benefit? Greg Shah

08:13 AM Bug #7656: investigate performance of context-local helpers (e.g. TransactionManager.TransactionH...

Greg Shah wrote:
> Did you test the change for the TM to see what the performance difference is?
Not yet. I'm work... Constantin Asofiei

08:11 AM Bug #7656: investigate performance of context-local helpers (e.g. TransactionManager.TransactionH...

The lesson here is that compiler generated helpers are often costly. Greg Shah

08:10 AM Bug #7656: investigate performance of context-local helpers (e.g. TransactionManager.TransactionH...

Good find. This seems like it should have good potential.
Did you test the change for the TM to see what the perf... Greg Shah

08:10 AM Bug #7656: investigate performance of context-local helpers (e.g. TransactionManager.TransactionH...

The same happens when accessing a private field: access to a private field from a non-defining class will be done via... Constantin Asofiei

06:38 AM Bug #7656: investigate performance of context-local helpers (e.g. TransactionManager.TransactionH...

In the profiler, I noticed that for i.e. @TransactionManager$TransactionHelper.deregisterOutputParameterAssigner@, th... Constantin Asofiei

06:22 AM Bug #7656 (New): investigate performance of context-local helpers (e.g. TransactionManager.Transa...

Constantin Asofiei

11:17 AM Feature #6493: add equivalent support for the command line parameters -ucc and -utf8intatchar

If this support requires "wide character" support in ncurses, then we may need some changes to ensure our support mat... Greg Shah

10:10 AM Feature #6431: implement support to allow CPINTERNAL set to UTF-8 and CP936

> As I understand, we are trying to emulate a keyboard based on the -cpinternal?
Not exactly. In the 4GL, the @C... Greg Shah

08/02/2023

05:41 PM Feature #6431 (WIP): implement support to allow CPINTERNAL set to UTF-8 and CP936

Theodoros Theodorou

05:36 PM Feature #6431: implement support to allow CPINTERNAL set to UTF-8 and CP936

I would like to ask for some guidance for this task. As I understand, we are trying to emulate a keyboard based on th... Theodoros Theodorou

04:00 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception

Alexandru Lungu wrote:
> > There is another approach: in a FOR EACH block, is the navigation always done via query.n... Constantin Asofiei

03:55 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception

> There is another approach: in a FOR EACH block, is the navigation always done via query.next()? If yes, we can:
> ... Alexandru Lungu

08/01/2023

10:07 AM Support #4032: block/transaction processing context-local and other performance improvements

Code Review Task Branch 4032a Revisions 14680 through 14682
The changes look good. They also seem fairly low risk. Greg Shah

08:05 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception

> Now that such iterating statement is removed, the block is unaware of the iteration type
I think I understand yo... Greg Shah

07:46 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception

*Created 7045b.*
Committed conversion changes to 7045b/rev. 14678. This is still a prototype as there is no run-ti... Alexandru Lungu

07/31/2023

01:32 PM Feature #6410: implement additional built-in OO classes/interfaces

OK, then let's get this patch into trunk using a different branch. Greg Shah

01:10 PM Feature #6410: implement additional built-in OO classes/interfaces

Greg Shah wrote:
> Do you think we are ready to merge 6410a to trunk?
No. @Method:invoke@ and @Class:getMethod(s)@... Constantin Asofiei

12:47 PM Feature #6410: implement additional built-in OO classes/interfaces

Do you think we are ready to merge 6410a to trunk? Greg Shah

11:30 AM Feature #6410: implement additional built-in OO classes/interfaces

Greg, I have this patch for an issue with @Progress.Lang.Class:invoke@ - we need to re-throw the structured exception... Constantin Asofiei

08:41 AM Bug #7643 (Test): memory leak in AssociatedThread

Greg Shah

08:29 AM Bug #7643: memory leak in AssociatedThread

Branch 7643a was merged to trunk rev 14677 and archived. Constantin Asofiei

08:27 AM Bug #7643: memory leak in AssociatedThread

Code Review Task Branch 7643a Revision 14677
It looks good.
You can merge to trunk. Greg Shah

08:24 AM Bug #7643: memory leak in AssociatedThread

Fixed in 7643a rev 14677. Constantin Asofiei

08:18 AM Bug #7643: memory leak in AssociatedThread

Agreed. Greg Shah

08:17 AM Bug #7643: memory leak in AssociatedThread

I'm thinking of keeping this new ThreadLocal field set only once; previously, for a thread, the value got changed eac... Constantin Asofiei

07:41 AM Bug #7643: memory leak in AssociatedThread

Constantin Asofiei wrote:
> Greg Shah wrote:
> > Code Review Task Branch 7643a Revision 14676
> >
> > Shouldn't ... Greg Shah

07:36 AM Bug #7643: memory leak in AssociatedThread

Greg Shah wrote:
> Code Review Task Branch 7643a Revision 14676
>
> Shouldn't we just change @endContext()@ to re... Constantin Asofiei

07:23 AM Bug #7643: memory leak in AssociatedThread

Code Review Task Branch 7643a Revision 14676
Shouldn't we just change @endContext()@ to remove the thread id as pa... Greg Shah

05:56 AM Bug #7643 (Review): memory leak in AssociatedThread

Created task branch 7643a from trunk rev 14675.
The fix is in 7643a rev 14676. Constantin Asofiei

08:41 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception

> Constantin's solution looks clean, so I will start working around it.
Agreed. Go ahead. Greg Shah

08:35 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception

Greg Shah wrote:
> > introduce a new lambda block between @Init@ and @Body@, namely @Iter@. This will be a boolean s... Alexandru Lungu

08:01 AM Bug #7644 (Test): load all legacy enums on FWD server startup

Greg Shah

07:57 AM Bug #7644: load all legacy enums on FWD server startup

Branch 7644a was merged to trunk rev 14676 and archived. Constantin Asofiei

07:41 AM Bug #7644: load all legacy enums on FWD server startup

If you want to merge 7644a to trunk, you can go ahead. Greg Shah

07:37 AM Bug #7644: load all legacy enums on FWD server startup

Greg Shah wrote:
> Code Review Task Brnach 7644a Revision 14676
>
> I don't have an objection, however I do wonde... Constantin Asofiei

07:32 AM Bug #7644: load all legacy enums on FWD server startup

Code Review Task Brnach 7644a Revision 14676
I don't have an objection, however I do wonder if there is a performa... Greg Shah

06:20 AM Bug #7644 (Review): load all legacy enums on FWD server startup

Created task branch 7644a from trunk rev 14675.
The fix is in rev 14676 - the @public static final@ field definiti... Constantin Asofiei

07/28/2023

04:27 PM Bug #7644: load all legacy enums on FWD server startup

The legacy enum values are defined as Java 'static final' - and created in a static constructor. The problem here is... Constantin Asofiei

04:21 PM Bug #7644 (Test): load all legacy enums on FWD server startup

Constantin Asofiei

03:09 PM Bug #7643: memory leak in AssociatedThread

In a large GUI application, in a thread dump @SecurityManager.threadIds@ is reported to have a size of 120k (for refe... Constantin Asofiei

02:58 PM Bug #7643 (Test): memory leak in AssociatedThread

Constantin Asofiei

11:09 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception

Constantin Asofiei wrote:
> There is another approach: in a @FOR EACH@ block, is the navigation always done via @que... Alexandru Lungu

09:55 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception

@FOR FIRST@ and @FOR LAST@ get converted to @forBlock@. I think it should work by injecting the @P2JQuery@ instance ... Constantin Asofiei

09:48 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception

There is another approach: in a @FOR EACH@ block, is the navigation always done via @query.next()@? If yes, we can:
... Constantin Asofiei

09:37 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception

> introduce a new lambda block between @Init@ and @Body@, namely @Iter@. This will be a boolean supplier. If this boo... Greg Shah

08:59 AM Feature #7045: re-implement "normal" (non-abend) stack unwinding to avoid throwing an exception

I analyzed the problem and understand most of the issue. I also have designed a solution here. However, I still have ... Alexandru Lungu

07/27/2023

10:56 AM Bug #7604 (Test): Switch does not honor case sensitivity of character

Greg Shah

10:54 AM Bug #7604 (Review): Switch does not honor case sensitivity of character

Merged 7464a in trunk as rev. 14673. Archived 7464a. Alexandru Lungu

07/25/2023

01:04 PM Bug #7460: implement Progress.Lang.Class:getMethods

Did pushed local changes in #6410a - rev 14556, probably missed release notes.
The idea was to directly use the Ja... Marian Edu

07:41 AM Feature #7045 (WIP): re-implement "normal" (non-abend) stack unwinding to avoid throwing an excep...

Greg Shah