FWD » Core Development » Database

05/28/2019

10:15 AM Feature #3815: more temp-table options

Stanislav Lomany wrote:
> I've found that the @CODEPAGE@ can be inherited from another permanent or temp-table throu... Eric Faulhaber

05/27/2019

06:01 PM Feature #3815: more temp-table options

I've found that the @CODEPAGE@ can be inherited from another permanent or temp-table through either table- or field-l... Stanislav Lomany

05/24/2019

06:27 PM Feature #3809: ProDataSet support

Ovidiu Maxiniuc wrote:
> As an intermediary conclusion, I believe they are some kind of dynamic buffers. Not sure ho... Constantin Asofiei

06:24 PM Feature #3809: ProDataSet support

I am struggling to make before-buffer conversion work correctly.
Apparently FWD converts correct these tables/buffer... Ovidiu Maxiniuc

01:25 PM Feature #3813: misc DB features part deux

@BUFFER-VALIDATE@ method is required for the current large app work. Constantin Asofiei

05/23/2019

05:35 PM Feature #3815: more temp-table options

> The only good reason is that they are hidden.
Sorry for bothering you, I wrote a code in @DumpTree@ which prints... Stanislav Lomany

12:53 PM Feature #3815: more temp-table options

> Do you have an idea why the middle part is not iterated? Nodes are not hidden.
The only good reason is that they... Greg Shah

11:58 AM Feature #3815: more temp-table options

Guys, I need some technical help. I was going to emit a function that sets the codepage in @convert/buffer_definition... Stanislav Lomany

05/22/2019

10:22 AM Feature #3809: ProDataSet support

> For the first two parameters they are parented to the passed object. In the last two cases, we find them parented t... Greg Shah

10:17 AM Feature #3815: more temp-table options

> I don't think we should support this at this point.
The problem here is that if we don't support this now, we wi... Greg Shah

05/21/2019

03:55 PM Feature #3809: ProDataSet support

The problem is not just RUN statement: for function/method calls, there is no PARAMETER node - so every option (APPEN... Constantin Asofiei

03:49 PM Feature #3809: ProDataSet support

The problem with parameter options is that FWD uses a mixed tree locations. Here is an example:... Ovidiu Maxiniuc

02:26 PM Feature #3815: more temp-table options

It is interesting that when @COLUMN-CODEPAGE@ is calculated _initial_ value of the parameter is used instead of the p... Stanislav Lomany

05/20/2019

06:03 PM Feature #3815: more temp-table options

> Please check to confirm that the expression is only evaluated once
It is evaluated once. Stanislav Lomany

02:33 PM Feature #3815: more temp-table options

> Ok. When I wrote this I thought these changes were needed for the same projects as 3751a.
It is but for differen... Greg Shah

02:32 PM Feature #3815: more temp-table options

Stanislav Lomany wrote:
> Guys, do we need to support arbitrary expressions as a @COLUMN-CODEPAGE@? As far as I get,... Greg Shah

02:29 PM Feature #3815: more temp-table options

Constantin Asofiei wrote:
> Eric Faulhaber wrote:
> > Stanislav Lomany wrote:
> > > Into what branch should I comm... Eric Faulhaber

02:25 PM Feature #3815: more temp-table options

Eric Faulhaber wrote:
> Stanislav Lomany wrote:
> > Into what branch should I commit?
> Probably 3751a. Constantin... Constantin Asofiei

02:21 PM Feature #3815: more temp-table options

Stanislav Lomany wrote:
> In this case we'll have to pass the expression to some function at the beginning of the ex... Constantin Asofiei

02:20 PM Feature #3815: more temp-table options

Guys, do we need to support arbitrary expressions as a @COLUMN-CODEPAGE@? As far as I get, it can be combination of:
... Stanislav Lomany

05/18/2019

12:49 PM Feature #3809: ProDataSet support

Constantin Asofiei wrote:
> There are issues with by-reference, append and bind options for caller's DATASET-HANDLE,... Ovidiu Maxiniuc

10:00 AM Feature #3815: more temp-table options

> Beyond the immediate scope of your changes
I'm not sure about this. I don't know when we would do this otherwise. Greg Shah

01:29 AM Feature #3815: more temp-table options

Stanislav Lomany wrote:
> Into what branch should I commit?
Probably 3751a. Constantin, is that ok with you?
> A... Eric Faulhaber

05/17/2019

03:19 PM Feature #3809: ProDataSet support

There are issues with by-reference, append and bind options for caller's DATASET-HANDLE, DATASET, TABLE-HANDLE, TABLE... Constantin Asofiei

09:12 PM Feature #3815: more temp-table options

Into what branch should I commit?
Am I right that there is no runtime part for @COLUMN-CODEPAGE@? Stanislav Lomany

05/16/2019

09:22 AM Feature #3816: table and table handle parameter options

Status update on 4GL testcases development:
> Subject: Feature #3816 - table and table handle parameter options
>... Greg Shah

05/15/2019

09:05 AM Support #4080: AWS Aurora support

A customer has asked this question:
>We want to move our customers to AWS. When moving to AWS it would be great to... Greg Shah

08:52 AM Support #4080 (New): AWS Aurora support

Greg Shah

05/14/2019

05:34 PM Feature #3813: misc DB features part deux

@DEFAULT-STRING@ conversion and runtime stub was added in 3751a rev 11322 Constantin Asofiei

04:19 PM Feature #3809: ProDataSet support

I committed the conversion support for you as r11321 in 3751a. Ovidiu Maxiniuc

04:07 PM Feature #3809: ProDataSet support

Constantin Asofiei wrote:
> I need compiler constant support for the ROW-STATE possible values... planing to add the... Ovidiu Maxiniuc

04:02 PM Feature #3809: ProDataSet support

I need compiler constant support for the ROW-STATE possible values... planing to add them now to 3751a. Constantin Asofiei

01:53 PM Feature #3809: ProDataSet support

I'm posting this here, but we may need to work it in a different branch; when calling a TABLE argument with the BY-R... Constantin Asofiei

04:10 PM Feature #3815: more temp-table options

Constantin Asofiei wrote:
> What is the state of BEFORE-TABLE option at conversion time? I need this for my current... Ovidiu Maxiniuc

10:42 AM Feature #3815: more temp-table options

What is the state of BEFORE-TABLE option at conversion time? I need this for my current work... Constantin Asofiei

05/13/2019

05:31 PM Feature #3815 (WIP): more temp-table options

@WRITE-XML/JSON@ as well as buffer @DISPLAY/UPDATE@ use order by @ORDER@.
I couldn't make @BUFFER-COPY@ and @BUFFER-... Stanislav Lomany

12:42 PM Feature #3809: ProDataSet support

Please update 3751a to r11316.
I could not test on Dataset project because of some other issue.
Ovidiu Maxiniuc

11:13 AM Feature #3809: ProDataSet support

I will add this in a couple of hours. Ovidiu Maxiniuc

11:11 AM Feature #3809: ProDataSet support

Ovidiu, I need some help; this is valid code (both runtime and compile) in OpenEdge 11.7:... Constantin Asofiei

05/10/2019

12:01 PM Feature #3809: ProDataSet support

Related to previous discussion about @BEFORE-TABLE@ in #3815, entries 7 to 10.
We had a problem with the attribute ... Ovidiu Maxiniuc

11:38 AM Feature #3815: more temp-table options

Greg Shah wrote:
> Hmm. I suspect the order of processing inside @BUFFER-COPY@ and @BUFFER-COMPARE@ may also depend... Eric Faulhaber

11:07 AM Feature #3815: more temp-table options

Hmm. I suspect the order of processing inside @BUFFER-COPY@ and @BUFFER-COMPARE@ may also depend on @ORDER@ or @POSI... Greg Shah

11:04 AM Feature #3815: more temp-table options

Stanislav Lomany wrote:
> Do you know other function than @RAW-TRANSFER@ and @BUFFER-FIELD@ that depend on fields or... Ovidiu Maxiniuc

10:43 AM Feature #3815: more temp-table options

I don't know the answer. Ovidiu and/or Eric may have some ideas there.
I think searching the 4GL docs and knowled... Greg Shah

10:33 AM Feature #3815: more temp-table options

Do you know other function than @RAW-TRANSFER@ and @BUFFER-FIELD@ that depend on fields order? Stanislav Lomany

05/09/2019

07:32 PM Feature #3815: more temp-table options

If we need this information for runtime purposes, then we do need to add this to annotations. Greg Shah

05:09 PM Feature #3815: more temp-table options

OK, we have @fieldId@, @order@ and @position@ fields, and the last two are optional. ... Stanislav Lomany

10:42 AM Feature #3815: more temp-table options

This seems to be a "feature" of the @BUFFER-FIELD()@ method.
Please see this article:
https://knowledgebase.pro... Greg Shah

09:09 AM Feature #3815: more temp-table options

As far as I tested, @LIKE@ and @LIKE-SEQUENTIAL@ behave in the same way (like @LIKE-SEQUENTIAL@). E.g. we have table ... Stanislav Lomany

03:29 AM Feature #3815: more temp-table options

Yes, the @BEFORE-TABLE@ option is working for me. I created a small testcase:... Ovidiu Maxiniuc

04:57 PM Bug #4071: FWD server stops allowing new web client logins after abend

Greg Shah wrote:
> > > Otherwise we probably need to ensure any rollback is executed on the conversation thread.
> ... Constantin Asofiei

04:55 PM Bug #4071: FWD server stops allowing new web client logins after abend

> > Otherwise we probably need to ensure any rollback is executed on the conversation thread.
>
> Agreed. I didn't... Greg Shah

10:46 AM Feature #3809: ProDataSet support

I do agree that using "HANDLE" seems incorrect for these cases. Aren't these specific types which cannot be intermix... Greg Shah

09:19 AM Feature #3809: ProDataSet support

Yes, I though the annotation is incorrect. The type attribute is @"HANDLE"@, too, in @name-map@. I understand they ha... Ovidiu Maxiniuc

08:11 AM Feature #3809: ProDataSet support

Ovidiu Maxiniuc wrote:
> Constantin, I have the following definitions in a object:
> [...]
> They are converted in... Constantin Asofiei

08:10 AM Feature #3809: ProDataSet support

Constantin, I have the following definitions in a object:... Ovidiu Maxiniuc

04:39 AM Feature #3809: ProDataSet support

Please check out the 3751a/11308. Ovidiu Maxiniuc

05/08/2019

06:13 PM Feature #3809: ProDataSet support

Constantin Asofiei wrote:
> Ovidiu, @dataset-handle@ parameters need to convert as @DatasetParameter@ (same as @tabl... Constantin Asofiei

04:06 PM Feature #3815: more temp-table options

Stanislav: Don't do anything on @BEFORE-TABLE@. It will be handled in #3809.
> However, it seems to be already wo... Greg Shah

03:48 PM Feature #3815: more temp-table options

Eric Faulhaber wrote:
> Ovidiu, please see previous posts. Is conversion of BEFORE-TABLE already handled as part of ... Ovidiu Maxiniuc

03:28 PM Feature #3815: more temp-table options

Ovidiu, please see previous posts. Is conversion of BEFORE-TABLE already handled as part of #3809? Eric Faulhaber

02:42 PM Feature #3815: more temp-table options

Do I need to work on @BEFORE-TABLE@ as well? Stanislav Lomany

02:51 PM Feature #4056: instrument FWD and supporting libraries to measure database performance

Eric Faulhaber wrote:
> Hynek Cihlar wrote:
> > How about to exclude the FWD classes in aop.xml? So that the FWD cl... Hynek Cihlar

02:30 PM Feature #4056: instrument FWD and supporting libraries to measure database performance

Hynek Cihlar wrote:
> How about to exclude the FWD classes in aop.xml? So that the FWD classes are only weaved by th... Eric Faulhaber

02:01 PM Feature #4056: instrument FWD and supporting libraries to measure database performance

Eric Faulhaber wrote:
> Hynek Cihlar wrote:
> > I'm not sure how well you can combine load-time and compile-time we... Hynek Cihlar

01:56 PM Feature #4056: instrument FWD and supporting libraries to measure database performance

Hynek Cihlar wrote:
> I'm not sure how well you can combine load-time and compile-time weaving.
Actually, I did not... Eric Faulhaber

01:40 PM Feature #4056: instrument FWD and supporting libraries to measure database performance

Eric Faulhaber wrote:
> Thank you for the guidance.
>
> I am using compile-time weaving only because that seemed ... Hynek Cihlar

01:32 PM Feature #4056: instrument FWD and supporting libraries to measure database performance

Thank you for the guidance.
I am using compile-time weaving only because that seemed to be the most direct way to ... Eric Faulhaber

01:14 PM Feature #4056: instrument FWD and supporting libraries to measure database performance

Here is more info on load time weaving http://www.eclipse.org/aspectj/doc/released/devguide/ltw.html. Hynek Cihlar

01:12 PM Feature #4056: instrument FWD and supporting libraries to measure database performance

Hynek Cihlar wrote:
> Eric Faulhaber wrote:
> > Hynek, I am using AspectJ to capture method-level profile data (jus... Hynek Cihlar

01:07 PM Feature #4056: instrument FWD and supporting libraries to measure database performance

Eric Faulhaber wrote:
> Hynek, I am using AspectJ to capture method-level profile data (just number of invocations a... Hynek Cihlar

12:51 PM Feature #4056 (WIP): instrument FWD and supporting libraries to measure database performance

Hynek, I am using AspectJ to capture method-level profile data (just number of invocations and aggregate elapsed time... Eric Faulhaber

05/06/2019

03:18 PM Bug #4071: FWD server stops allowing new web client logins after abend

Greg Shah wrote:
> Otherwise we probably need to ensure any rollback is executed on the conversation thread.
Agre... Eric Faulhaber

03:14 PM Bug #4071: FWD server stops allowing new web client logins after abend

I'm not saying you are incorrect. It may be a new "feature" of the UNIX socket code. But I think it is an invalid a... Greg Shah

02:29 PM Bug #4071: FWD server stops allowing new web client logins after abend

The reason I assumed it was something in the new socket code is that we've had many abnormal ends in the past, where ... Eric Faulhaber

02:26 PM Bug #4071: FWD server stops allowing new web client logins after abend

This is persistence context cleanup code. At the persistence level, we don't have any control as to the thread on whi... Eric Faulhaber

02:13 PM Bug #4071: FWD server stops allowing new web client logins after abend

> Possibly just a bug in the org.newsclub.net socket code?
Yes, it is possible.
*However*, this database proces... Greg Shah

04:42 AM Bug #4071: FWD server stops allowing new web client logins after abend

Eric Faulhaber wrote:
> Is the hung Reader thread preventing port 7449 from being released, because context cleanup ... Constantin Asofiei

12:54 AM Bug #4071: FWD server stops allowing new web client logins after abend

I've recently noticed that after some abnormal ends, I can no longer log into the web client without restarting the F... Eric Faulhaber

12:32 AM Bug #4071 (New): FWD server stops allowing new web client logins after abend

Eric Faulhaber

06:22 AM Feature #3816: table and table handle parameter options

Yes, this one is more urgent since we are about to start work on it. Greg Shah

01:59 AM Feature #3816: table and table handle parameter options

Greg Shah wrote:
> Please document the specific questions we have that will require 4GL testcases. I don't know if ... Marian Edu

05/02/2019

10:24 AM Feature #3809: ProDataSet support

Ovidiu, @dataset-handle@ parameters need to convert as @DatasetParameter@ (same as @table-handle@ does). Currently t... Constantin Asofiei

05/01/2019

04:58 PM Feature #3761 (Closed): add DYNAMIC-FUNCTION() support to the dynamic database WHERE clause inter...

As of trunk revision 11302, this functionality is complete. Greg Shah

02:14 PM Support #4063: linking looping FIND FIRST/NEXT and FIND LAST/PREV in a set oriented operation

After speaking with Eric, some thoughts:
* The current approach is handled exclusively at runtime.
* The more adv... Greg Shah

02:04 PM Support #4063: linking looping FIND FIRST/NEXT and FIND LAST/PREV in a set oriented operation

From Eric:
> Yes. This has been implemented since 2006 as @AdaptiveFind@.
>
> It doesn't detect that it is in ... Greg Shah

02:02 PM Support #4063: linking looping FIND FIRST/NEXT and FIND LAST/PREV in a set oriented operation

From Constantin:
> An idea here is to add an UUID at conversion time, to a FIND. This way the runtime can uniquely... Greg Shah

02:01 PM Support #4063: linking looping FIND FIRST/NEXT and FIND LAST/PREV in a set oriented operation

Can @FIND FIRST@ and @FIND NEXT@ with same buffer and same @WHERE@ be matched together into a set oriented approach? ... Greg Shah

02:00 PM Support #4063 (New): linking looping FIND FIRST/NEXT and FIND LAST/PREV in a set oriented operation

Greg Shah

01:58 PM Feature #4033: avoid query for FIND unique when the current record in the buffer would match the ...

I think that in addition to @FIND unique@, there are other cases (@FIND FIRST/LAST@) where we can track enough state ... Greg Shah

01:17 PM Support #4060: investigate converting extent fields to array columns

Note that this task is created in the context of investigating changes that may improve performance, but I don't know... Eric Faulhaber

01:08 PM Support #4060: investigate converting extent fields to array columns

We have found on occasion that our default approach of normalizing extent fields of like extent into a secondary tabl... Eric Faulhaber

01:01 PM Support #4060 (WIP): investigate converting extent fields to array columns

Eric Faulhaber

01:00 PM Support #4058: consider denormalizing tables as the default approach

I don't really want to make denormalization the default, as this can result in a huge number of columns in a table, h... Eric Faulhaber

10:49 AM Support #4058: consider denormalizing tables as the default approach

We know that denormalizing for all tables is not always a performance improvement, so there may need to be a better a... Greg Shah

10:48 AM Support #4058 (New): consider denormalizing tables as the default approach

Greg Shah

10:46 AM Bug #4057: h2 performance degradation

From Eric:
> I'd rather not. 197 has a lot of fixes for regressions in the last few point releases before it.
>
... Greg Shah

10:46 AM Bug #4057: h2 performance degradation

As an interim measure, can we go back to 1.4.196 (and add our customizations)? Greg Shah

10:45 AM Bug #4057: h2 performance degradation

More from the mailing list:
-------- Forwarded Message --------
Subject: [h2] Re: Recent drops in performance
D... Greg Shah

10:44 AM Bug #4057: h2 performance degradation

From Eric:
> We're on (a customized) 1.4.197 (not depicted, but probably somewhere between cyan and red). Greg Shah

10:43 AM Bug #4057: h2 performance degradation

From the H2 mailing list:
-------- Forwarded Message --------
Subject: [h2] Re: Recent drops in performance
Dat... Greg Shah

10:42 AM Bug #4057 (New): h2 performance degradation

Greg Shah

02:21 AM Feature #4056: instrument FWD and supporting libraries to measure database performance

In order to improve the performance of the persistence layer of FWD, we first need to measure the time spent at sever... Eric Faulhaber

02:11 AM Feature #4056 (WIP): instrument FWD and supporting libraries to measure database performance

Eric Faulhaber

02:03 AM Feature #4055: optimize temp-table output parameter copying

This idea warrants further investigation. Eric Faulhaber

02:02 AM Feature #4055: optimize temp-table output parameter copying

Eric wrote:
---
Constantin Asofiei wrote:
> Eric Faulhaber wrote:
> > We could get past the first issue with ... Eric Faulhaber

02:01 AM Feature #4055: optimize temp-table output parameter copying

Constantin wrote:
---
Eric Faulhaber wrote:
> We could get past the first issue with some clever insert statem... Eric Faulhaber

02:01 AM Feature #4055: optimize temp-table output parameter copying

Eric wrote:
---
Constantin Asofiei wrote:
> Eric, did you notice the time spent during temp-table output param... Eric Faulhaber

02:00 AM Feature #4055: optimize temp-table output parameter copying

Constantin wrote:
---
Eric, did you notice the time spent during temp-table output parameter copying, during pr... Eric Faulhaber

01:59 AM Feature #4055: optimize temp-table output parameter copying

The following discussion between Constantin and myself is copied from another task... Eric Faulhaber

01:58 AM Feature #4055 (WIP): optimize temp-table output parameter copying

Eric Faulhaber

01:32 AM Feature #4054: only track specific tables in dirty database

The idea is to improve performance of the persistence layer by only tracking certain tables within the dirty database... Eric Faulhaber

01:28 AM Feature #4054 (New): only track specific tables in dirty database

Eric Faulhaber