FWD » Core Development » Database

09/14/2019

05:39 PM Bug #4325: Not flushing the newly created record to the database

In the following sequence ... Stanislav Lomany

04:59 PM Bug #4325 (New): Not flushing the newly created record to the database

Stanislav Lomany

11:59 PM Feature #4307: DMO conversion changes

Well done.
As noted originally, however, the DMO implementation classes should _not_ be emitted by default. The de... Eric Faulhaber

09/13/2019

07:35 PM Feature #4323: parse full FQL statements and convert them to SQL

An intermediary status. I added support for gathering data from tables/fields annotations in a cached data structure ... Ovidiu Maxiniuc

09/12/2019

04:26 PM Feature #4323: parse full FQL statements and convert them to SQL

I have investigated how much Hibernate is blended in FWD code: it's a lot: maybe 100 classes. Because of this, I thin... Ovidiu Maxiniuc

09/11/2019

06:44 PM Feature #4307: DMO conversion changes

Now the DMO Impl classes are not generated any more and DMO interfaces are constructed with DMO annotations.
To see ... Ovidiu Maxiniuc

03:08 PM Feature #4307: DMO conversion changes

Ovidiu Maxiniuc wrote:
> At a closer inspection, I see that the @extends DataModelObject@ is not present any more in... Eric Faulhaber

02:25 PM Feature #4307 (WIP): DMO conversion changes

At a closer inspection, I see that the @extends DataModelObject@ is not present any more in the new interface. Is thi... Ovidiu Maxiniuc

04:57 PM Feature #4323: parse full FQL statements and convert them to SQL

Ovidiu Maxiniuc wrote:
> A few questions before starting work:
> * we do not have a FWDQL generator. Should I use t... Eric Faulhaber

04:22 PM Feature #4323: parse full FQL statements and convert them to SQL

A few questions before starting work:
* we do not have a FWDQL generator. Should I use the current conversion (HQL) ... Ovidiu Maxiniuc

02:59 PM Feature #4323: parse full FQL statements and convert them to SQL

As part of the persistence performance overhaul, we are removing Hibernate from FWD. However, we still need an object... Eric Faulhaber

09/10/2019

03:39 PM Feature #4323 (WIP): parse full FQL statements and convert them to SQL

Eric Faulhaber

03:36 PM Feature #4012: replacing TCP/IP (AF_INET) sockets with Unix Domain Sockets (AF_UNIX) as PostgreSQ...

I think #4071 is caused by the use of AF_UNIX sockets, at least with the implementation noted above. This is a best g... Eric Faulhaber

09/09/2019

03:22 AM Feature #4307: DMO conversion changes

As part of our persistence performance overhaul, we are no longer going to emit source code for DMO implementation cl... Eric Faulhaber

03:10 AM Feature #4307 (Closed): DMO conversion changes

Eric Faulhaber

09/04/2019

05:54 PM Feature #3809: ProDataSet support

Ovidiu Maxiniuc wrote:
> You can drop it (all the lines). It was just for debug. I saw it, but you have already star... Constantin Asofiei

05:53 PM Feature #3809: ProDataSet support

Constantin Asofiei wrote:
> Ovidiu, you've added this to @dmo_common.rules@:
> [...]
> This produces lots of noise... Ovidiu Maxiniuc

05:51 PM Feature #3809: ProDataSet support

Ovidiu, you've added this to @dmo_common.rules@:... Constantin Asofiei

09/03/2019

03:42 PM Support #3871: determine how to change codepages/locales during import

>> 1. How do we control the encoding when we read the .d inputs? Do we have work here?
>
> Normally, one should not... Greg Shah

09/02/2019

06:15 PM Feature #3809: ProDataSet support

Constantin Asofiei wrote:
> Ovidiu, another question; for a data-source query like:
> [...]
>
> the 'next-rowid'... Ovidiu Maxiniuc

08/30/2019

11:11 AM Feature #2156: datetime-tz and hibernate

I would go for an unified solution for all dialects, even this duplicating the offset information for H2 and MSSQL: t... Ovidiu Maxiniuc

09:18 AM Feature #2156: datetime-tz and hibernate

It does look like PostgreSQL has a bad design for this use case (storing and honoring the offset):
https://dba.sta... Greg Shah

07:10 AM Feature #2156: datetime-tz and hibernate

I believe the big problem to implement single-column for DATETIME-TZ was the lack of support from H2.
Relative recen... Ovidiu Maxiniuc

08/29/2019

04:29 PM Bug #3379 (Review): do not copy index definitions when defining WORKFILE like TABLE

Vladimir Tsichevski

04:29 PM Bug #3379 (WIP): do not copy index definitions when defining WORKFILE like TABLE

The fix committed to the 4103a branch. The new revno is 11328. Vladimir Tsichevski

03:55 PM Feature #2156: datetime-tz and hibernate

All of our current databases support a single column that stores both the timestamp and offset:
* PostgreSQL has @... Greg Shah

02:20 PM Feature #2156: datetime-tz and hibernate

There's one other thing to add for @datetime-tz@ - its usage in indexes and sort clauses. Being a multi-column Hiber... Constantin Asofiei

08/28/2019

06:09 PM Feature #3809: ProDataSet support

As noted in a previous email, @__error-flag__@, @__origin-rowid__@, @__error-string__@, @__after-rowid__@ and @__row-... Ovidiu Maxiniuc

03:16 PM Bug #3379: do not copy index definitions when defining WORKFILE like TABLE

See the explanation and a proposed fix in #4103. Vladimir Tsichevski

01:53 AM Feature #3816: table and table handle parameter options

Constantin Asofiei wrote:
> An addition here: not only the default-buffer, any explicit buffer defined (non-dynamic)... Marian Edu

08/27/2019

09:17 AM Bug #3379: do not copy index definitions when defining WORKFILE like TABLE

Good, that makes it simpler. Please dig into @addEntries()@ to see how the indexes get copied. The @SchemaDictionar... Greg Shah

07:44 AM Bug #3379: do not copy index definitions when defining WORKFILE like TABLE

Greg Shah wrote:
> The 4GL docs don't show that the @USE-INDEX@ clause is allowed for @DEFINE WORK-TABLE@ but perh... Vladimir Tsichevski

07:29 AM Bug #3379: do not copy index definitions when defining WORKFILE like TABLE

@like_clause@ uses @SymbolResolver.addFieldsFrom()@ to search for the @LIKE table@ and load the fields using @SchemaD... Greg Shah

09:00 AM Feature #3816: table and table handle parameter options

@BIND@ can be used to bind a reference only table in the caller only if the called routine has an @OUTPUT@ parameter ... Marian Edu

08:26 AM Feature #3816: table and table handle parameter options

Marian Edu wrote:
> Constantin, the before-table as well as the default buffer handle (and the before-table's defaul... Constantin Asofiei

07:12 AM Feature #3816: table and table handle parameter options

Constantin, the before-table as well as the default buffer handle (and the before-table's default buffer handle) are ... Marian Edu

08/26/2019

07:22 PM Bug #3379: do not copy index definitions when defining WORKFILE like TABLE

Greg, what do you think? Assuming all indices are to be ignored in the copy as we were told, does the fix belong in t... Eric Faulhaber

01:30 PM Feature #3816: table and table handle parameter options

There's a new issue to consider; if the temp-table has a before-table, when passing this table (or its dataset) aroun... Constantin Asofiei

08/25/2019

05:09 PM Feature #3809: ProDataSet support

Ovidiu, another issue: DATASET:ACCEPT-CHANGES is not working for static datasets, as you are emulating this via the @... Constantin Asofiei

08/23/2019

03:14 PM Feature #3809: ProDataSet support

Found another issue: bind can be bi-directional, from the destination to source, if the source is BY-REFERENCE. I'm w... Constantin Asofiei

09:01 PM Feature #3809: ProDataSet support

Constantin,
Please see revision 11483. Further stabilized. Now the 8.9MB xml you sent me is read successfully and, w... Ovidiu Maxiniuc

08/21/2019

06:35 PM Feature #3809: ProDataSet support

Constantin, please see revision 11479. Further stabilized and with support for indexes and relations.
Ovidiu Maxiniuc

12:03 PM Feature #3809: ProDataSet support

Constantin, please see revision 11477. I stabilized it so it should handle the simple xml as the one you sent to me.
... Ovidiu Maxiniuc

04:15 AM Feature #3809: ProDataSet support

Ovidiu, please stabilize and commit your READ-XML changes; as I mentioned before, the urgent part is reading a single... Constantin Asofiei

04:46 PM Feature #1587: implement full support for word indexes

The following are notes about "word break tables" in the 4GL. The tables are a simple mapping between a character an... Greg Shah

05:21 AM Support #3871: determine how to change codepages/locales during import

Greg Shah wrote:
> An underlying assumption of our approach so far is that the definition of a codepage is an intern... Marian Edu

08/20/2019

05:58 AM Feature #3809: ProDataSet support

Constantin Asofiei wrote:
> Ovidiu, when serializing a dataset to JSON, what is the order of the records on a child ... Ovidiu Maxiniuc

08/19/2019

05:55 PM Feature #3809: ProDataSet support

Ovidiu, when serializing a dataset to JSON, what is the order of the records on a child table (associated with a pare... Constantin Asofiei