FWD » Core Development » Database

10/30/2019

10:39 AM Support #4376 (New): test replication trigger implementation

Greg Shah

10:38 AM Feature #4375: add support for -pf <profile_filename> option for CONNECT statement

The @CONNECT@ statement can have a @.pf@ file specified as an option. At runtime, this must be read from the filesys... Greg Shah

10:34 AM Feature #4375 (Closed): add support for -pf <profile_filename> option for CONNECT statement

Greg Shah

10/25/2019

01:22 PM Feature #4369: implement stateless FWD server clustering

> For me, running in a stateless mode means that the clients connect and execute independent (quick) request on one o... Greg Shah

01:19 PM Feature #4369: implement stateless FWD server clustering

> I think the in-server appserver agents would be a good addition to this (I know we plan to add this).
I think s... Greg Shah

01:02 PM Feature #4369: implement stateless FWD server clustering

Ovidiu Maxiniuc wrote:
> How about the persistent procedures?
> As I see this, they are allowed as long they are u... Constantin Asofiei

12:44 PM Feature #4369: implement stateless FWD server clustering

Constantin Asofiei wrote:
> An alternative for the meta tables would be to store it in a physical database, to which... Ovidiu Maxiniuc

12:30 PM Feature #4369: implement stateless FWD server clustering

First, pleat let me know if I got this tight.
For me, running in a stateless mode means that the clients connect and... Ovidiu Maxiniuc

10/24/2019

04:08 PM Feature #4369: implement stateless FWD server clustering

Most (if not all) of the non-persistence runtime is context-local. One part I see that may require improvements (if ... Constantin Asofiei

03:50 PM Feature #4369: implement stateless FWD server clustering

We also should consider the potential duplication of dynamic proxies and other dynamically generated interfaces and c... Eric Faulhaber

03:20 PM Feature #4369: implement stateless FWD server clustering

Constantin/Ovidiu: Can you think of any other persistence or even non-persistence state which we need to share across... Greg Shah

03:19 PM Feature #4369: implement stateless FWD server clustering

At this time I am assuming the Admin Console is not changing, but we must at least make it work with the directory ap... Greg Shah

03:17 PM Feature #4369: implement stateless FWD server clustering

One thing to consider is that we don't want these server instances to be hard to configure. At one time we considere... Greg Shah

03:14 PM Feature #4369: implement stateless FWD server clustering

For data structures which cannot be backed by the database AND which must be shared across JVM instances in the FWD c... Greg Shah

03:09 PM Feature #4369: implement stateless FWD server clustering

After a discussion with Eric, we have this list of items that are persistence related.
* dirty share database
** ... Greg Shah

02:53 PM Feature #4369: implement stateless FWD server clustering

The core idea here is that if customers have written their application to be stateless in the 4GL, then today they ca... Greg Shah

02:03 PM Feature #4369 (Test): implement stateless FWD server clustering

Greg Shah

03:06 PM Feature #1879: push lock manager back into database

Based on a discussion with Eric, he notes that the primary differences in our in-memory implementation versus the Pos... Greg Shah

10:19 AM Feature #4323: parse full FQL statements and convert them to SQL

Eric Faulhaber wrote:
> The one thing I really don't like, however, is the presence of casts to @TemporaryBuffer@ wi... Ovidiu Maxiniuc

10/23/2019

04:17 PM Feature #4323: parse full FQL statements and convert them to SQL

The one thing I really don't like, however, is the presence of casts to @TemporaryBuffer@ within @RecordBuffer@ metho... Eric Faulhaber

12:40 PM Feature #4323: parse full FQL statements and convert them to SQL

Ovidiu Maxiniuc wrote:
> The problem is @RecordBuffer@, where some methods like @copy()@, which, in order to minimiz... Eric Faulhaber

08:47 AM Feature #4323: parse full FQL statements and convert them to SQL

Eric Faulhaber wrote:
> Ovidiu, in your last commit to 4011a, you removed all the default methods from @Record@ whic... Ovidiu Maxiniuc

12:19 PM Feature #1587: implement full support for word indexes

Some useful references:
* "ElasticSearch":https://en.wikipedia.org/wiki/Elasticsearch - this is the most popular f... Greg Shah

05:57 AM Support #3871: determine how to change codepages/locales during import

I used the similar diff to the following one in order to setup the UTF-8 database with INPUT_ENCODING 1252 (WINDOWS-1... Sergey Ivanovskiy

10/22/2019

07:59 PM Feature #4323: parse full FQL statements and convert them to SQL

Eric Faulhaber wrote:
> ... I'd like to change it back.
FYI, I've done this in my local working copy. So in case ... Eric Faulhaber

07:20 PM Feature #4323: parse full FQL statements and convert them to SQL

Ovidiu, in your last commit to 4011a, you removed all the default methods from @Record@ which were based on your earl... Eric Faulhaber

11:35 AM Support #3871: determine how to change codepages/locales during import

Sergey Ivanovskiy wrote:
> I found that 4GL supports aliases for charsets that are not supported by @java.nio.charse... Greg Shah

11:33 AM Support #3871: determine how to change codepages/locales during import

> No, it is not the case that was described. I considered the case when DATABASE_ENCODING is not the same as INPUT_EN... Greg Shah

11:10 AM Support #3871: determine how to change codepages/locales during import

I found that 4GL supports aliases for charsets that are not supported by @java.nio.charset.Charset.forName@. For an e... Sergey Ivanovskiy

11:07 AM Support #3871: determine how to change codepages/locales during import

The part of this issue is related to support @DBCODEPAGE@ function because answering this question gets the runtime i... Sergey Ivanovskiy

02:21 AM Feature #2156: datetime-tz and hibernate

Ovidiu Maxiniuc wrote:
> What was the case you encountered with the customer code? TEMP-TABLE or permanent database?... Constantin Asofiei

10/21/2019

06:19 PM Support #3871: determine how to change codepages/locales during import

Greg Shah wrote:
> Sergey Ivanovskiy wrote:
> > At this moment I suppose now that PostgreSQL database has been conf... Sergey Ivanovskiy

02:48 PM Support #3871: determine how to change codepages/locales during import

Sergey Ivanovskiy wrote:
> At this moment I suppose now that PostgreSQL database has been configured to have @UTF-8@... Greg Shah

06:11 PM Feature #2156: datetime-tz and hibernate

Constantin Asofiei wrote:
> BTW, when you say ISO strings, you mean what is in the WHERE clauses?
I was referring... Eric Faulhaber

05:23 PM Feature #2156: datetime-tz and hibernate

> I am not sure I understand what you mean. Each DTZ has its own timezone. You can obtain it with @getOffset()@ and c... Stanislav Lomany

05:06 PM Feature #2156: datetime-tz and hibernate

Constantin,
I inspected how DTZ are converted (with slightly modified FWD code, to aviod some NPEs):
* I see the @d... Ovidiu Maxiniuc

03:15 PM Feature #2156: datetime-tz and hibernate

Eric Faulhaber wrote:
> Constantin, aside from the broken support in indices and sorting, have you noticed any perfo... Constantin Asofiei

02:41 PM Feature #2156: datetime-tz and hibernate

Constantin, aside from the broken support in indices and sorting, have you noticed any performance problems related t... Eric Faulhaber

02:37 PM Feature #2156: datetime-tz and hibernate

In my point of view, the high-priority outstanding issue is using a datetime-tz in indexes and BY clauses. FWD is br... Constantin Asofiei

02:32 PM Feature #2156: datetime-tz and hibernate

For the current implementation is in @DatetimeTzUserType@.
However, I would do a step back and see where we stand.... Ovidiu Maxiniuc

02:23 PM Feature #2156: datetime-tz and hibernate

> my understanding is that you were suggesting we use the timestamp without time zone type (even for databases which ... Greg Shah

02:14 PM Feature #2156: datetime-tz and hibernate

Stanislav Lomany wrote:
> Eric, if I understand it correctly, timestamp with timezone and normal timestamp are both ... Eric Faulhaber

01:39 PM Feature #2156: datetime-tz and hibernate

I am not sure I understand what you mean. Each DTZ has its own timezone. You can obtain it with @getOffset()@ and con... Ovidiu Maxiniuc

11:43 AM Feature #2156: datetime-tz and hibernate

Are we going to get the current time zone via @TimeZone.getDefault()@ (H2) or @show timezone;@ (PostgreSQL) which wou... Stanislav Lomany

11:22 AM Feature #2156: datetime-tz and hibernate

Eric, if I understand it correctly, timestamp with timezone and normal timestamp are both mapped to @org.hibernate.ty... Stanislav Lomany

09:34 PM Feature #2156: datetime-tz and hibernate

Stanislav Lomany wrote:
> > either we force the same-zone column for all dialects
>
> I suppose we are going that... Eric Faulhaber

10/20/2019

07:31 AM Feature #3815: more temp-table options

> Everything is complete in this task?
I think yes. Code page is properly resolved and set, but data it is not act... Stanislav Lomany

10/17/2019

04:33 PM Support #3871: determine how to change codepages/locales during import

At this moment I suppose now that PostgreSQL database has been configured to have @UTF-8@ encoding. If another one-by... Sergey Ivanovskiy

10/16/2019

01:31 PM Support #3871: determine how to change codepages/locales during import

Yes, these are the only three supported dialects (see implementations of @P2JDialect@) at this moment. Ovidiu Maxiniuc

12:00 PM Support #3871: determine how to change codepages/locales during import

Ovidiu, thank you for help. Should we support only PostgreSql, H2 and MS SQL Server? Sergey Ivanovskiy

10:22 AM Support #3871: determine how to change codepages/locales during import

It seems that not all database developers have chosen to honour ANSI standard.
For H2 (I updated my h2 test envir... Ovidiu Maxiniuc

04:26 AM Support #3871: determine how to change codepages/locales during import

Ovidiu Maxiniuc wrote:
> Review of r11346 / 3871a.
>
> Generally I am OK with the new code as it seem logical, ex... Sergey Ivanovskiy

07:04 AM Feature #2156: datetime-tz and hibernate

> * In what use case(s) is the UTC timestamp sufficient?
Also comparison and WHERE statement. Stanislav Lomany

06:46 AM Feature #2156: datetime-tz and hibernate

> * In what use case(s) do we need the timezone that originally was known when the datum was stored?
When @datetim... Stanislav Lomany

10/15/2019

07:29 PM Feature #2156: datetime-tz and hibernate

Sorry, I've gotten lost in all the circles here. Can we take a step back? Please help me understand the following fun... Eric Faulhaber

05:00 PM Feature #2156: datetime-tz and hibernate

> either we force the same-zone column for all dialects
I suppose we are going that way. And basically I was wonde... Stanislav Lomany

04:38 PM Feature #2156: datetime-tz and hibernate

OK, I understand, thanks. I was mislead because of the final convert. But the idea is the same. Ovidiu Maxiniuc

04:35 PM Feature #2156: datetime-tz and hibernate

>The additional column is added only because of PSQL. It converts all DTZs to local (not UTC) DT and store them that ... Stanislav Lomany

04:10 PM Feature #2156: datetime-tz and hibernate

Actually, all supported db engines (H2, PSQL and MSSQL) will handle the mixed-timezone timestamps. Please see note 23... Ovidiu Maxiniuc

03:41 PM Feature #2156: datetime-tz and hibernate

Stanislav Lomany wrote:
> In addition: now we have timestamps with different timezones in one column (correct me if ... Constantin Asofiei

03:38 PM Feature #2156: datetime-tz and hibernate

In addition: now we have timestamps with different timezones in one column (correct me if I'm wrong). I suggest to us... Stanislav Lomany

03:22 PM Feature #2156: datetime-tz and hibernate

Currently when we create a @datetime-tz@ instance with, e.g. @TIMESTAMP+05@, it stores @TIMESTAMP(+05)@ timestamp and... Stanislav Lomany

02:24 PM Feature #2156: datetime-tz and hibernate

> for PostgreSQL timezone is a session parameter
Are you assuming that all @datetime-tz@ instances have the same ... Greg Shah

12:24 PM Feature #2156: datetime-tz and hibernate

> We are going to use 2 columns. BUT the 2nd column is only for us to restore the offset when we use it in the JVM. ... Stanislav Lomany

04:29 PM Support #3871: determine how to change codepages/locales during import

Review of r11346 / 3871a.
Generally I am OK with the new code as it seem logical, except for the query from @Impor... Ovidiu Maxiniuc

02:15 PM Feature #3814: more schema metadata

Additional metadata requirements:
Need conversion and runtime support:... Eric Faulhaber

02:07 PM Feature #4176: make certain metadata tables aware of live runtime state

Add to the list of metadata which must be made "live":... Eric Faulhaber

12:46 PM Feature #3815: more temp-table options

Everything is complete in this task?
Is the gap marking up to date as well? Greg Shah

10/14/2019

04:44 PM Support #4080: AWS Aurora support

Greg Shah wrote:
> Some thoughts from Eric:
>
> In regard to pljava, as long as this can be selected/enabled for ... Eric Faulhaber

04:12 PM Support #4080: AWS Aurora support

Some thoughts from Eric:
In regard to pljava, as long as this can be selected/enabled for an Aurora instance, we s... Greg Shah

10/13/2019

01:59 PM Feature #3809: ProDataSet support

Ovidiu, please add history entries to the files you modified in 3809e. Thanks! Constantin Asofiei

10/11/2019

06:32 AM Support #3871: determine how to change codepages/locales during import

3871a was rebased up to revision 11346 over 11337 trunc. Sergey Ivanovskiy

10/10/2019

04:40 PM Feature #4323: parse full FQL statements and convert them to SQL

Ovidiu Maxiniuc wrote:
> I am trying to clean-up the code after DDL changes and I am not sure whether we need the @m... Eric Faulhaber

02:50 PM Feature #4323: parse full FQL statements and convert them to SQL

Eric,
I am trying to clean-up the code after DDL changes and I am not sure whether we need the @merge-dmo-root@ se... Ovidiu Maxiniuc

04:04 PM Feature #3815 (Review): more temp-table options

3816a has been merged into the trunk as bzr revision 11337. Stanislav Lomany

03:10 PM Support #3871: determine how to change codepages/locales during import

Committed revision 11344 (3871a) fixed missed comment. Sergey Ivanovskiy

03:01 PM Support #3871 (Review): determine how to change codepages/locales during import

The committed revision 11343 (3871a) should fix the cases when the source code page is UTF8 or a different code page ... Sergey Ivanovskiy

03:53 AM Support #3871: determine how to change codepages/locales during import

This it the test example. The question for me is that BOM (https://en.wikipedia.org/wiki/Byte_order_mark) is generate... Sergey Ivanovskiy

03:46 AM Support #3871: determine how to change codepages/locales during import

I found that the UTF-8 source dump file is imported incorrectly now. The problem here is @FileStream@ is supposed to ... Sergey Ivanovskiy

10/09/2019

08:49 AM Support #3871: determine how to change codepages/locales during import

Greg, please review the committed rev 11338 (3871a). It was tested with UTF-8 PostgresSql database. Sergey Ivanovskiy

10/08/2019

03:00 PM Feature #2156: datetime-tz and hibernate

Constantin Asofiei wrote:
> There's one other thing to add for @datetime-tz@ - its usage in indexes and sort clauses... Constantin Asofiei

01:47 PM Feature #2156: datetime-tz and hibernate

Stanislav Lomany wrote:
> Could you please elaborate what should be done for this task? datetime-tz seems to be work... Constantin Asofiei

12:21 PM Feature #2156: datetime-tz and hibernate

Could you please elaborate what should be done for this task? datetime-tz seems to be working in general. Stanislav Lomany

10:37 AM Support #3871: determine how to change codepages/locales during import

I doesn't matter what one byte encoding is used. If it is internally mapped into UTF-8, then this issue can be observ... Sergey Ivanovskiy

10:32 AM Support #3871: determine how to change codepages/locales during import

Sergey Ivanovskiy wrote:
> OK, it seems that the Hibernate uses UTF-8 internally because I experimented with US-ASCI... Sergey Ivanovskiy

06:52 AM Support #3871: determine how to change codepages/locales during import

OK, it seems that the Hibernate uses UTF-8 internally because I experimented with US-ASCII encoding as a default enco... Sergey Ivanovskiy

06:42 AM Support #3871: determine how to change codepages/locales during import

If there is no valid codepage conversion for the INPUT_ENCODING to the DATABASE_ENCODING, then you should throw an ex... Greg Shah

06:26 AM Support #3871: determine how to change codepages/locales during import

This is that imported file encoded as WINDOWS-1251. Sergey Ivanovskiy

06:24 AM Support #3871: determine how to change codepages/locales during import

I don't know the way how to convert parameters for HSQL queries like this one... Sergey Ivanovskiy

06:16 AM Support #3871: determine how to change codepages/locales during import

Sergey Ivanovskiy wrote:
> Greg Shah wrote:
> > > Then the client will read the imported data from this database as... Greg Shah

06:10 AM Support #3871: determine how to change codepages/locales during import

>Where are the UTF-8 strings coming from? Are they from the import data?
UTF-8 is the default encoding for my java e... Sergey Ivanovskiy

05:59 AM Support #3871: determine how to change codepages/locales during import

Greg Shah wrote:
> > Then the client will read the imported data from this database as LATIN-1 strings because the d... Sergey Ivanovskiy

05:47 AM Support #3871: determine how to change codepages/locales during import

> Then the client will read the imported data from this database as LATIN-1 strings because the database has LATIN-1 ... Greg Shah

05:25 AM Support #3871: determine how to change codepages/locales during import

Sergey Ivanovskiy wrote:
> I think that there is only one way to save UTF-8 strings into LATIN-1 database is to conv... Sergey Ivanovskiy

05:15 AM Support #3871: determine how to change codepages/locales during import

I think that there is only one way to save UTF-8/Unicode strings into LATIN-1 database that is to convert UTF-8 strin... Sergey Ivanovskiy

10/07/2019

02:17 PM Support #3871: determine how to change codepages/locales during import

I encountered the following issue that the WINDOWS-1251 string decoded into the UTF-8 string (if UTF-8 is a default j... Sergey Ivanovskiy

06:26 AM Support #3871: determine how to change codepages/locales during import

Let us consider this case in which the database has LATIN-1 encoding and the import table has WINDOWS-1251 encoding. ... Sergey Ivanovskiy

10:21 AM Feature #4323: parse full FQL statements and convert them to SQL

These are generateDDL() calls for the metadata database. I am working on it right now. Ovidiu Maxiniuc

10/04/2019

05:35 PM Feature #4323: parse full FQL statements and convert them to SQL

Are there any other project-level changes that need to be made, beside those in your last note? I'm trying to convert... Eric Faulhaber

04:01 PM Feature #4323: parse full FQL statements and convert them to SQL

I committed latest changes to 4011a/rr11344.
If you have an old ddl directory, please backup it. The conversion wi... Ovidiu Maxiniuc

02:16 PM Feature #4323: parse full FQL statements and convert them to SQL

Is there a new class @DDLGeneratorWorker@ that was supposed to be checked in?... Eric Faulhaber

01:59 PM Feature #4323: parse full FQL statements and convert them to SQL

I forgot to CTRL+F5 :(. Ovidiu Maxiniuc

01:57 PM Feature #4323: parse full FQL statements and convert them to SQL

Sorry, Eric. I cleaned the .g file only looking at the @ExpressionParser@. Indeed, I was too eager to drop the @java.... Ovidiu Maxiniuc

01:54 PM Feature #4323: parse full FQL statements and convert them to SQL

BTW, Greg has decreed that "FWDQL" should be "FQL" :)
Please work that name change back into the code, file names,... Eric Faulhaber

01:50 PM Feature #4323: parse full FQL statements and convert them to SQL

Ovidiu, I'm getting the following compile errors with 4011a/11342:... Eric Faulhaber

07:02 AM Bug #3471: incremental conversion is regressed

> for OO, the tricky part is keeping the Java converted method name consistent. If you recall, we need to keep this c... Greg Shah

10/03/2019

05:34 PM Bug #3471: incremental conversion is regressed

I'm still checking, but the current issues are:
* for OO, the tricky part is keeping the Java converted method name ... Constantin Asofiei