FWD » Core Development » Base Language

03/07/2014

09:06 PM Feature #1634: implement full native library (.so or DLL) support

The update for you to review to fix @name_map.xml@ entries. Only @name_map.xml@ generation rule is involved. So no ne... Eugenie Lyzenko

03/05/2014

04:53 PM Feature #1634: implement full native library (.so or DLL) support

Some finding for execution the test in 4GL systems:
Source:... Eugenie Lyzenko

04:48 PM Feature #1634: implement full native library (.so or DLL) support

> And what about native library path names? For example @"library_calls/testapi/testapi.dll"@. Here we also need Linu... Greg Shah

04:05 PM Feature #1634: implement full native library (.so or DLL) support

> We want to implement all name_map.xml entries in Linux form, even if the app was written with Windows file seps. In... Eugenie Lyzenko

03:52 PM Feature #1634: implement full native library (.so or DLL) support

Greg Shah wrote:
> Constantin: please provide your thoughts on where exactly to put this code. I know we need to re... Constantin Asofiei

03:43 PM Feature #1634: implement full native library (.so or DLL) support

We want to implement all name_map.xml entries in Linux form, even if the app was written with Windows file seps. In ... Greg Shah

03:22 PM Feature #1634: implement full native library (.so or DLL) support

> I understand. But my concern is that changing the p2j.cfg.xml has other implications that may make this a bad idea.... Eugenie Lyzenko

01:25 PM Feature #1634: implement full native library (.so or DLL) support

No, you are right. The @p2j.cfg.xml.win@ should have @case-sensitive==false@. Keep this change. Greg Shah

12:59 PM Feature #1634: implement full native library (.so or DLL) support

> I understand. But my concern is that changing the p2j.cfg.xml has other implications that may make this a bad idea.... Eugenie Lyzenko

12:24 PM Feature #1634: implement full native library (.so or DLL) support

> This change was made according to note #96 above.
I understand. But my concern is that changing the p2j.cfg.xml... Greg Shah

12:11 PM Feature #1634: implement full native library (.so or DLL) support

>... Eugenie Lyzenko

11:40 AM Feature #1634: implement full native library (.so or DLL) support

Code Review 0305a
The change to @collect_names.rules@ is fine. The @null@ check is not really needed, because if ... Greg Shah

11:19 AM Feature #1634: implement full native library (.so or DLL) support

Code Review 0305b
The change to these values is concerning:... Greg Shah

10:47 AM Feature #1634: implement full native library (.so or DLL) support

Two updates for you to review, one for P2J code, other - for testcases(@p2j.cfg.xml.win@ to convert testcases). As fo... Eugenie Lyzenko

09:22 AM Feature #1634: implement full native library (.so or DLL) support

Please post a real update so that I can use diffing tools (like meld) to compare your changes. The changes you have ... Greg Shah

03/04/2014

05:31 PM Feature #1634: implement full native library (.so or DLL) support

The changes for you to review to convert/compile/run testcases in Windows. The result - there are no messages during ... Eugenie Lyzenko

12:49 PM Feature #1634: implement full native library (.so or DLL) support

> Using libname = ExpressionConversionWorker.progressToJavaString(libname) causes the exception:
Please search the... Greg Shah

12:42 PM Feature #1634: implement full native library (.so or DLL) support

The strange thing I'm struggling with. Using libname = ExpressionConversionWorker.progressToJavaString(libname) cause... Eugenie Lyzenko

03/03/2014

04:33 PM Feature #1599: shift EXP implementation (MathOps.pow()) to BigDecimal

Code Review 0227a
It looks fine. Assuming it passes testing, you can check it in and distribute it. Greg Shah

02/28/2014

10:06 AM Feature #1634: implement full native library (.so or DLL) support

> The place to fix is:
Yes. Instead of the call to @removeQuotes()@, it should call @ExpressionConversionWorker.p... Greg Shah

09:20 AM Feature #1634: implement full native library (.so or DLL) support

> ... But in this case the text is stored into the name_map.xml during annotations/collect_names.rules and the string... Eugenie Lyzenko

08:55 AM Feature #1634: implement full native library (.so or DLL) support

> Greg, please advise: is this a conversion bug in the library name conversion, or a problem in load_library_native, ... Greg Shah

05:56 AM Feature #1634: implement full native library (.so or DLL) support

Eugenie Lyzenko wrote:
> This is just fragment, the others - inside attached
> [...]
The @~\@ comes from the 4GL c... Constantin Asofiei

05:42 AM Feature #1634: implement full native library (.so or DLL) support

This is just fragment, the others - inside attached... Eugenie Lyzenko

04:52 AM Feature #1634: implement full native library (.so or DLL) support

Eugenie Lyzenko wrote:
> > What I think we need to do is to keep the program names in name_map.xml with the linux-st... Constantin Asofiei

04:48 AM Feature #1634: implement full native library (.so or DLL) support

> What I think we need to do is to keep the program names in name_map.xml with the linux-style separator.
After ch... Eugenie Lyzenko

02:55 AM Feature #1634: implement full native library (.so or DLL) support

Greg Shah wrote:
> No, this is not a conversion problem. We can't change the source code since it is valid 4GL and ... Constantin Asofiei

02/27/2014

06:06 PM Feature #1634: implement full native library (.so or DLL) support

Another similar issue is library name transformation from requested @"library_calls\\testapi\\testapi.dll"@ to @"libr... Eugenie Lyzenko

03:38 PM Feature #1634: implement full native library (.so or DLL) support

In the 4GL, this code works without problem on both Linux and Windows.
This means there is a runtime problem in ou... Greg Shah

02:54 PM Feature #1634: implement full native library (.so or DLL) support

The first thing I've faced is the issue with hardcoded test file path name to run for example:... Eugenie Lyzenko

11:19 AM Feature #1634: implement full native library (.so or DLL) support

> The Linux copy has the error I need to clarify:
Please compare your results to the logs in note 79 to see if err... Greg Shah

10:30 AM Feature #1634: implement full native library (.so or DLL) support

I'm executing test_runner.p in both Linux and Windows to compare behavior/results. The Linux copy has the error I nee... Eugenie Lyzenko

08:17 AM Feature #1634: implement full native library (.so or DLL) support

> I think what you are missing is copy the p2jspi.jar in your JRE's ext lib folder.
Yes, you are right, the server... Eugenie Lyzenko

07:54 AM Feature #1634: implement full native library (.so or DLL) support

Eugenie Lyzenko wrote:
> Ti enable persistence for server the following changes should be done in directory.xml file... Constantin Asofiei

07:52 AM Feature #1634: implement full native library (.so or DLL) support

Regarding previous note error. Do I need to set up the SQL server running on Windows to execute the tests? Eugenie Lyzenko

07:49 AM Feature #1634: implement full native library (.so or DLL) support

Trying to start the server on Windows for native library with persistence enabled. But the error is happening.
Ti ... Eugenie Lyzenko

03:06 PM Feature #1599: shift EXP implementation (MathOps.pow()) to BigDecimal

Fixed error handling issue in MathOps.pow(BigDecimal, BigDecimal). Regression test is in progress. Hynek Cihlar

02:55 PM Bug #2133: fix precision for decimal, dynamic-extent variables

Constantin Asofiei wrote:
> When the number of array elements is missing, then we have a dynamic extent. And looks l... Hynek Cihlar

12:00 PM Bug #2133: fix precision for decimal, dynamic-extent variables

Hynek Cihlar wrote:
> Setting decimal precision in the @DEFINE VARIABLE@ statement of an @EXTENT@ with the missing n... Constantin Asofiei

11:13 AM Bug #2133: fix precision for decimal, dynamic-extent variables

Setting decimal precision in the @DEFINE VARIABLE@ statement of an @EXTENT@ with the missing number of array elements... Hynek Cihlar

02:52 PM Bug #2250: Conversion error when decimal precision defined on dynamic extent variable

Setting decimal precision in the @DEFINE VARIABLE@ statement of an @EXTENT@ with the missing number of array elements... Hynek Cihlar

02:51 PM Bug #2250 (Closed): Conversion error when decimal precision defined on dynamic extent variable

Hynek Cihlar

02/26/2014

03:35 PM Feature #1599 (Closed): shift EXP implementation (MathOps.pow()) to BigDecimal

Greg Shah

03:17 PM Feature #1599: shift EXP implementation (MathOps.pow()) to BigDecimal

0220a regression tested and checked in to bzr revision 10481. Hynek Cihlar

02/25/2014

12:57 PM Feature #1608: implement full appserver support (from 4GL clients only)

About static resources generated in a appserver procedure ran persistent (using @RUN ... PERSISTENT ON SERVER@ or via... Constantin Asofiei

02/24/2014

02:26 PM Bug #2133: fix precision for decimal, dynamic-extent variables

Greg Shah wrote:
> > Can it ever happen to have the array reference registered to multiple scopes on the scope stack... Hynek Cihlar

02:18 PM Bug #2133: fix precision for decimal, dynamic-extent variables

> Can it ever happen to have the array reference registered to multiple scopes on the scope stack?
Yes. I believe... Greg Shah

01:47 PM Bug #2133: fix precision for decimal, dynamic-extent variables

When the dynamic array is resized (@EXTENT@ statement is encountered), old array reference is "deregistered" and new ... Hynek Cihlar

12:11 PM Feature #2145 (Closed): eliminate program-name limitations

Greg Shah

11:50 AM Feature #2145: eliminate program-name limitations

Greg Shah wrote:
> Have you retested the "standalone" super proc/persistent proc testcases?
Yes, the tests are OK.
... Constantin Asofiei

10:49 AM Feature #2145: eliminate program-name limitations

Code Review 0222b/e
I am fine with the changes. This is another nice improvement/refactoring of the code such tha... Greg Shah

02/23/2014

03:15 PM Feature #1599 (WIP): shift EXP implementation (MathOps.pow()) to BigDecimal

Hynek Cihlar

03:13 PM Bug #2133 (WIP): fix precision for decimal, dynamic-extent variables

Hynek Cihlar

03:12 PM Bug #2133: fix precision for decimal, dynamic-extent variables

The decimal precision is effectively lost when (1) unfixed decimal array is defined and later fixed with the @EXTENT@... Hynek Cihlar

04:18 AM Feature #2145 (Review): eliminate program-name limitations

Constantin Asofiei

04:14 AM Feature #2145: eliminate program-name limitations

The 0222b.zip and 0222e.zip (from #2248) have passed regression testing, after checking the cumulative regression tes... Constantin Asofiei

02/22/2014

09:47 AM Feature #2145: eliminate program-name limitations

Failures during testing of 0222b.zip: there are cases like this (MAJIC uses them):... Constantin Asofiei

05:05 AM Feature #2145: eliminate program-name limitations

This version allows java-style calls for internal and external procedures and centralizes the maintenance of procedur... Constantin Asofiei

02/21/2014

09:15 AM Feature #2145: eliminate program-name limitations

Greg Shah wrote:
> Option 3 is the best solution, IF it is possible.
I think I will go ahead with this one. The onl... Constantin Asofiei

08:31 AM Feature #2145: eliminate program-name limitations

Option 3 is the best solution, IF it is possible.
A modified option 1 may be the next best approach. The idea is ... Greg Shah

04:10 AM Feature #2145 (WIP): eliminate program-name limitations

Regression testing showed problems related to calling the hand-written java code: we need to determine the policy of ... Constantin Asofiei

02/20/2014

04:45 PM Feature #1625: implement OUTPUT TO PRINTER on Windows

The Linux/UNIX support for OUTPUT TO PRINTER is already done from the work in #1631. The remaining work for this tas... Greg Shah

04:37 PM Feature #1824 (Closed): client platform delegate implementation

The design choices implemented in #1811 eliminate the need for this work. Greg Shah

11:05 AM Feature #2145: eliminate program-name limitations

> The only dependence on reflection is the "java-style calls of legacy user-def functions", when we need to determine... Greg Shah

10:56 AM Feature #2145 (Review): eliminate program-name limitations

Greg Shah wrote:
> We don't have much choice in coding so much dependence on reflection and the old approach was als... Constantin Asofiei

09:58 AM Feature #2145: eliminate program-name limitations

Code Review 0220d
I am OK with the changes. As always with the procedure processing, this stuff is pretty nasty. ... Greg Shah

09:21 AM Feature #2145 (WIP): eliminate program-name limitations

This solves the OOME and the SOURCE- and TARGET-PROCEDURE deviations. Is going through testing now. Constantin Asofiei

08:21 AM Feature #1599: shift EXP implementation (MathOps.pow()) to BigDecimal

Code Review 0220a
Everything looks good. If it passes runtime regression testing, you can check it in and distrib... Greg Shah

03:58 AM Feature #1599: shift EXP implementation (MathOps.pow()) to BigDecimal

I also moved the @log@ private methods down to the private methods section. The attached changes are submitted for re... Hynek Cihlar

01:26 AM Feature #1624: add support for PUT CONTROL

Evgeny Kiselev wrote:
> Do I need to commit testcases ?
Of course, the testcases need to be in the repository. Constantin Asofiei

11:00 PM Feature #1624: add support for PUT CONTROL

Do I need to commit testcases ? Evgeny Kiselev

11:59 PM Feature #2208: create a java application running web services for testing purposes

I've committed intermediate version of test server to the @testcases/uast/web_services/server/@.
To run this from ID... Evgeny Kiselev

02/19/2014

03:41 PM Feature #1599: shift EXP implementation (MathOps.pow()) to BigDecimal

Code Review 0219a
The code looks good. My only comments relate to some coding standards:
1. Lines 608 and 614 a... Greg Shah

02:52 PM Feature #1599: shift EXP implementation (MathOps.pow()) to BigDecimal

Please review the attached changes. Testcases checked in to testcases/uast/math/exp.p. Hynek Cihlar

12:15 PM Feature #1624 (Closed): add support for PUT CONTROL

Greg Shah

11:52 AM Feature #1624: add support for PUT CONTROL

update evk_upd20140214a.zip has been passed regression testing.
Committed to bzr revision 10470. Evgeny Kiselev

11:13 PM Feature #1624 (Review): add support for PUT CONTROL

Evgeny Kiselev

02/18/2014

01:50 PM Feature #2145: eliminate program-name limitations

This should solve the SOURCE- and TARGET-PROCEDURE cases (when invoking an external program or a trigger). Also, it i... Constantin Asofiei

11:02 AM Feature #1634: implement full native library (.so or DLL) support

Almost.
> take the testcases from @native_library_calls_4gl_..._sample_20131107.zip@
No, that is just a slightly... Greg Shah

10:51 AM Feature #1634: implement full native library (.so or DLL) support

Greg,
The question about task #2234. The goal is to take the testcases from @native_library_calls_4gl_..._sample_2... Eugenie Lyzenko

02/17/2014

11:53 AM Support #2233 (Closed): test and fix the native API support on Linux

Greg Shah

02/16/2014

08:01 AM Feature #1624: add support for PUT CONTROL

Check it in and distribute it. Greg Shah

03:20 AM Feature #1624: add support for PUT CONTROL

Evgeny Kiselev wrote:
> Final update. Regression is running now.
Passed regression testing Evgeny Kiselev

10:04 PM Feature #1634: implement full native library (.so or DLL) support

Between 2 regression testing runs, all tests have passed except for GSO 167 (and TC JOB 002 which fails as expected i... Greg Shah

02/15/2014

01:24 PM Feature #2145: eliminate program-name limitations

Greg Shah wrote:
> > my concern is that, if the PROPATH gets modified, we might resolve the relative name to a compl... Constantin Asofiei

01:20 PM Feature #2145: eliminate program-name limitations

> my concern is that, if the PROPATH gets modified, we might resolve the relative name to a completely different prog... Greg Shah

09:06 AM Feature #2145: eliminate program-name limitations

This one proves a little more challenging: the procedure names in error messages and the procedure's @FILE-NAME@ and ... Constantin Asofiei

08:59 AM Feature #1634: implement full native library (.so or DLL) support

This update is merged up to bzr revision 10466. I am going back through runtime regression testing. The first run f... Greg Shah

08:56 AM Feature #1598 (Closed): shift LOG implementation (MathOps.log()) to BigDecimal

Greg Shah

07:32 AM Feature #1598: shift LOG implementation (MathOps.log()) to BigDecimal

Committed to bzr revision 10466. Hynek Cihlar

07:41 PM Feature #1624: add support for PUT CONTROL

Final update. Regression is running now. Evgeny Kiselev

02/14/2014

05:47 PM Feature #1598: shift LOG implementation (MathOps.log()) to BigDecimal

Code Review 0214a
The changes look good. You can check them in and distribute them. Greg Shah

07:13 AM Feature #1598: shift LOG implementation (MathOps.log()) to BigDecimal

Introduced @toBigDecimal@ abstract method in @NumberType@ and fixed the file headers. The code changes passed regress... Hynek Cihlar

05:26 PM Feature #1624: add support for PUT CONTROL

Update the copyright dates and post the final update here. Then start regression testing (both conversion and runtim... Greg Shah

07:45 PM Feature #1624: add support for PUT CONTROL

Greg Shah wrote:
> Code Review 0208a
>
> 1. Does the @Steam.generateFormattedText()@ have implications when the s... Evgeny Kiselev

02:16 AM Feature #2208: create a java application running web services for testing purposes

Evgeny Kiselev wrote:
> Where I can upload update ?
> I've ~20mb zip archive (with axis2 libs and configs).
Commi... Constantin Asofiei

07:23 PM Feature #2208: create a java application running web services for testing purposes

Where I can upload update ?
I've ~20mb zip archive (with axis2 libs and configs).
Evgeny Kiselev

02/13/2014

01:30 PM Feature #1634: implement full native library (.so or DLL) support

All testcase updates and captured logs/output are checked into the testcases project.
The code has passed conversi... Greg Shah

01:03 PM Feature #1634: implement full native library (.so or DLL) support

As can be seen from the logs, the remaining diffs are:
1. A missing error 86 in one call to output_float_from_buff... Greg Shah

12:36 PM Feature #1634: implement full native library (.so or DLL) support

This incorporates many fixes and the code review feedback. It allows the API tests to run properly on Linux. The on... Greg Shah

02/12/2014

06:27 PM Feature #1598: shift LOG implementation (MathOps.log()) to BigDecimal

Code Review 0212a
It looks pretty good. Fix these minor issues and get it tested:
1. I'd prefer to add an abst... Greg Shah

04:33 PM Feature #1598: shift LOG implementation (MathOps.log()) to BigDecimal

The attached file contains @log@ function implementation and related changes for review.
The implemented @log@ fun... Hynek Cihlar

05:23 PM Feature #1634: implement full native library (.so or DLL) support

> or better:... Greg Shah

03:59 PM Feature #1597 (Closed): shift SQRT implementation (MathOps.sqrt()) to BigDecimal

Greg Shah

02:04 PM Feature #1597: shift SQRT implementation (MathOps.sqrt()) to BigDecimal

Committed to revision 10461. Hynek Cihlar

10:37 AM Feature #1597: shift SQRT implementation (MathOps.sqrt()) to BigDecimal

Code Review 0209a
The code looks good. Please commit and distribute it. Greg Shah

10:23 AM Feature #1597: shift SQRT implementation (MathOps.sqrt()) to BigDecimal

The attached file is the final version.
Note, that the previously added condition of @sqrt(?) == 1@ was removed. ... Hynek Cihlar

01:26 AM Feature #2208: create a java application running web services for testing purposes

Evgeny Kiselev wrote:
> Yes, you are right.
If you think you can solve this in 2-4 hours, then go ahead and solve i... Constantin Asofiei

02/11/2014

05:48 PM Feature #2208: create a java application running web services for testing purposes

Yes, you are right. Evgeny Kiselev

02:12 PM Feature #2208: create a java application running web services for testing purposes

OK, I think I understand your problem: you are you using embedded AXIS2 web server, which does not support HTTPS. An... Constantin Asofiei

08:17 PM Feature #2208: create a java application running web services for testing purposes

I have problem with SSL. It's not possible to do it in the internal Axis2 web server(SimpleHttpServer).
Here is a ... Evgeny Kiselev

05:47 PM Feature #1624: add support for PUT CONTROL

Code Review 0208a
1. Does the @Steam.generateFormattedText()@ have implications when the statement is not PUT CONT... Greg Shah

02/10/2014

08:09 PM Feature #1624: add support for PUT CONTROL

1) Added conversion for @COPY-LOB@ from file to memptr
2) Added implementation for @COPY-LOB@
3) Added for @RAW@ ty... Evgeny Kiselev

02/06/2014

10:17 AM Feature #1597: shift SQRT implementation (MathOps.sqrt()) to BigDecimal

Looks great! Greg Shah

10:02 AM Feature #1597: shift SQRT implementation (MathOps.sqrt()) to BigDecimal

Yes, the "4GL" is a typo, thanks for catching it. I am attaching a fix and sending it to regression test. Hynek Cihlar

08:15 AM Feature #1597: shift SQRT implementation (MathOps.sqrt()) to BigDecimal

Code Review 0205a
This looks good.
My only question is about this table:... Greg Shah

07:20 PM Feature #1597: shift SQRT implementation (MathOps.sqrt()) to BigDecimal

The input of one is covered by the initial approximation, but yes it won't hurt if stated explicitly.
New in the a... Hynek Cihlar

02/05/2014

09:03 AM Feature #1624: add support for PUT CONTROL

> except wrong conversion of @copy-lob from file "test_put_control.txt" to m.@. P2J convert it's in uncompilable code... Greg Shah