firebird-devel — List for general discussion among IB developers

Re: [Firebird-devel] Any recent sources available by anything but CVS?

[Mike N. <ta...@al...>]

Leyne, Sean wrote:
>
> > http://cvs.sourceforge.net/cvstarballs/firebird-cvsroot.tar.gz
>
> I use IE 6.0 (this is also is true of v5.x) and have never been able to
> use this method to download the tarball, the download never ends....  (I
> just tried it, and gave up after 100MB -- the file said that it was 34MB
> in size)

Couldn't this be the case where MS once again knows better than you and
automagically decompress that gzipped file for you on the fly? Isn't it
great with user friendliness! :-<

You might try to right click the file and choose "Save Target As...".

/Mike

RE: [Firebird-devel] Any recent sources available by anything but CVS?

[Rick F. <Ric...@qu...>]

I've got a call in with our MIS department to see if they can put a hole
into 
the firewall.. I'm not particularly optimistic, but I'll keep my fingers
crossed.
In the meantime, I'm downloading that aforementioned 34M CVS file -- it's
currently
stuck at 11.8Mb.. Perhaps I'll have to figure out a way to do this from home
via
DSL.. Aren't firewalls great!

-----Original Message-----
From: Leyne, Sean [mailto:sl...@at...]
Sent: Monday, December 10, 2001 3:39 PM
To: Firebird-Devel (E-mail)
Subject: RE: [Firebird-devel] Any recent sources available by anything
but CVS?


Mike,

> If all else fails, you could use the following link
> 
> http://cvs.sourceforge.net/cvstarballs/firebird-cvsroot.tar.gz

I use IE 6.0 (this is also is true of v5.x) and have never been able to
use this method to download the tarball, the download never ends....  (I
just tried it, and gave up after 100MB -- the file said that it was 34MB
in size)

So, unfornately this is not always an available option...


Sean

_______________________________________________
Firebird-devel mailing list
Fir...@li...
https://lists.sourceforge.net/lists/listinfo/firebird-devel

RE: [Firebird-devel] Any recent sources available by anything but CVS?

[Leyne, S. <sl...@at...>]

Mike,

> If all else fails, you could use the following link
>=20
> http://cvs.sourceforge.net/cvstarballs/firebird-cvsroot.tar.gz

I use IE 6.0 (this is also is true of v5.x) and have never been able to
use this method to download the tarball, the download never ends....  (I
just tried it, and gave up after 100MB -- the file said that it was 34MB
in size)

So, unfornately this is not always an available option...


Sean

Re: [Firebird-devel] Any recent sources available by anything but CVS?

[Mike N. <ta...@al...>]

Rick Flower wrote:
>

> Unfortunately, we currently can't do anonymous
> CVS due to a firewall issue beyond my control.  Is there any other way to
> get at the latest version of the firebird v1 source tree?

Can you get SSH through the firewall? In that case it might work with SSH
tunneling/port forwarding for CVS.

If all else fails, you could use the following link

http://cvs.sourceforge.net/cvstarballs/firebird-cvsroot.tar.gz

but I can't really recommend it since this is not only the Firebird 1 tree,
but _everything_ that is in our CVS (details at
http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/firebird/).

/Mike

[Firebird-devel] Any recent sources available by anything but CVS?

[Rick F. <Ric...@qu...>]

Hi.. I'm a firebird newbie.. I just figured out that the "nightly" snapshot
source tarballs available at
http://firebird.sourceforge.net/download_Source_Code.php haven't been
updated since Sept. 29th.. Unfortunately, we currently can't do anonymous
CVS due to a firewall issue beyond my control.  Is there any other way to
get at the latest version of the firebird v1 source tree?  
 
Thanks.. 

[Firebird-devel] libc 2.1.3: can I build with it?

[Roman R. <rro...@ya...>]

hi,

firebird RC1 build for Debian requires libc 2.2. my
current debian has 2.1.3. Is it possible to build
firebird from sources using this libc or should I
upgrade my debian to libc 2.2?

TIA,
Roman Rokytskyy

__________________________________________________
Do You Yahoo!?
Send your FREE holiday greetings online!
http://greetings.yahoo.com

Re: [Firebird-devel] ChangeLog + build number

[John B. <bel...@cs...>]

On Monday, December 10, 2001, at 11:45 AM, Frank Schlottmann-Goedde 
wrote:

> On Saturday,  8. December 2001 23:54, John Bellardo wrote:
>> Frank,
>
>> I've included a patch at the end of this message.  It should append the
>> build number at the time the change log is generated, to the message.
>> That means all entries that get generated at the same time will have 
>> the
>> same build number.  But that is better than what we have now, I guess.
>
> Thank you,
> but in the meantime I've written a postprocessor in Free Pascal, that
> generates headers like:
>
> 2001/12/09 20:12 nmcc Build Nr. : 659
>
> This should be good enough, and I don't have to fiddle around with 
> cvs2cl.pl

Great!  Do you think we should start storing these little helper 
scripts/programs in CVS?  Perhaps we should create a module named 
"cvs_utilities" or some such?

-John

Re: [Firebird-devel] ChangeLog + build number

[Frank Schlottmann-G. <sch...@t-...>]

On Saturday,  8. December 2001 23:54, John Bellardo wrote:
> Frank,

> I've included a patch at the end of this message.  It should append the
> build number at the time the change log is generated, to the message.
> That means all entries that get generated at the same time will have the
> same build number.  But that is better than what we have now, I guess.

Thank you, 
but in the meantime I've written a postprocessor in Free Pascal, that 
generates headers like:

2001/12/09 20:12 nmcc Build Nr. : 659

This should be good enough, and I don't have to fiddle around with cvs2cl.pl

Frank 

-- 
"Fascinating creatures, phoenixes. They can carry immensely heavy loads,
 their tears have healing powers and they make highly faithful pets."
          - J.K. Rowling                         http://firebirdsql.org

[Firebird-devel] grant bug

[Ann W. H. <aha...@ib...>]

I found the grant bug - a typo that caused it to save
the wrong request handle.

Regards,

Ann

[Firebird-devel] protecting metadata in IB 6.5

[Dmitry K. <di...@de...>]

Hello, All!

Protectng metadata is good feature in IB 6.5, but not new.
I'm ready to dispute about header in readmeta.sql, blindmeta.sql, writemeta.sql:

/*
 * Copyright (C) 2001 Borland Software Corporation
 * All Rights Reserved.
 * Contributor(s): ______________________________________.
 */

This can't be copyrighted, since one person named Michael Podstrechny
found this solution for IB 4.x and sent to me.
And I've PUBLISHED it in 1997:
(sorry that text in Russian, but that's as document was published
originally. We are only interested only in grant/revoke commands and idea itself)
http://ib.demo.ru/DevInfo/sysprot.htm

Original file date was lost, sorry, but anyway last change date is
29 october 1998.

So, I'm proud to say that EVERYONE can use this idea, without any
mention of Borland Software Corp. (with my respect to it).
I understand Borland's motivation, i.e. all included in distribution
must be marked. But in this situation they have no (c).

p.s. If I apply readmeta.sql on FB database, every user can create his
own tables. But he can't delete or edit records directly in RDB$.
(that's why I don't understand purpose of writemeta.sql - to enable
every user directly run 'delete from rdb$user_privileges' ?).
Blindmeta.sql hides everything (except for SYSDBA) - RDB$ are not
accessible and thus user can't select any data from any table.

So, as I understand, this feature (new 6.5 metadata protection) is in switching 
off privileges check for system tables RDB$ ? In this case, I don't understand
why they need to be checked at all. Let engine check privileges for any objects,
except system.

p.p.s. it is much easier to cut some part of code, rather than to write new. Isn't it? :-)

-- 
Dmitry Kuzmenko, Epsylon Technologies.

RE: [Firebird-devel] Aggregate doesn't work inside UDF (new case) more clear

[Christian P. <c_p...@ya...>]

Hello Neil

I didn't see the post in firebird-checkins and my last build is 647.
Sorry and thank you for your patience.

Christian

Neil McCalden <nm...@zi...> escribió en el mensaje de noticias
200...@zi......
> On Mon, Dec 10, 2001 at 12:03:08AM -0300, Christian Pradelli wrote:
> >
> [opinions on select and group by]
> >
> > SELECT CUSTOMER, DIVIDE(ABS(TOTAL),TOTAL), ABS(SUM(TOTAL))
> > FROM SALES
> > GROUP BY CUSTOMER, DIVIDE(ABS(TOTAL),TOTAL)
> >
> > here you have an error.
> >
>
> If you are not subscribed to firebird-checkins I can see my message was
> probably not clear enough, this case is fixed as well now - build 658.
>
>
> --
> Neil McCalden                @home     nm...@zi...
>
> _______________________________________________
> Firebird-devel mailing list
> Fir...@li...
> https://lists.sourceforge.net/lists/listinfo/firebird-devel
>

Re: [Firebird-devel] Aggregate doesn't work inside UDF (new case) more clear

[Neil M. <nm...@zi...>]

On Mon, Dec 10, 2001 at 12:03:08AM -0300, Christian Pradelli wrote:
> 
[opinions on select and group by]
> 
> SELECT CUSTOMER, DIVIDE(ABS(TOTAL),TOTAL), ABS(SUM(TOTAL))
> FROM SALES
> GROUP BY CUSTOMER, DIVIDE(ABS(TOTAL),TOTAL)
> 
> here you have an error.
> 

If you are not subscribed to firebird-checkins I can see my message was
probably not clear enough, this case is fixed as well now - build 658. 


-- 
Neil McCalden                @home     nm...@zi...     

Re: [Firebird-devel] Building in HP-UX 11

[Paul B. <pa...@wa...>]

<<I´ve mangged to get a machine with HP-UX 11 to try to make a build
from this monday on. I got the document "Boot build on HP-UX and
Solaris" at IBPhoenix´s site. Does anybody have a hint to give me?
Any recommendation?>>

I haven't tried this, although I am aware of it. I am trying to get a build
up and running using Danny Mavromatis' test HP-UX build and the firebird
build scripts.
Although the information Martin has provided is very useful. Danny's build
however doesn't have a copy of gdef, which is needed for the build script,
so I will have to create this seperately.

A couple of comments though.

1. The "source" command is only available as a built in function for the C
shell, so you will need to define certain environment variables by using a
different mechanism to the firebird build script Configure_SetupEnv.sh
2. The firebird build script doesn't really know about the HP10 define, some
changes are needed to get it to run.
3. HP Make is "pants" you will need gmake
4. There are a number of changes that need to be made to the code, as the
HP-UX compiler is very ANSI and doesn't like certain syntax and the way the
code is currently. I intend to log all of these changes and commit the
changes back to CVS when I get a sucessful build.


Regards
Paul

RE: [Firebird-devel] ibeng32.dll

[Claudio V. C. <cv...@us...>]

> -----Original Message-----
> From: Dmitri Zakharov [mailto:dz...@mi...]
> Sent: Domingo 9 de Diciembre de 2001 23:19
>
> This "engine in dll" was an official Borland binary in IB 4.0 for NT.
> By the way, it was alone engine at all. Remote access was provied by
> another wrapper program (ibremote.exe), which also called gds32.dll,
> and worked as a service.

I used IB4.0. But I remember that it was a 16-BIT RELEASE, so gds32 didn't
exist. I just rescued the CD from the dust and these are the DLLs:
RUNIMAGE\IBLOCAL\BIN\DSQL.DLL
RUNIMAGE\IBLOCAL\BIN\FILEIO.DLL
RUNIMAGE\IBLOCAL\BIN\GBAK.DLL
RUNIMAGE\IBLOCAL\BIN\GDS.DLL
RUNIMAGE\IBLOCAL\BIN\INTL.DLL
RUNIMAGE\IBLOCAL\BIN\IUTLS.DLL
RUNIMAGE\IBLOCAL\BIN\JRD.DLL
RUNIMAGE\IBLOCAL\BIN\REMOTE.DLL
RUNIMAGE\IBLOCAL\BIN\STACK.DLL

At that time, InstallShield was Stirling Technologies, Inc. or they
purchased that company, don't remember.
;-)

That's a 16-bit artifact that was abandoned around 1995, since Delphi 2 came
with IB_4.1 that was already 32-bit:
http://www.cvalde.com/IbRoadmap.htm

The alternative is that you're speaking about IB4.00 WorkGroup Server for NT
3.50, a scarce release for the museum. :-) Was it really for 32 bits? In the
meantime, I noticed that several links from my URL show above are now
invalid: Borland deleted or hid any press release before 1998 in the
previous months and the IK subsidiary did the same. Previously, you could
find some old press releases only in Borland UK.

Or maybe you are speaking about ib4.1? It has effectively an ibremote.exe
program, gds32.dll, ibres.dll and was the first 32 bit but the last classic
release on Windows, designed for NT 3.51 and tested with NT 4 beta. I can't
find a reference to ibeng or ibeng32 in it.

Anyway, my point is that your pet ibeng32.DLL was deprecated before 1997
(since IB4.2 appeared circa August 1996) and probably even before 1996, so
you'll have to review the #define's and the like to be sure it's current.

C. (aka the IB paleontologist)
---------
Claudio Valderrama C.
Ingeniero en Informática - Consultor independiente
http://www.cvalde.com - http://www.firebirdSQL.org

[Firebird-devel] Type of nod_udf

[Claudio V. C. <cv...@us...>]

Neil, regarding your latest changes through fb-checkins:

string = (STR) node->nod_arg [e_udf_name];
+ /* how are we supposed to tell which type of nod_udf this is ?? */
+ if (string->str_data[0] == NULL)
+ PRINTF("%s\"\n",((UDF) node->nod_arg[0])->udf_name);
+ else
+ PRINTF ("%s\"\n", string->str_data);

What's the pattern for the two types of UDFs? I thought the name always
comes in a nod_arg at index e_udf_name, but your code proves other fact.
Isn't there another place in the code when the udf name is pushed into the
BLR stream? Since e_udf_name is zero in practice, I understand your problem
is between
char *s = ((STR) node->nod_arg [0])->str_data;
and
char *s = ((UDF)node->nod_arg[0])->udf_name

Strange problem. One possible source could be our dsql parser that might be
generating different nodes. The other explanation is that you're in a map
when you have aggregate functions or sub-selects, have a look at pass1.c
when a name is assigned to each column, there's a long if-else-if
construction.

In practice, you can determine the type of the node by looking at
nod->nod_type (high level type)
and
nod->nod_header.blk_type (what the allocator knows)

The types that the DSQL allocator knows are in dsql\blk.h, you could test
such info; there are type_udf and type_str. I'm still wondering why other
places of the code don't seem to have that same problem.

C.
---------
Claudio Valderrama C.
Ingeniero en Informática - Consultor independiente
http://www.cvalde.com - http://www.firebirdSQL.org

[Firebird-devel] [ firebird-Bugs-451944 ] Trigger Act/Deact increases meta counter

[<no...@so...>]

Bugs item #451944, was opened at 2001-08-16 20:16
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=109028&aid=451944&group_id=9028

Category: Core Engine
Group: Fixed v1.0
>Status: Closed
Resolution: Fixed
Priority: 3
Submitted By: Sean Leyne (seanleyne)
>Assigned to: Ann W. Harrison (awharrison)
Summary: Trigger Act/Deact increases meta counter

Initial Comment:
Activating/Deactivating trigger should not increment 
the metadata counter.

----------------------------------------------------------------------

Comment By: Claudio Valderrama C. (robocop)
Date: 2001-12-09 22:59

Message:
Logged In: YES 
user_id=62823

Ann's changes compare the old and the new table formats and 
only increment the counter if there's a change in the table 
structure. The counter is not incremented with trigger 
creation, deletion, activation or deactivation. Only 
rdb$runtime is refreshed.


----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=109028&aid=451944&group_id=9028

[Firebird-devel] [ firebird-Bugs-233025 ] Server hangs when executing Stored Proc more than once

[<no...@so...>]

Bugs item #233025, was opened at 2001-02-18 15:58
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=109028&aid=233025&group_id=9028

>Category: DSQL
>Group: Fixed v1.0
>Status: Closed
>Resolution: Fixed
Priority: 9
Submitted By: Helen Borrie (helebor)
>Assigned to: Ann W. Harrison (awharrison)
Summary: Server hangs when executing Stored Proc more than once

Initial Comment:
When executing a SP inside a loop that accepts input parameters, the server hangs on the second iteration if the statement is not unprepared and re-prepared between iterations.  This is an old bug which goes back to v. 5.x as well.

Note that, if you add a SUSPEND statement at the end of the SP and return a dummy parameter, you can multi-call the SP with SELECT instead of EXECUTE without the need to do the unprepare/prepare when passing subsequent sets of inputs.

----------------------------------------------------------------------

>Comment By: Claudio Valderrama C. (robocop)
Date: 2001-12-09 23:13

Message:
Logged In: YES 
user_id=62823

Ann posted a temporary fix that stops the confusion in the 
DSQL layer regarding the correct statement to be executed 
with the "immed" API calls.


----------------------------------------------------------------------

Comment By: Nobody/Anonymous (nobody)
Date: 2001-09-25 22:39

Message:
Logged In: NO 

pba...@ho... provided the following testcase
for this bug in 9nmhq1$buk$1...@ne...


Hi,

We have a test case that reproduces the following bug :
http://sourceforge.net/tracker/index.php?
func=detail&aid=233025&group_id=902
8&atid=109028
there is the descritpion of a bug of priority 9 :
"This is an old bug which goes back to v. 5.x as well"

We have a gdb, a program written in C and another in delphi 
that reproduces
the bug:

For this, you can downlad the gdb sample on :
http://www.vienneinfo.org/delphi/zip/3.gdbIBBug.zip
(50 Ko)
and the source of the sample in C :
http://www.vienneinfo.org/delphi/zip/tstc.cpp
(8 Ko)
This is a translation of my Delphi sample that uses IBX 
4.62 :
http://www.vienneinfo.org/delphi/zip/bugib.zip
(52 Ko : Delphi 5 source + gdb sample)


What we know today :

in 3.gdb, there are two stored procedures: SP 002 and SP 011

if we start with SP 002 it works since the last input 
parameter is an
integer.
if we start with SP 011 where there is only one VARCHAR 
input parameter,
then we alternate with SP 002 then SP 011 again => crash.
if we only run SP 011 N times : it works well...

with IB local (without TCP/IP) : no trouble in all cases

with IB Server AND Client on the same computer, with TCP/IP 
=> trouble
The C Sample has got the same trouble on Windows or Linux...

More strange :

With one computer as Server (under Windows) and another 
computer as Client
(under Linux) : No trouble ! :-(
It also works with Linux as server and Windows as client.

Summarry :

client    server   protocol  result
---------------------------------------
windows   windows   tcp/ip      crash
linux     linux     tpc/ip      crash
windows   linux     tcp/ip      ok
linux     windows   tcp/ip      ok
windows   windows   local       ok
linux     linux     local       ok


if you need more information, you can contact me on 
pba...@ho...


kloo


Jean Lacoste, the author of the C sample comments :
When debugging in Linux, when the error happens, the 
debugger displays an
error of
typesegmentation fault at line 61 of memcopy.c (found in 
sysdeps/generic).
It seems that memcpy
is called with its 2nd parameter, sources, set to NULL.



----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=109028&aid=233025&group_id=9028

[Firebird-devel] [ firebird-Bugs-451944 ] Trigger Act/Deact increases meta counter

[<no...@so...>]

Bugs item #451944, was opened at 2001-08-16 20:16
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=109028&aid=451944&group_id=9028

>Category: Core Engine
>Group: Fixed v1.0
Status: Open
>Resolution: Fixed
Priority: 3
Submitted By: Sean Leyne (seanleyne)
Assigned to: Nobody/Anonymous (nobody)
Summary: Trigger Act/Deact increases meta counter

Initial Comment:
Activating/Deactivating trigger should not increment 
the metadata counter.

----------------------------------------------------------------------

>Comment By: Claudio Valderrama C. (robocop)
Date: 2001-12-09 22:59

Message:
Logged In: YES 
user_id=62823

Ann's changes compare the old and the new table formats and 
only increment the counter if there's a change in the table 
structure. The counter is not incremented with trigger 
creation, deletion, activation or deactivation. Only 
rdb$runtime is refreshed.


----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=109028&aid=451944&group_id=9028

[Firebird-devel] [ firebird-Bugs-409769 ] UDF argument can't be query-parameter

[<no...@so...>]

Bugs item #409769, was opened at 2001-03-19 09:03
You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=109028&aid=409769&group_id=9028

Category: DSQL
>Group: Fixed v1.0
>Status: Closed
>Resolution: Fixed
Priority: 5
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Claudio Valderrama C. (robocop)
Summary: UDF argument can't be query-parameter

Initial Comment:
Query kind of:

  SELECT 
     Field1, Field2, ... ,
     UDF1(Field1, ?Param1)
  FROM 
     TABLE1 

don't work - message is "unknown data type"  
         

----------------------------------------------------------------------

>Comment By: Claudio Valderrama C. (robocop)
Date: 2001-12-09 22:22

Message:
Logged In: YES 
user_id=62823

Now it works. For this, the DSQL layer has to load 
information about UDF arguments that previously it didn't 
have available.
With udfs that use descriptors, the recognized parameter 
type is always the formal parameter; the engine cannot 
guess the actual parameter wanted by the user. Either 
redeclare that udf with a different SQL name and use the 
desired formal parameter or try adjusting the scale of the 
input query parameter when the format argument is int and 
the current parameter is numeric(9,4), for example. If a 
UDF is declared to receive INT by descriptor, the engine 
can pass it happily a numeric(p,s) quantity, then it's to 
the udf to test the scale.


----------------------------------------------------------------------

Comment By: Claudio Valderrama C. (robocop)
Date: 2001-12-03 00:14

Message:
Logged In: YES 
user_id=62823

Our syntax allows that. However, the DSQL layer doesn't 
retrieve information about parameters to know what their 
type is. A fix is being tested, but the interaction with 
descriptor-based UDFs will surprise developers.


----------------------------------------------------------------------

You can respond by visiting: 
http://sourceforge.net/tracker/?func=detail&atid=109028&aid=409769&group_id=9028

RE: [Firebird-devel] Aggregate doesn't work inside UDF (new case) more clear

[Christian P. <c_p...@ya...>]

Neil McCalden <nm...@zi...> escribió en el mensaje de noticias
200...@zi......
>
> I think that grouping by fields/udf without them being in the select
> list is a bad idea but as it allowed for straight fields it should be
> allowed for udfs. Making aggregates a special case was where I was
> heading with aggregate_in_list but the udf test was blocking it first in
> this case.
In my opinion grouping by fields/udf without them being in the select list
should be valid, BUT in this case,
not including the grouping in the select was my mistake, this is not the
point here.
I'll try to explain it again, here is my example with a modification:

SELECT CUSTOMER, DIVIDE(ABS(TOTAL),TOTAL), SUM(TOTAL)
FROM SALES
GROUP BY CUSTOMER, DIVIDE(ABS(TOTAL),TOTAL)

this example is working FINE and the result is as spected.
But if you add the ABS to the SUM, to get all values as positive, then you
get the error:

SELECT CUSTOMER, DIVIDE(ABS(TOTAL),TOTAL), ABS(SUM(TOTAL))
FROM SALES
GROUP BY CUSTOMER, DIVIDE(ABS(TOTAL),TOTAL)

here you have an error.

Look that the possible values for the structure DIVIDE(ABS(TOTAL),TOTAL) are
only two "1" if TOTAL > 0 and "-1" if TOTAL > 0. I use this to separate
negative values from positive values.

----

I made some tests and I found that the problem ocurr when the field inside
the aggregate and the field inside the UDF in the GROUP BY is the same. For
a better explain of this, the following example works FINE:

SELECT CUSTOMER, ANYUDF(XFIELD), ABS(SUM(TOTAL))
FROM SALES
GROUP BY CUSTOMER, ANYUDF(XFIELD)

where XFIELD is other than TOTAL

----

Of course this is not a compatibility issue at all, since UDF in GROUP BY
are not allowed in INTERBASE, but it restricts some interesting things that
you could be do with FIREBIRD.

with best regards
Christian

c_p...@ya...

Re: [Firebird-devel] Grant bug

[Neil M. <nm...@zi...>]

On Sat, Dec 08, 2001 at 09:08:03PM -0400, Claudio Valderrama C. wrote:
> >
> > I do not have a fix but I can reproduce the problem with LI-T1.0.0.655
> > Classic and point to the area of code causing the problem which may
> > help.
> 
> 
> Is there any chance that GPRE doesn't work the same depending on classic v/s
> SS? Although Ded found the problem in WI-T6.2.632.
> 

I don't have a SS build around a the moment to try but as it happens on
Win I would doubt it.

> 
> Pay attention to the structs that GPRE defined at the beginning of
> set_field_class_name:
> 
>    struct {
>           TEXT  jrd_12 [32];	/* RDB$SECURITY_CLASS */
>    } jrd_11;
> [snip]
>    struct {
>           TEXT  jrd_20 [32];	/* RDB$SECURITY_CLASS */
>           SSHORT jrd_21;	/* gds__utility */
>           SSHORT jrd_22;	/* gds__null_flag */
>    } jrd_19;
> 

 Both the same here.

> Now, let's see the call:
> 
> >     gds__vtov (jrd_19.jrd_20, jrd_11.jrd_12, 32);
> 
> At least in my case (my generated DYN.C), I see that both source and target
> at 32 bytes.
> 
> >     EXE_start (tdbb, (REQ)request2, gbl->gbl_transaction);
> >     EXE_send (tdbb, (REQ)request2, 0, 32, (UCHAR*)&jrd_11);
> 
> Here, jrd_11 was passed completely, but this struct has only jrd_12 inside
> that a string of 32 bytes. Don't understand what's the supposed mismatch.
> 
[snip end of proc]
> 
> > In EXE_send it has a test which compares the length passed in as a
> > parameter 32 and the expected length of the of the statement in the node
> > it finds in request2 which in this case is 64. The 32 matches the size
> > of jrd_11 so looks correct.
> 
> What's 64, Neil. I can't see it. Does your GPRE generates different struct
> definitions?
> 

In EXE_send

if (length != format->fmt_length)
	ERR_post (gds__port_len,.....
 
with values  32 != 64

In a new isql session this happens on the second field, once failed if
you try running the grant again it fails on the first field.


-- 
Neil McCalden                @home     nm...@zi...     

Re: [Firebird-devel] Aggregate doesn't work inside UDF (new case) more clear

[Neil M. <nm...@zi...>]

On Sun, Dec 09, 2001 at 01:58:54AM -0400, Claudio Valderrama C. wrote:
> > -----Original Message-----
> > From: fir...@li...
> > [mailto:fir...@li...]On Behalf Of
> > Christian Pradelli
> > Sent: Mircoles 5 de Diciembre de 2001 8:07
> >
> > BUT the following combination DOESN'T work:
> >
> > SELECT CUSTOMER, ABS(SUM(TOTAL))
> > FROM SALES
> > GROUP BY CUSTOMER, DIVIDE(ABS(TOTAL),TOTAL)
> > // this is the like the last query but returns all values like positives
> >
> > invalid column reference
> 
> At first glance, it's invalid. The grouping happens on a different UDF than
> the SELECT part. However, what makes the case special is the presence of
> aggregate functions. I regret to accept that's a valid construction. I don't
> know what Neil may suggest, but the only idea I have is:
> 
> - Given a concatenation of udfs, if we hit an aggregate function (count,
> sum, avg, min, max), such branch is game over and we don't worry anymore: no
> invalid reference. In other words,

I think that grouping by fields/udf without them being in the select
list is a bad idea but as it allowed for straight fields it should be
allowed for udfs. Making aggregates a special case was where I was
heading with aggregate_in_list but the udf test was blocking it first in
this case. 

> 
> - Someone may write
> select customer, abs(trunc(sum(total)+field))
> from sales
> group by customer, abs(total)
> that's invalid, but I prefer that this case goes undetected if it's a
> nightmare, because
> select customer, abs(trunc(sum(total)*2+3))
> should be accepted.
> 

Adding a check for that was straight forward as it is already recursing
the item to check for an aggregate and I don't think I have broken
anything else in the process. 


-- 
Neil McCalden                @home     nm...@zi...     

[Firebird-devel] [client-java] Escaped syntax for CallableStatement: need your help/input

[Roman R. <rro...@ya...>]

Hi,

currently I am trying to implement the correct support
of escaped syntax in the callable statement. but,
seems to me that the code i'm trying to get work might
be an overkill for the FB.

Escaped syntax of form {call proc_name(?, ?,...)}
allows both IN and OUT parameters in the call.
Unfortunately, Firebird (and InterBase) do allow only
IN parameters in procedure call, OUT are represented
by the ResultSet. The consequence of this is, that
escaped syntax containing IN and OUT params cannot be
prepared directly.

Since I don't have big experience with stored
procedures throught JDBC, I need your input, how IN
and OUT params look there:

- first go IN, then OUT?
- mixed IN and OUT?
- can constant expression (not "?") be an OUT
parameter?

Thanks in advance,
Roman Rokytskyy

__________________________________________________
Do You Yahoo!?
Buy the perfect holiday gifts at Yahoo! Shopping.
http://shopping.yahoo.com

[Firebird-devel] CRLF2LF [was: isql\show.e spoiled]

[Mike N. <ta...@al...>]

Claudio Valderrama C. wrote:
> 
> Now, could you have pity on me and recommend a good CR stripper, please?

Not good, but the attached one works.

/Mike

RE: [Firebird-devel] ibeng32.dll

[Claudio V. C. <cv...@us...>]

> -----Original Message-----
> From: fir...@li...
> [mailto:fir...@li...]On Behalf Of Dmitri
> Zakharov
>
>   Recently I tried to use ibeng32.dll from FireBird 0.94
>   It was build via MS Visual C++ 6.0.
>   (it was renamed to gds32.dll of cource)
>   as a local data engine.
>   It seems the code contain some int 3 inside. (I am not sure...)

You should know that this "engine in a DLL" has not been used in years.
Don't know how successful it was when it was tested and why it was
abandoned, but certainly it's not in any official Borland binary
distribution.


>   Backup through isc_service behaves the same way (int 3),
>   Though backup file is created correctly.
>
>   Can anybody hint (advise) me a solution of this trouble?

I think that your option is to try to debug it. Have a small C program that
loads the DLL and makes the API calls (you can start from one of the API*.C
examples) and debug it with MSVC, until you jump inside the DLL and the
debugger asks for the source file(s).

C.