> but I see what hapened
> Whe should decide about buffer alignment
> I can use "packed record" instead of "record" then data are as is
> But what is your buffer alignment when you fill data in that record?
>
>>Search for type_alignments:
>>https://github.com/FirebirdSQL/core/blob/master/src/jrd/a
Adriano,
could you check also inheritance in classes in FBApi
Is this ok that base class have function implemented?
In child class this function is redeclared as virtual; abstract
I suppose that this should be in reverse order
In base class function as virtual; abstract and in child implemented a
17.03.2015 9:06, liviusliv...@poczta.onet.pl wrote:
> i tried all alignments Byte, Word, DWord, QWord
Alignments in the buffer are irregular, you can't get it right with these
options. You
have to use packed record and byte arrays as fillers.
--
WBR, SD.
17.03.2015 10:06, liviusliv...@poczta.onet.pl wrote:
>> but I see what hapened
>> Whe should decide about buffer alignment
>> I can use "packed record" instead of "record" then data are as is
>> But what is your buffer alignment when you fill data in that record?
>>
>
>>> Search for type_alignments
17.03.2015 9:06, liviusliv...@poczta.onet.pl wrote:
> i tried all alignments Byte, Word, DWord, QWord
>> Alignments in the buffer are irregular, you can't get it right with
>> these options. You
>>have to use packed record and byte arrays as fillers.
>>--
>> WBR, SD.
Hi,
but how to work w
>> After re-reading docs
>>http://docwiki.embarcadero.com/RADStudio/XE6/en/Internal_Data_Formats#Record_Types
>> i'd say you should not use nor special(explicit) alignment nor packed
>>records. Just ensure you have {$A+} when record type is declared and
>>you use record members with data types
17.03.2015 10:48, liviusliv...@poczta.onet.pl wrote:
> but how to work with this? I should have FB source code?
Yes.
> I supposed that this appi was for simplify usage.
You supposed wrong.
--
WBR, SD.
--
Dive
> but how to work with this? I should have FB source code?
>> Yes.
>>--
>> WBR, SD.
Can you tell me where in FB source is info about this?
regards,
Karol Bieniaszewski
--
Dive into the World of Parallel Progra
17.03.2015 11:54, liviusliv...@poczta.onet.pl wrote:
>>>After re-reading docs
>
>>> http://docwiki.embarcadero.com/RADStudio/XE6/en/Internal_Data_Formats#Record_Types
>
>>> i'd say you should not use nor special(explicit) alignment nor packed
>>> records. Just ensure you have {$A+} when recor
-Oryginalna wiadomość-
From: Vlad Khorsun
Sent: Tuesday, March 17, 2015 11:24 AM
To: firebird-devel@lists.sourceforge.net
Subject: Re: [Firebird-devel] Firebird 3 API demo usage with Pascal
17.03.2015 11:54, liviusliv...@poczta.onet.pl wrote:
>>>After re-reading docs
>
>>> http://docw
17.03.2015 11:15, liviusliv...@poczta.onet.pl wrote:
> Can you tell me where in FB source is info about this?
You should look at fbutil::sqlTypeToDSC() routine.
--
WBR, SD.
--
Dive into the World of Parallel Progr
If you use {$A+} all fields should appear to aligned to 8 byte boundaries
(The default).
I think would be easier to make it controllable, to use packed record and
add some "Foo Bar" padding variables if needed.
If the original structure is continuous then packed record with correct
field sizes s
Hello Dimitry,
On 17.03.2015 8:30, Dmitry Yemanov wrote:
> 17.03.2015 01:02, Lauri Zoova wrote:
>>
>> After further investigation, it turns out the problem is NOT with
>> context values getting lost but rather in the ON CONNECT database
>> trigger not firing when connecting. Furthermore, so far it
17.03.2015 12:45, liviusliv...@poczta.onet.pl wrote:
> i assign it to whole record not for member by member
So, you don't know what values should be in what member ?
> but how offsets can hange any in this representation?
It show us exactly how compiler works
> I should do for one integer
Hi Alex,
please can you take your tpcc.fdb, backup, 7zip and send to me?
I will use as reference database.
Thank you.
Slavek
Ing. Slavomir Skopalik
Executive Head
Elekt Labs s.r.o.
Collection and evaluation of data from machines and laboratories
by means of system MASA (http://www.elektlabs.cz
17.03.2015 14:37, Vlad Khorsun wrote:
>> >think about this simplified record declaration - how will you declare this?
>> >
>> >GenRowsInMessage = record
>> >iValue: Integer;
>> >iValueNull: WordBool;
>> >sValue: SmallInt;
>> >sValueNull: WordBool;
>> >iValue2: Integer;
>> >iValue2Null: WordBool;
>>
On 03/17/15 16:50, Slavomir Skopalik wrote:
> Hi Alex,
> please can you take your tpcc.fdb, backup, 7zip and send to me?
> I will use as reference database.
>
Done.
Confirm that you've received it pls.
--
Dive into the
On 17-03-2015 06:40, Vlad Khorsun wrote:
> 17.03.2015 10:06, liviusliv...@poczta.onet.pl wrote:
>>> but I see what hapened
>>> Whe should decide about buffer alignment
>>> I can use "packed record" instead of "record" then data are as is
>>> But what is your buffer alignment when you fill data in t
On 17-03-2015 05:06, liviusliv...@poczta.onet.pl wrote:
>> but I see what hapened
>> Whe should decide about buffer alignment
>> I can use "packed record" instead of "record" then data are as is
>> But what is your buffer alignment when you fill data in that record?
>>
>
>>> Search for type_alignm
>>I think it must be more or less like Win32 records declared in Pascal.
>>They follow standard C alignment rules, so as Firebird.
>>Adriano
Hi,
and how it look in C Win 64 - because i use Win64 compiler and 64bit
fbclient.dll
PS. did you look at inheritance in fbapi?
regards,
Karol Bieniaszew
> i assign it to whole record not for member by member
>> So, you don't know what values should be in what member ?
i know exactly what should be but
> but how offsets can hange any in this representation?
>> It show us exactly how compiler works
this as you say mamory dump was not from recor
-Oryginalna wiadomość-
From: Dimitry Sibiryakov
Sent: Tuesday, March 17, 2015 11:47 AM
To: For discussion among Firebird Developers
Subject: Re: [Firebird-devel] Firebird 3 API demo usage with Pascal
17.03.2015 11:15, liviusliv...@poczta.onet.pl wrote:
> Can you tell me where in FB source
On 17-03-2015 13:11, liviusliv...@poczta.onet.pl wrote:
>>> I think it must be more or less like Win32 records declared in Pascal.
>>> They follow standard C alignment rules, so as Firebird.
>>> Adriano
>
> Hi,
>
> and how it look in C Win 64 - because i use Win64 compiler and 64bit
> fbclient.d
>>"virtual; abstract;" methods in Impl classes hides concrete
>>implementation from non-impl classes.
>>Concrete implementations calls the virtual methods.
>>This way we can have standard classes and you can override the real
>>methods, not ones with fake names.
>>Adriano
Ok if this is as desi
Messages "Error in isc_release_request() ... when working with legacy security
database" appear in firebird.log for CLASSIC server when connecting with
'AuthClient = Legacy_Auth' in client-side firebird.conf
-
>>It would be easy if you share your current code in a github.
>>Adriano
Nothing to share for this little problem ;-)
try this:
//bufor
GenRowsInMessage = Array[0..100] of Byte;
//proc in db
create procedure gen_rows_pascal (
iValue integer not null,
sValue Smallint not null,
i
17.03.2015 23:08, liviusliv...@poczta.onet.pl wrote:
>
> iValue integer not null,
> sValue Smallint not null,
> iValue2 integer not null,
> iValue3 integer not null
>
> ($CC, $BB, $AA, $11, 0, 0, $DD, $33, 0, 0, 0, 0, $DD, $CC, $BB, $22, 0, 0,
> 0, 0, $DD, $CC, $BB, $22, 0, 0, 0
17.03.2015 22:08, liviusliv...@poczta.onet.pl wrote:
>>> It would be easy if you share your current code in a github.
>
>
>>> Adriano
>
> Nothing to share for this little problem ;-)
> try this:
>
> //bufor
> GenRowsInMessage = Array[0..100] of Byte;
>
> //proc in db
> create procedure gen_row
$CC, $BB, $AA, $11 -- iValue (aligned at 4-bytes boundary)
0, 0, -- null flag for iValue (aligned at 2-bytes boundary)
$DD, $33 -- sValue (aligned at 2-bytes boundary)
0, 0, -- null flag for sValue (aligned at 2-bytes boundary)
0, 0, -- padding to align iValue2 properly <= i understand why here we
>>$CC, $BB, $AA, $11 -- iValue (aligned at 4-bytes boundary)
>>0, 0, -- null flag for iValue (aligned at 2-bytes boundary)
>>$DD, $33 -- sValue (aligned at 2-bytes boundary)
>>0, 0, -- null flag for sValue (aligned at 2-bytes boundary)
>>0, 0, -- padding to align iValue2 properly <= i understand wh
17.03.2015 21:30, Vlad Khorsun wrote:
> Looks like 2-byte null-indicator's after every input parameter
Yes, message buffer contains null indicator for each value, including those
which are
not nullable.
--
WBR, SD.
I started writing this email (in quotes) and i saw that results was
illogical form record structure POV
"Hi,
i do not understand what happened
this is memory dump from record
$CC $BB $AA $11 $00 $00 $DD $33 $00 $00 $00 $00 $DD $CC $BB $22 $00 $00 $00
$00 $DD $CC $BB $22 $00 $00 $00 $00 $00 $00
Hi Alex,
I made full restore of tpcc.fbk with loging of each compresion request.
Result is here:
http://www.elektlabs.cz/tpcc.7z
Record size is original length of request for compresion.
New RLE is RLE that I was developed for FB3, but was rejected.
LZ4 is fast version, not HC, just original d
33 matches
Mail list logo
