Menu
▾
▴
letodb-developers — Leto DB developers
You can subscribe to this list here.
| 2008 |
Jan
|
Feb
(5) |
Mar
(52) |
Apr
(12) |
May
(2) |
Jun
(6) |
Jul
(11) |
Aug
(3) |
Sep
|
Oct
|
Nov
|
Dec
(2) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2009 |
Jan
|
Feb
(2) |
Mar
(2) |
Apr
(20) |
May
(19) |
Jun
(21) |
Jul
(6) |
Aug
(1) |
Sep
|
Oct
(7) |
Nov
(5) |
Dec
(2) |
| 2010 |
Jan
(8) |
Feb
(2) |
Mar
|
Apr
(6) |
May
(5) |
Jun
(26) |
Jul
(14) |
Aug
(5) |
Sep
(3) |
Oct
(3) |
Nov
(4) |
Dec
(1) |
| 2011 |
Jan
(1) |
Feb
(4) |
Mar
(1) |
Apr
(13) |
May
(10) |
Jun
(10) |
Jul
(7) |
Aug
(5) |
Sep
(2) |
Oct
(6) |
Nov
(1) |
Dec
|
| 2012 |
Jan
(4) |
Feb
(8) |
Mar
(6) |
Apr
(1) |
May
(2) |
Jun
(4) |
Jul
(2) |
Aug
(8) |
Sep
(6) |
Oct
(1) |
Nov
(5) |
Dec
(6) |
| 2013 |
Jan
(5) |
Feb
|
Mar
(12) |
Apr
|
May
(1) |
Jun
(2) |
Jul
|
Aug
(3) |
Sep
(4) |
Oct
(2) |
Nov
(6) |
Dec
(27) |
| 2014 |
Jan
(13) |
Feb
(6) |
Mar
(8) |
Apr
(2) |
May
|
Jun
(6) |
Jul
(2) |
Aug
(1) |
Sep
(1) |
Oct
(2) |
Nov
(5) |
Dec
(1) |
| 2015 |
Jan
(6) |
Feb
(7) |
Mar
(7) |
Apr
(16) |
May
(2) |
Jun
(10) |
Jul
(21) |
Aug
(3) |
Sep
|
Oct
(1) |
Nov
|
Dec
|
| 2016 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
(6) |
Oct
(15) |
Nov
(2) |
Dec
(4) |
|
From: Alexander S.K. <al...@be...> - 2008-04-04 08:53:13
|
2008-04-04 12:55 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
+ tests
+ tests/test_ta.prg
+ Added a sample, which demonstrates how to use transactions with
Leto db server
2008-04-04 12:45 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/client/leto1.c
! Additional fix, related to records unlocking after transaction
commit or rollback.
2008-04-04 12:35 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/client/leto1.c
* source/server/letofunc.c
! Bugs fixed, related to transactions
* leto_Rollback([lUnlock]) now unlocks all locked records, if
lUnlock == Nil or .T.
* source/server/server.prg
! Small bug fixed in hb_createtable()
Regards,
Alexander
|
|
From: Alexander S.K. <al...@be...> - 2008-04-03 11:19:20
|
2008-04-03 15:16 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/client/leto1.c
* source/server/letofunc.c
! Few Bugs fixed, related to transactions and records locking
Regards,
Alexander.
|
|
From: Alexander S.K. <al...@be...> - 2008-04-03 06:25:04
|
2008-04-03 10:25 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
+ make_xhb.bat
* Added bat file for xHarbour Builder
* Readme.txt
* Added info about make_xhb.bat
|
|
From: Alexander S.K. <al...@be...> - 2008-04-03 06:13:37
|
2008-04-03 10:05 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/common/hbip.c
* hb_iprecv() and hb_ipsend() has been simplified - unused modes
removed.
* Changing of send/receive buffer, introduced in previous post, is
isolated to separate function hb_ipSetBufSize(), buffer size is
reduced
to 16K. Changing of buffer size is moved from hb_ipServer() to
hb_ipAccept()
* include/funcleto.h
* HB_SENDRECV_BUFFER_SIZE is increased to 16K
* source/client/leto1.c
* source/server/letofunc.c
* Changed calls to hb_iprecv(), hb_ipsend()
Miguel, there is no any need to change a buffer of a socket, created in
hb_ipServer(), because this socket is used to establish connection only
and it doesn't send or receive big amount of data - some small portions
of internal tcp/ip info only. So I moved the changing of a buffer size
to the hb_ipAccept() - this function creates server side sockets, which
really exchanges data with client.
I think that the same fix should be done in xHarbour's hb_inet.c.
Also, I think that the 64K buffer is too large for our purposes, most of
packets are much smaller - so I reduced it to 16K.
Regards,
Alexander.
|
|
From: Miguel A. M. <mig...@ma...> - 2008-04-01 11:39:53
|
2008-04-01 13:36 UTC+0100 Miguel Angel marchuet Frutos (miguelangel/at/marchuet.net) * source/common/hbip.c % increased SND and RCV buffers to 64k and use 1/2 of its capacities to increase speed transfer. In the future this may be configurable settings. Best regards, Miguel Angel Marchuet |
|
From: Miguel A. M. <mig...@ma...> - 2008-03-31 09:24:11
|
We can add on connect
DWORD value;
int len = sizeof(value);
GetSockOpt(SO_SNDBUF, &value, &len);
if (value < 65536)
{
value = 65536;
SetSockOpt(SO_SNDBUF, &value, sizeof(value));
}
Socket->SndBufSize = value;
GetSockOpt(SO_RECBUF, &value, &len);
if (value < 65536)
{
value = 65536;
SetSockOpt(SO_RECBUF, &value, sizeof(value));
}
Socket->RecBufSize = value;
and modify structure:
/* #define HB_SENDRECV_BUFFER_SIZE 1400 */ never use more this
typedef struct tag_HB_SOCKET_STRUCT
{
ULONG sign;
HB_SOCKET_T com;
char *errorDesc;
int errorCode;
struct sockaddr_in remote;
LONG count;
int timeout;
int timelimit;
PHB_ITEM caPeriodic;
DWORD SndBufSize;
DWORD RecBufSize;
} HB_SOCKET_STRUCT;
Alexander S.Kresin escribió:
> Miguel Angel Marchuet writes:
>
>> compiler uses hardcoded
>> #define HB_SENDRECV_BUFFER_SIZE 1400
>> instead of use
>> getsockopt (Windows Sockets) to retrieve the value of socket option SO_MAX_MSG_SIZE
>> It makes very slow some transactions, for example FTP, Leto RDD, etc.
>> Any body knows why we are using hardcoded 1400 ?.
>>
> I left it in hbip.c as it was in hbinet.c, because I supposed that
> someone who introduced this, had reasons for it, and I simply hadn't
> time yet to review this.
>
>> We can change this ?
>>
>
> Sure.
>
> Regards,
> Alexander.
>
> -------------------------------------------------------------------------
> Check out the new SourceForge.net Marketplace.
> It's the best place to buy or sell services for
> just about anything Open Source.
> http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace
> _______________________________________________
> Letodb-developers mailing list
> Let...@li...
> https://lists.sourceforge.net/lists/listinfo/letodb-developers
>
|
|
From: Alexander S.K. <al...@be...> - 2008-03-31 08:01:25
|
Miguel Angel Marchuet writes: > compiler uses hardcoded > #define HB_SENDRECV_BUFFER_SIZE 1400 > instead of use > getsockopt (Windows Sockets) to retrieve the value of socket option SO_MAX_MSG_SIZE > It makes very slow some transactions, for example FTP, Leto RDD, etc. > Any body knows why we are using hardcoded 1400 ?. > I left it in hbip.c as it was in hbinet.c, because I supposed that someone who introduced this, had reasons for it, and I simply hadn't time yet to review this. > > We can change this ? > Sure. Regards, Alexander. |
|
From: Miguel A. M. <mig...@ma...> - 2008-03-31 07:47:30
|
compiler uses hardcoded #define HB_SENDRECV_BUFFER_SIZE 1400 instead of use getsockopt (Windows Sockets) to retrieve the value of socket option SO_MAX_MSG_SIZE It makes very slow some transactions, for example FTP, Leto RDD, etc. Any body knows why we are using hardcoded 1400 ?. We can change this ? best regards, Miguel Angel marchuet |
|
From: Alexander S.K. <al...@be...> - 2008-03-28 10:40:21
|
2008-03-28 13:40 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/server/letofunc.c
* source/client/leto1.c
* Transactions support is ready. Not properly tested yet, but simplest
tests works.
2008-03-28 09:40 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/server/letofunc.c
* source/client/leto1.c
* Continued implementation of transactions support.
Regards,
Alexander
|
|
From: Miguel A. M. <mig...@ma...> - 2008-03-28 10:22:01
|
HB_FT_TIMESTAMP and HB_FT_DAYTIME are the same type "@" but HB_FT_MODTIME is type "=" used in files with extension ADT (ADS) and HB_DATETIME is type "T" used in files with extension DBF (VFP) "=" was incorrectly introduced, this type was borrowed from ADT files to DBF files. but is not supported make an incompatible DBF for the rest of compilers or applications. Sometimes i explain very bad :( Best regards, Miguel Angel Marchuet > |
|
From: Pavel T. <tp...@ma...> - 2008-03-27 21:17:00
|
Hello, Miguel! MAM> at xharbour MAM> HB_FT_DAYTIME is named HB_FT_TIMESTAMP MAM> HB_FT_MODTIME exists yet MAM> HB_FT_TIME exists yet (but incompatible with DBF at harbour) MAM> HB_FT_ROWVER exists yet MAM> HB_FT_AUTOINC exists yet MAM> HB_FT_ANY exists yet (but incompatible with DBF at harbour) MAM> this 2 FT only exist at RDD with the purpose to read inexistent DBF MAM> ar fields of ADT created with ADSADT but not are DBFCDX standard MAM> compatible DBF. MAM> instead of MAM> HB_FT_TIMESTAMP and HB_FT_DATETIME MAM> of xharbour are VFP complatible MAM> HB_FT_ANY MAM> of ADSADT breaks compatibility with DBF because V meanings MAM> "VarLenght" ar DBF and "Any" at ADT MAM> HB_FT_DAYTIME is named HB_FT_TIMESTAMP MAM> HB_FT_MODTIME exists yet MAM> HB_FT_TIME exists yet MAM> HB_FT_ROWVER exists yet MAM> HB_FT_AUTOINC exists yet MAM> HB_FT_ANY exists yet MAM> This types were borrowed from ADT format, but are not DBF format MAM> instead of MAM> HB_FT_DATETIME MAM> HB_FT_TIMESTAMP MAM> HB_FT_CURRENCY MAM> HB_FT_VARLENGTH MAM> HB_FT_INTEGER MAM> HB_FT_DOUBLE MAM> HB_FT_FLOAT MAM> HB_FT_OLE MAM> HB_FT_PICTURE MAM> HB_FT_STRING (binary and normal) MAM> HB_FT_MEMO (binary and normal) MAM> I think that the implementation of xharbour respect all formats MAM> At xharbour autoincremental (flag and FT) works properly, but not MAM> at harbour. IMHO this FT identifiers is used only under letodb. HB_FT_DATETIME and HB_FT_TIMESTAMP is the synonym of HB_FT_DAYTIME and HB_FT_MODTIME With best regards, Pavel Tsarenko. |
|
From: Pavel T. <tp...@ma...> - 2008-03-27 21:16:51
|
2008-03-27 21:30 UTC+0200 Pavel Tsarenko (tpe2/at/mail.ru) * source/client/leto1.c ! fixed several bugs for binary fields + add missing SELF_SYNCCHILDREN for letoGoTop and letoGoBottom + added letoSetBlankRecord function. It's need to fill record by empty values before commit operation With best regards, Pavel Tsarenko. |
|
From: Alexander S.K. <al...@be...> - 2008-03-27 13:32:17
|
2008-03-27 12:44 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/server/letofunc.c
* source/client/leto1.c
* Continued implementation of transactions support.
New functions has been added:
leto_BeginTransaction()
leto_RollBack()
leto_CommitTransaction()
Regards,
Alexander.
|
|
From: Miguel A. M. <mig...@ma...> - 2008-03-27 11:00:31
|
at xharbour HB_FT_DAYTIME is named HB_FT_TIMESTAMP HB_FT_MODTIME exists yet HB_FT_TIME exists yet (but incompatible with DBF at harbour) HB_FT_ROWVER exists yet HB_FT_AUTOINC exists yet HB_FT_ANY exists yet (but incompatible with DBF at harbour) this 2 FT only exist at RDD with the purpose to read inexistent DBF ar fields of ADT created with ADSADT but not are DBFCDX standard compatible DBF. instead of HB_FT_TIMESTAMP and HB_FT_DATETIME of xharbour are VFP complatible HB_FT_ANY of ADSADT breaks compatibility with DBF because V meanings "VarLenght" ar DBF and "Any" at ADT HB_FT_DAYTIME is named HB_FT_TIMESTAMP HB_FT_MODTIME exists yet HB_FT_TIME exists yet HB_FT_ROWVER exists yet HB_FT_AUTOINC exists yet HB_FT_ANY exists yet This types were borrowed from ADT format, but are not DBF format instead of HB_FT_DATETIME HB_FT_TIMESTAMP HB_FT_CURRENCY HB_FT_VARLENGTH HB_FT_INTEGER HB_FT_DOUBLE HB_FT_FLOAT HB_FT_OLE HB_FT_PICTURE HB_FT_STRING (binary and normal) HB_FT_MEMO (binary and normal) I think that the implementation of xharbour respect all formats At xharbour autoincremental (flag and FT) works properly, but not at harbour. Best regards, Miguel Angel Marchuet Pavel Tsarenko escribió: > 2008-03-26 21:20 UTC+0200 Pavel Tsarenko (tpe2/at/mail.ru) > * include/funcleto.h > * source/server/letofunc.c > * source/client/leto1.c > + added support for additional field types: > HB_FT_INTEGER > HB_FT_DOUBLE > HB_FT_FLOAT > HB_FT_DAYTIME > HB_FT_MODTIME > + added hb_strncpyLower function for xHarbour > * source/server/server.prg > + added requests for codepages > * small change in hs_createindex > > With best regards, Pavel Tsarenko > > > ------------------------------------------------------------------------- > Check out the new SourceForge.net Marketplace. > It's the best place to buy or sell services for > just about anything Open Source. > http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace > _______________________________________________ > Letodb-developers mailing list > Let...@li... > https://lists.sourceforge.net/lists/listinfo/letodb-developers > |
|
From: admin <al...@be...> - 2008-03-27 10:22:50
|
2008-03-27 12:44 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/client/leto1.c
* Started implementation of transactions support
2008-03-27 10:30 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/common/hbip.c
* Added SO_KEEPALIVE option to the server side socket.
* Whatsnew.txt
* Added build 6 info.
2008-03-27 10:25 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/client/leto1.c
* hb_strncpyLower is replaced by other functions
|
|
From: Pavel T. <tp...@ma...> - 2008-03-26 20:21:30
|
2008-03-26 21:20 UTC+0200 Pavel Tsarenko (tpe2/at/mail.ru) * include/funcleto.h * source/server/letofunc.c * source/client/leto1.c + added support for additional field types: HB_FT_INTEGER HB_FT_DOUBLE HB_FT_FLOAT HB_FT_DAYTIME HB_FT_MODTIME + added hb_strncpyLower function for xHarbour * source/server/server.prg + added requests for codepages * small change in hs_createindex With best regards, Pavel Tsarenko |
|
From: Alexander S.K. <al...@be...> - 2008-03-26 07:41:29
|
2008-03-26 10:32 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/server/server.prg
! Bug fixed
* source/client/leto1.c
* ordBagExt() now returns ".NTX" if the dbf is opened via "DBFNTX"
Regards,
Alexander
|
|
From: Alexander S.K. <al...@be...> - 2008-03-26 06:39:31
|
2008-03-26 09:38 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/server/letofunc.c
* source/server/server.prg
! Few bugs fixed
* Now it is possible to define [DATABASE] section and the following
options for it:
DataPath = some_path
Driver = CDX/NTX
Regards,
Alexander
|
|
From: Alexander S.K. <al...@be...> - 2008-03-25 18:55:51
|
2008-03-25 21:55 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/server/letofunc.c
* DATABASE structures and appropriate code has been added.
Regards,
Alexander
http://kresin.belgorod.su
|
|
From: Alexander S.K. <al...@be...> - 2008-03-25 17:10:38
|
2008-03-25 20:15 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* include/rddleto.h
* source/client/leto1.c
* source/server/letofunc.c
* source/server/server.prg
* Support of DBFNTX has been added.
For now, to force Letodb work via DBFNTX driver, you need to write
in letodb.ini: DEFAULT_DRIVER = NTX
2008-03-25 17:10 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/client/leto1.c
! Small fix
Regards,
Alexander
http://kresin.belgorod.su
|
|
From: Alexander S.K. <al...@be...> - 2008-03-25 13:51:47
|
2008-03-25 16:50 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* include/funcleto.h
* source/client/leto1.c
* source/server/letofunc.c
* source/server/server.prg
* letoOrderListClear() has been implemented.
2008-03-25 12:00 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/client/leto1.c
* source/server/letofunc.c
* Changed LETOTAG structure
* All bagnames and tagnames are kept now in lower case
Regards,
Alexander.
|
|
From: Alexander S.K. <al...@be...> - 2008-03-24 08:33:01
|
2008-03-24 11:30 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/client/leto1.c
* some changes in the protocol:
SET DELETED flag is transferred now in all move commands instead of
'intro' and 'set;01;'
* source/client/letomgmn.c
* leto_setDeleted() has been removed.
* source/server/letofunc.c
* changes, related to SET DELETED handling
* leto_goto(), leto_skip() and leto_seek() has been changed to use the
leto_getParam() function.
|
|
From: Alexander S.K. <al...@be...> - 2008-03-24 06:24:14
|
2008-03-24 09:24 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/server/letofunc.c
* Some optimizations in leto_goto() and leto_skip()
Regards,
Alexander.
|
|
From: Alexander S.K. <al...@be...> - 2008-03-23 16:02:13
|
2008-03-23 19:00 UTC+0300 Alexander Kresin (alex/at/belacy.belgorod.su)
* source/server/letofunc.c
! Bugs fixed in leto_goto() and leto_SetFilter()
* leto_SendAnswer() has been optimized.
Regards,
Alexander
http://kresin.belgorod.su
|
|
From: Pavel T. <tp...@ma...> - 2008-03-23 07:24:27
|
2008-03-23 09:30 UTC+0200 Pavel Tsarenko (tpe2/at/mail.ru) * source/server/letofunc.c ! fixed bug in leto_Skip if filter is defined % memory allocations has been optimized in leto_Skip With best regards, Pavel Tsarenko. |
15 messages has been excluded from this view by a project administrator.
