Menu
▾
▴
firebird-devel
|
From: Tagangout, A. <Abd...@me...> - 2010-09-29 15:54:02
|
Hi Dmitry, many thanks for your answere. i find the location where the communication with the TCP-Driver will be solved. 1- When i try only a attach-Command to an Firebird Server from a client, where will be the attach-String saved to be send to the TCP-Driver? 2- Is there any Documentation for this issue? How the isql Commands are Managed in a Packet, that would be transmitted to the TCP-Driver? Many thanks Abdo MEP Kutsch GmbH & Co. KG Flaßkuhle 1 58452 Witten Tel. : 02302 28262-41 Fax.: 02302 28262-99 mailto:abd...@me... http://www.mep.de -----Ursprüngliche Nachricht----- Von: Dimitry Sibiryakov [mailto:sd...@ib...] Gesendet: Mittwoch, 29. September 2010 12:36 An: For discussion among Firebird Developers Betreff: Re: [Firebird-devel] Porting Firebird 29.09.2010 12:28, Tagangout, Abderrahim wrote: > Can you make only one or two changes on this file so i can folow them? Or tell me where should i make this changes and to give the Y-valve only the remote provider? It is done automatically if macro SUPERCLIENT is defined during build. This macro is defined in Makefile.libfbclient. > sorry for this banal question, but what does mean remote-provider? This is Firebird module which provides ability to work with remote Firebird server. Mostly placed in src/remote dir. -- SY, SD. ------------------------------------------------------------------------------ Start uncovering the many advantages of virtual appliances and start using them to simplify application deployment and accelerate your shift to cloud computing. http://p.sf.net/sfu/novell-sfdev2dev Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel |
|
From: Dimitry S. <sd...@ib...> - 2010-09-29 16:11:58
|
29.09.2010 17:53, Tagangout, Abderrahim wrote: > 1- When i try only a attach-Command to an Firebird Server from a client, where will be the attach-String saved to be send to the TCP-Driver? Nowhere. It is parsed and host-part used for connect. DB-part is sent to server in op_attach packet. > 2- Is there any Documentation for this issue? How the isql Commands are Managed in a Packet, that would be transmitted to the TCP-Driver? Isql commands are managed by isql completely on client side. Isql sends SQL commands to server using standard FB API. -- SY, SD. |
|
From: Tagangout, A. <Abd...@me...> - 2010-09-29 16:20:54
|
Hi Dmitry, > It is parsed and host-part used for connect. DB-part is sent to server in op_attach packet. Sorry but Where can find Documentation about attach-String or connect-String? Many thanks Abdo MEP Kutsch GmbH & Co. KG Flaßkuhle 1 58452 Witten Tel. : 02302 28262-41 Fax.: 02302 28262-99 mailto:abd...@me... http://www.mep.de -----Ursprüngliche Nachricht----- Von: Dimitry Sibiryakov [mailto:sd...@ib...] Gesendet: Mittwoch, 29. September 2010 18:11 An: For discussion among Firebird Developers Betreff: Re: [Firebird-devel] Porting Firebird 29.09.2010 17:53, Tagangout, Abderrahim wrote: > 1- When i try only a attach-Command to an Firebird Server from a client, where will be the attach-String saved to be send to the TCP-Driver? Nowhere. It is parsed and host-part used for connect. DB-part is sent to server in op_attach packet. > 2- Is there any Documentation for this issue? How the isql Commands are Managed in a Packet, that would be transmitted to the TCP-Driver? Isql commands are managed by isql completely on client side. Isql sends SQL commands to server using standard FB API. -- SY, SD. ------------------------------------------------------------------------------ Start uncovering the many advantages of virtual appliances and start using them to simplify application deployment and accelerate your shift to cloud computing. http://p.sf.net/sfu/novell-sfdev2dev Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel |
|
From: Enrico W. <we...@me...> - 2010-09-30 17:46:47
|
Hi folks, IMHO it would be a better approach to fork into new branch and step by step trim down until only the client library ist left. (that fork will be frequently rebased to upstream to keep up). *After* then, care about crosscompiling. cu -- ---------------------------------------------------------------------- Enrico Weigelt, metux IT service -- http://www.metux.de/ phone: +49 36207 519931 email: we...@me... mobile: +49 151 27565287 icq: 210169427 skype: nekrad666 ---------------------------------------------------------------------- Embedded-Linux / Portierung / Opensource-QM / Verteilte Systeme ---------------------------------------------------------------------- |
|
From: Enrico W. <we...@me...> - 2010-10-01 17:47:57
|
* Alex Peshkoff <pes...@ma...> schrieb: Hi, > Almost it is done (but not committed yet) in trunk. yValve is present > (but actually this is no danger), but the overall process of library > build became absolutely clear. where can I get this tree ? > And (what should be most interesting from > this thread POV) library is built as the first step, right after > building gpre_boot. > > *After* then, care about crosscompiling. > It can looks as follows: > 1. Build host gpre_boot (can be avoided if one needs no blobs and arrays). > 2. Preprocess 2 epp files. > 3. Cross-build firebird library. What exactly is gpre_boot required for ? Is it still so, that preprocessing the .epp files requires an running database instance ? cu -- ---------------------------------------------------------------------- Enrico Weigelt, metux IT service -- http://www.metux.de/ phone: +49 36207 519931 email: we...@me... mobile: +49 151 27565287 icq: 210169427 skype: nekrad666 ---------------------------------------------------------------------- Embedded-Linux / Portierung / Opensource-QM / Verteilte Systeme ---------------------------------------------------------------------- |
|
From: Alex P. <pes...@ma...> - 2010-10-04 07:18:15
|
On 10/01/10 21:41, Enrico Weigelt wrote: > * Alex Peshkoff <pes...@ma...> schrieb: > > Hi, > >> Almost it is done (but not committed yet) in trunk. yValve is present >> (but actually this is no danger), but the overall process of library >> build became absolutely clear. > where can I get this tree ? > It will be committed in trunk soon (you know we changed from CVS to SVN). If you want it as fast as possible, just let me know - I can send it to you privately. >> And (what should be most interesting from >> this thread POV) library is built as the first step, right after >> building gpre_boot. >>> *After* then, care about crosscompiling. >> It can looks as follows: >> 1. Build host gpre_boot (can be avoided if one needs no blobs and arrays). >> 2. Preprocess 2 epp files. >> 3. Cross-build firebird library. > What exactly is gpre_boot required for ? Some high-level ISC API calls (blobs and arrays support) are implemented using gdml. > Is it still so, that preprocessing the .epp files requires an running > database instance ? > For engine build process we do not need to preprocess .epp files with anything else except gpre_boot, which does not require database instance. But to build some plugins (like security database management) and some utilities (qli) full gpre and therefore running instance are required. Must say that both can be rewritten to use SQL. |
|
From: Alex P. <pes...@ma...> - 2010-09-30 13:48:17
|
On 09/29/10 13:53, Dimitry Sibiryakov wrote: > As I already said it is possible to bypass Y-valve and build client > containing only > "remote" provider, but for current codebase it requires quite a few modification (which is > mostly copy-paste). > This will not be what we call 'isc API' now. It's missing all functions, related with SQLDA (thks to Borland). I doubt that people wish to use only messages API on AVR. |
|
From: Dimitry S. <sd...@ib...> - 2010-09-30 14:12:38
|
30.09.2010 15:48, Alex Peshkoff wrote: >> "remote" provider, but for current codebase it requires quite a few modification (which is >> mostly copy-paste). >> > > This will not be what we call 'isc API' now. It's missing all functions, > related with SQLDA (thks to Borland). I doubt that people wish to use > only messages API on AVR. Exactly these copy-paste from why.cpp I had on mind. Couple of functions, containing calls of UTLD_parse_sqlda() should be enough for topic starter who wanted only "connect to database and get some information from it". BTW, if call of one of parse_sqlda functions fail, sqlda_sup leaks. To fix it, call of release_dsql_support must be moved into sqlda_sup destructor. -- SY, SD. |
|
From: Alex P. <pes...@ma...> - 2010-10-01 07:34:04
|
On 09/30/10 18:12, Dimitry Sibiryakov wrote: > > BTW, if call of one of parse_sqlda functions fail, sqlda_sup leaks. To fix it, call of > release_dsql_support must be moved into sqlda_sup destructor. > Please add it to the tracker. |
|
From: Alex P. <pes...@ma...> - 2010-10-01 07:41:14
|
On 09/29/10 18:17, Enrico Weigelt wrote: > Hi folks, > > > IMHO it would be a better approach to fork into new branch and > step by step trim down until only the client library ist left. > (that fork will be frequently rebased to upstream to keep up). > Almost it is done (but not committed yet) in trunk. yValve is present (but actually this is no danger), but the overall process of library build became absolutely clear. And (what should be most interesting from this thread POV) library is built as the first step, right after building gpre_boot. > *After* then, care about crosscompiling. It can looks as follows: 1. Build host gpre_boot (can be avoided if one needs no blobs and arrays). 2. Preprocess 2 epp files. 3. Cross-build firebird library. Seems trivial, specially if one agrees to skip steps 2 and 3 - i.e. for blob-less client one can simply start cross build of library. I will commit this soon, and may be client from version 3 will be the most simple solution. |
|
From: Tagangout, A. <Abd...@me...> - 2010-10-01 08:00:21
|
Hi Alex, Many thanks, and it will be vary beutiful if there is a Documentation for this issue. Abdo MEP Kutsch GmbH & Co. KG Flaßkuhle 1 58452 Witten Tel. : 02302 28262-41 Fax.: 02302 28262-99 mailto:abd...@me... http://www.mep.de -----Ursprüngliche Nachricht----- Von: Alex Peshkoff [mailto:pes...@ma...] Gesendet: Freitag, 1. Oktober 2010 09:41 An: fir...@li... Betreff: Re: [Firebird-devel] Porting Firebird On 09/29/10 18:17, Enrico Weigelt wrote: > Hi folks, > > > IMHO it would be a better approach to fork into new branch and step by > step trim down until only the client library ist left. > (that fork will be frequently rebased to upstream to keep up). > Almost it is done (but not committed yet) in trunk. yValve is present (but actually this is no danger), but the overall process of library build became absolutely clear. And (what should be most interesting from this thread POV) library is built as the first step, right after building gpre_boot. > *After* then, care about crosscompiling. It can looks as follows: 1. Build host gpre_boot (can be avoided if one needs no blobs and arrays). 2. Preprocess 2 epp files. 3. Cross-build firebird library. Seems trivial, specially if one agrees to skip steps 2 and 3 - i.e. for blob-less client one can simply start cross build of library. I will commit this soon, and may be client from version 3 will be the most simple solution. ------------------------------------------------------------------------------ Start uncovering the many advantages of virtual appliances and start using them to simplify application deployment and accelerate your shift to cloud computing. http://p.sf.net/sfu/novell-sfdev2dev Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel |
|
From: Enrico W. <we...@me...> - 2010-09-27 03:07:20
|
* Alex Peshkoff <pes...@ma...> schrieb: Hi folks, > This definitely means that you've built it not with cross toolchain, but > with native one. How should it be fixed - no idea, sorry. As already discussed @ telco w/ Abdo this afternoon, the way to make a downstream fork that builds _only_ the client library, using the properly passed target (not host) toolchain ($CC etc). This new package (which will be _rebased_ to subsequent upstream releases) then can be cleanly built via sysroot'ed crosscompiler. (no, you surely don't want crosscompiling w/o sysroot ;-p) I'm not up-to-date, whether certain .epp -> .cpp file generation still requires a running database instance. In this case we'll have to fix it (such circular dependencies are clearly an major misdesign). A flat text file as primary data source and the generator written in shellscript or perl will suffice. cu -- ---------------------------------------------------------------------- Enrico Weigelt, metux IT service -- http://www.metux.de/ phone: +49 36207 519931 email: we...@me... mobile: +49 151 27565287 icq: 210169427 skype: nekrad666 ---------------------------------------------------------------------- Embedded-Linux / Portierung / Opensource-QM / Verteilte Systeme ---------------------------------------------------------------------- |
|
From: Adriano d. S. F. <adr...@gm...> - 2010-09-24 10:41:58
|
On 24/09/2010 03:33, Alex Peshkoff wrote: > On 09/23/10 18:10, Tagangout, Abderrahim wrote: >> >> Hi Alex, >> i try what you say, but i become an error. >> >> ../gen/firebird/bin/gpre_current -r -m -z -n ../src/dsql/array.epp ../temp/superclient/dsql/array.cpp >> make: ../gen/firebird/bin/gpre_current: Kommando nicht gefunden >> make: *** [../temp/superclient/dsql/array.cpp] Fehler 127 >> abdo@ubuntu804icnova:~/Firebird-2.1.3.18185-0/gen$ >> >> The error say, that the command "gpre_current" not found. don't forget that i delete all compilate under /gen/firebird/bin/* > Yes, I see. I've supposed that preprocessed (converted by gpre from epp > to cpp) files will remain here. Unfortunately, it's not staying as before. Seems that some generic makefile rule makes them to be generated as temporary targets. I even put something about secondary rules in the makefile that in a first moment appeared to fix it. Later I see nothing probably was changed. Adriano |
|
From: Alex P. <pes...@ma...> - 2010-09-24 12:48:46
|
>> Yes, I see. I've supposed that preprocessed (converted by gpre from epp >> to cpp) files will remain here. > Unfortunately, it's not staying as before. Seems that some generic > makefile rule makes them to be generated as temporary targets. > > I even put something about secondary rules in the makefile that in a > first moment appeared to fix it. Later I see nothing probably was changed. > Adriano, appears we both do not understand something with .SECONDARY target in gmake. I've tried various uses of it, but... Looks like files stay when doing regular builds, but after first build in clean tree they are deleted. |
|
From: Tagangout, A. <Abd...@me...> - 2010-09-23 16:40:02
|
Hi all, I want to know, how does firebird solve a connect-command in the TCP/IP layer? Which command will be used under Linux (for example) to make a connection to the Server? I know that it is possible to implement a connection to an Firebird Server, but how? Best Regards Abdo MEP Kutsch GmbH & Co. KG Flaßkuhle 1 58452 Witten Tel. : 02302 28262-41 Fax.: 02302 28262-99 mailto:abd...@me... http://www.mep.de -----Ursprüngliche Nachricht----- Von: Alex Peshkoff [mailto:pes...@ma...] Gesendet: Donnerstag, 23. September 2010 13:57 An: For discussion among Firebird Developers Betreff: Re: [Firebird-devel] Porting Firebird On 09/23/10 15:28, Tagangout, Abderrahim wrote: > Hi Alex, > so the build process is complete, and all directories from /temp are deleted, /gen/firebird/bin/* like so ../lib/* are deleted. > and when i try to cross compile i become this: > > > abdo@ubuntu804icnova:~/Firebird-2.1.3.18185-0/gen$ make -f > Makefile.libfbclient > g++ -DSUPERCLIENT -pthread -I../src/include/gen -I../src/include > g++ -I../src/vulcan -DNAMESPACE=Vulcan -I../extern/icu/source/common > g++ -I../extern/icu/source/i18n -DNDEBUG -ggdb > g++ -DFB_SEND_FLAGS=MSG_NOSIGNAL -DLINUX -pipe -MMD -fPIC > g++ -fmessage-length=0 -O3 -march=i586 -mcpu=i686 > g++ -fno-omit-frame-pointer -fno-builtin -DPROD_BUILD -c > g++ ../src/jrd/alt.cpp -o ../temp/superclient/jrd/alt.o > `-mcpu=' is deprecated. Use `-mtune=' or '-march=' instead. > Assembler messages: > Fatal error: can't create ../temp/superclient/jrd/alt.o: No such file > or directory > make: *** [../temp/superclient/jrd/alt.o] Fehler 2 > abdo@ubuntu804icnova:~/Firebird-2.1.3.18185-0/gen$ > > can you see any mistakes? You should not delete subdirs themself - only files from them. ------------------------------------------------------------------------------ Start uncovering the many advantages of virtual appliances and start using them to simplify application deployment and accelerate your shift to cloud computing. http://p.sf.net/sfu/novell-sfdev2dev Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel |
|
From: Alex P. <pes...@ma...> - 2010-09-24 06:35:56
|
On 09/23/10 20:39, Tagangout, Abderrahim wrote: > Hi all, > > I want to know, how does firebird solve a connect-command in the TCP/IP layer? > Which command will be used under Linux (for example) to make a connection to the Server? > I know that it is possible to implement a connection to an Firebird Server, but how? Sorry, the answer is too big to write it here in plain English. But there is c++ file src/remote/interface.cpp, which (together with some more files in remote, but this one is a key one) can give you such answer. We do not have another description of the wire protocol. |
|
From: Dimitry S. <sd...@ib...> - 2010-09-29 16:29:06
|
29.09.2010 18:20, Tagangout, Abderrahim wrote: > Sorry but Where can find Documentation about attach-String or connect-String? In Firebird Quick Start Guide. Chapter "Connection Strings". -- SY, SD. |
|
From: Tagangout, A. <Abd...@me...> - 2010-09-29 16:33:28
|
Hi Dmitry, Many thanks, i will read this Guide first to have an idea about this Structures. Many thanks for your help. Abdo MEP Kutsch GmbH & Co. KG Flaßkuhle 1 58452 Witten Tel. : 02302 28262-41 Fax.: 02302 28262-99 mailto:abd...@me... http://www.mep.de -----Ursprüngliche Nachricht----- Von: Dimitry Sibiryakov [mailto:sd...@ib...] Gesendet: Mittwoch, 29. September 2010 18:28 An: For discussion among Firebird Developers Betreff: Re: [Firebird-devel] Porting Firebird 29.09.2010 18:20, Tagangout, Abderrahim wrote: > Sorry but Where can find Documentation about attach-String or connect-String? In Firebird Quick Start Guide. Chapter "Connection Strings". -- SY, SD. ------------------------------------------------------------------------------ Start uncovering the many advantages of virtual appliances and start using them to simplify application deployment and accelerate your shift to cloud computing. http://p.sf.net/sfu/novell-sfdev2dev Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel |
×
Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.
Thanks for helping keep SourceForge clean.
X