[Firebird-devel] Re. Fix Firebird build against OS X SDK 10.11
new small patch from libo https://github.com/LibreOffice/core/commit/a799d7c9a66464f33bd8ee42d535e29a976d64ae Paul please apply if consider as necessarily There will have been a reason why that particular flag was used The change for nodefaultlibs was added to fix a problem with gcc 4.1 by the look of it. Since we are no longer using gcc 4.1, since MacOS 10.8 (currently used to build 2.x) is now running gcc 4.2, its probably OK to remove it. See original checkin below: Modified Files: Tag: B2_5_Release darwin.defaults Log Message: Fixed darwin built with gcc4.1 Index: darwin.defaults === RCS file: /cvsroot/firebird/firebird2/builds/posix/darwin.defaults,v retrieving revision 1.4 retrieving revision 1.4.2.1 diff -b -U3 -r1.4 -r1.4.2.1 --- darwin.defaults 11 Sep 2009 10:49:30 - 1.4 +++ darwin.defaults 12 Oct 2009 13:41:00 - 1.4.2.1 @@ -55,8 +55,15 @@ LINK_TRACE = $(LIB_LINK) $(LIB_BUNDLE_OPTIONS) -LINK_CLIENT = $(CLIENTLIB_LINK) $(LINK_FIREBIRD_CLIENT_SYMBOLS) $(LIB_LINK_OPTIONS) $(LIB_LINK_IMPLIB)\ +LINK_CLIENT = $(LIB_LINK) -nodefaultlibs $(LINK_FIREBIRD_CLIENT_SYMBOLS) $(LIB_LINK_OPTIONS) $(LIB_LINK_IMPLIB)\ $(LIB_CLIENT_LINK_OPTIONS) $(LIB_LINK_SONAME) +ifeq (@VOID_PTR_SIZE@,8) +GCCS=-lgcc_s.10.5 +else +GCCS=-lgcc_s.10.4 +endif +STATIC_CXX_SUPPORT = -lsupc++ $(GCCS) -lgcc_eh -lSystem + LINK_EMBED = $(LIB_LINK) $(LINK_FIREBIRD_EMBED_SYMBOLS) $(LIB_LINK_OPTIONS)\ $(LIB_EMBED_LINK_OPTIONS) $(LIB_LINK_SONAME) Paul -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel
Re: [Firebird-devel] Regular Expression Library
11.06.2015 13:44, Adriano dos Santos Fernandes wrote: On 11/06/2015 06:08, Dmitry Yemanov wrote: 10.06.2015 21:03, Adriano dos Santos Fernandes wrote: At a first look, boost.regex is the library who can do it. It's templatized and works with iterators and traits for customization. It has integration with ICU and MFC strings, so this is useful to understand how to customize it. Is it compatible with our implementation syntax wise? No. It will require transformation. My idea is not to completely remove our parser, which may feed tokens (or something like) at compilation phase to the external library. My idea is to remove our execution engine. Is it possible to do with ICU regexp facilities ? Regards, Vlad -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel
[Firebird-devel] [FB-Tracker] Created: (CORE-4838) DDL statement CREATE TABLE
DDL statement CREATE TABLE -- Key: CORE-4838 URL: http://tracker.firebirdsql.org/browse/CORE-4838 Project: Firebird Core Issue Type: Bug Components: API / Client Library Affects Versions: 2.5.4 Environment: FlameRobin 0.9.3 Unicode and 32-bit Embedded on Windows Server 2003 R2 Reporter: Andrew Klimov Preparing statement: CREATE TABLE file_usage ( fid INTEGER NOT NULL, module VARCHAR(255) DEFAULT '' NOT NULL, type VARCHAR(64) DEFAULT '' NOT NULL, id INTEGER DEFAULT 0 NOT NULL, count INTEGER DEFAULT 0 NOT NULL ) Error: *** IBPP::SQLException *** Context: Statement::Prepare( CREATE TABLE file_usage ( fid INTEGER NOT NULL, module VARCHAR(255) DEFAULT '' NOT NULL, type VARCHAR(64) DEFAULT '' NOT NULL, id INTEGER DEFAULT 0 NOT NULL, count INTEGER DEFAULT 0 NOT NULL ) ) Message: isc_dsql_prepare failed SQL Message : -104 Invalid token Engine Code: 335544569 Engine Message : Dynamic SQL Error SQL error code = -104 Token unknown - line 1, column 159 count -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://tracker.firebirdsql.org/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel
Re: [Firebird-devel] Regular Expression Library
On 11/06/2015 05:26, Mark Rotteveel wrote: That doesn't work: SQL regex is an SQL abomination that doesn't look like normal (perl or posix) regular expressions. Surely it has not normal perl/posix syntax. BTW, as far as I know the postgresql similar to is SQL standard compliant, but they also provide a regex that is more familiar to people. That was the abomination. PostgreSQL treat SIMILAR like a standard regex with the percent/underscore additions, at least the time I look. Adriano -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel
Re: [Firebird-devel] firebird 2.5.4 on osx 64 configure host
On 06/11/2015 10:51 AM, marius adrian popa wrote: I will do a test with B2_5_Release , with 2.5.4 seems to detect the darwin_i386 even if the changes with x*64-*-darwin* | x*64_*-darwin*) are in the configure script In case of problems please report here actual value of target that caused them. -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel
Re: [Firebird-devel] Regular Expression Library
10.06.2015 21:03, Adriano dos Santos Fernandes wrote: At a first look, boost.regex is the library who can do it. It's templatized and works with iterators and traits for customization. It has integration with ICU and MFC strings, so this is useful to understand how to customize it. Is it compatible with our implementation syntax wise? Understand and customize it will require time, however. And I do not want to lost time if it's going to be objected because: - It's external code - It's boost - It's a library which uses C++ as C++, i.e., uses the std library - Not only it's internal (like ICU) uses C++, but its interface too These are not showstopers, IMO. So what do you think on include such type of library in Firebird? I don't mind provided that the replacement happens post v3. I'm going to branch off the v3 code in August, so trunk will be open for new development at that point. Dmitry -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel
Re: [Firebird-devel] Regular Expression Library
On 11/06/2015 07:58, Vlad Khorsun wrote: 11.06.2015 13:44, Adriano dos Santos Fernandes wrote: On 11/06/2015 06:08, Dmitry Yemanov wrote: 10.06.2015 21:03, Adriano dos Santos Fernandes wrote: At a first look, boost.regex is the library who can do it. It's templatized and works with iterators and traits for customization. It has integration with ICU and MFC strings, so this is useful to understand how to customize it. Is it compatible with our implementation syntax wise? No. It will require transformation. My idea is not to completely remove our parser, which may feed tokens (or something like) at compilation phase to the external library. My idea is to remove our execution engine. Is it possible to do with ICU regexp facilities ? I don't think so. ICU is very Unicode focused, so we loose all of our collations attributions. Our engine should work even with binary data. Adriano -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel
[Firebird-devel] Any chance for realisation of CORE-2040?
Hello, any chance for a realisation in FB 3.0? Or is there a way to access the variables of an exception message in PSQL? Björn -- Björn Reimer - RRZE -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel
Re: [Firebird-devel] Regular Expression Library
My vote is to use the standard one or the boost.regex On Thu, Jun 11, 2015 at 10:07 AM, Egor Pugin egor.pu...@gmail.com wrote: By the way C++11 has native regex support. On 10 June 2015 at 21:03, Adriano dos Santos Fernandes adrian...@gmail.com wrote: Hi! When I created SIMILAR TO predicate, I research for libraries to do the real work and found no good alternative. PgSQL were using (AFAIR) Henry Spencer library, but its (PgSQL) SIMILAR TO were definitively non SQL standard. We needed a library to work with 1, 2 or 4 byte-length characters, actually their canonical representations of collations. So as no suitable library was found, I wrote myself one based on a CodeProject library. This library was simple and recursive, hence it had problems with BLOBs. Then I managed to recreate it with non-recursive code. This library is trick. First of all, I'm not experienced in regex matching. Bugs on it are difficult to fix, and there are cases which it's really slow (I didn't compared with others libraries, which may also be slow with bomb expressions). I'd want to replace it with something capable of working as we need, when customized (not editing their code). At a first look, boost.regex is the library who can do it. It's templatized and works with iterators and traits for customization. It has integration with ICU and MFC strings, so this is useful to understand how to customize it. Understand and customize it will require time, however. And I do not want to lost time if it's going to be objected because: - It's external code - It's boost - It's a library which uses C++ as C++, i.e., uses the std library - Not only it's internal (like ICU) uses C++, but its interface too So what do you think on include such type of library in Firebird? Adriano -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel -- Egor Pugin -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel
Re: [Firebird-devel] firebird 2.5.4 on osx 64 configure host
I will do a test with B2_5_Release , with 2.5.4 seems to detect the darwin_i386 even if the changes with x*64-*-darwin* | x*64_*-darwin*) are in the configure script I will double check and report On Wed, Jun 10, 2015 at 5:51 PM, Paul Beach pbe...@ibphoenix.com wrote: One issue is that bash firebird configure script detects host/target as i386-darwin on x86-64 osx so I had to use the 2.5.2 trick from libreoffice and force the build for darwin_x86_64 i*86-*-darwin*) MAKEFILE_PREFIX=darwin_x86_64 issue was libreoffice detects host to be darwin x86_64 and firebird configure detects it as i386 darwin so we had this error while linking ld: warning: ignoring file ../temp/boot/jrd/dsc.o, file was built for unsupported file format ( 0xCF 0xFA 0xED 0xFE 0x07 0x00 0x00 0x01 0x03 0x00 0x00 0x00 0x01 0x00 0x00 0x00 ) which is not the architecture being linked (i386): ../temp/boot/jrd/dsc.o http://ci.libreoffice.org/job/lo_gerrit_master/2971/Gerrit=Gerrit,Platform=MacOSX/consoleText This was fixed in B2_5_Release sometime ago. Configure detects both 64bit and 32bit using x*64-*-darwin* | x*64_*-darwin*) MAKEFILE_PRIFIX=darwin_x86_64 and i*86-*-darwin* | i*86_*-darwin*) MAKEFILE_PRIFIX=darwin_i386 Regards Paul -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel
Re: [Firebird-devel] Problem with MERGE
-Original Message- From: Dmitry Yemanov [mailto:firebi...@yandex.ru] Sent: Miércoles, 10 de Junio de 2015 15:40 With the old logic, the result is one. With the new logic, the result is two becaise we get the direct and indirect effects. I doubt the new logic would return two. These counters are per-request and trigger is a separate request, so its changes shouldn't be counted for the parent request. DSQL's get_request_info() retrieves not only the changes but the selected records, too so we still need to filter out those. Sure, the idea was to sum up only I/U/D counters. Has anybody thought what will we reported in the extreme event of more than 2^32 updated or deleted records? The counters are going to overflow because they are ULONG in the engine. What a cruel world ;-) Somebody will have to fix the cruel world the day it became common to update 2^33 records. :-) For the moment, I changed the code and I'm going to commit in the next days. BTW, it's exactly CORE-4817. C. -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel
Re: [Firebird-devel] Regular Expression Library
By the way C++11 has native regex support. On 10 June 2015 at 21:03, Adriano dos Santos Fernandes adrian...@gmail.com wrote: Hi! When I created SIMILAR TO predicate, I research for libraries to do the real work and found no good alternative. PgSQL were using (AFAIR) Henry Spencer library, but its (PgSQL) SIMILAR TO were definitively non SQL standard. We needed a library to work with 1, 2 or 4 byte-length characters, actually their canonical representations of collations. So as no suitable library was found, I wrote myself one based on a CodeProject library. This library was simple and recursive, hence it had problems with BLOBs. Then I managed to recreate it with non-recursive code. This library is trick. First of all, I'm not experienced in regex matching. Bugs on it are difficult to fix, and there are cases which it's really slow (I didn't compared with others libraries, which may also be slow with bomb expressions). I'd want to replace it with something capable of working as we need, when customized (not editing their code). At a first look, boost.regex is the library who can do it. It's templatized and works with iterators and traits for customization. It has integration with ICU and MFC strings, so this is useful to understand how to customize it. Understand and customize it will require time, however. And I do not want to lost time if it's going to be objected because: - It's external code - It's boost - It's a library which uses C++ as C++, i.e., uses the std library - Not only it's internal (like ICU) uses C++, but its interface too So what do you think on include such type of library in Firebird? Adriano -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel -- Egor Pugin -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel
[Firebird-devel] Fix Firebird build against OS X SDK 10.11
new small patch from libo https://github.com/LibreOffice/core/commit/a799d7c9a66464f33bd8ee42d535e29a976d64ae Paul please apply if consider as necessarily -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel
Re: [Firebird-devel] Regular Expression Library
That doesn't work: SQL regex is an SQL abomination that doesn't look like normal (perl or posix) regular expressions. BTW, as far as I know the postgresql similar to is SQL standard compliant, but they also provide a regex that is more familiar to people. - Reply message - Van: Egor Pugin egor.pu...@gmail.com Aan: For discussion among Firebird Developers firebird-devel@lists.sourceforge.net Onderwerp: [Firebird-devel] Regular Expression Library Datum: do, jun. 11, 2015 09:07 By the way C++11 has native regex support. On 10 June 2015 at 21:03, Adriano dos Santos Fernandes adrian...@gmail.com wrote: Hi! When I created SIMILAR TO predicate, I research for libraries to do the real work and found no good alternative. PgSQL were using (AFAIR) Henry Spencer library, but its (PgSQL) SIMILAR TO were definitively non SQL standard. We needed a library to work with 1, 2 or 4 byte-length characters, actually their canonical representations of collations. So as no suitable library was found, I wrote myself one based on a CodeProject library. This library was simple and recursive, hence it had problems with BLOBs. Then I managed to recreate it with non-recursive code. This library is trick. First of all, I'm not experienced in regex matching. Bugs on it are difficult to fix, and there are cases which it's really slow (I didn't compared with others libraries, which may also be slow with bomb expressions). I'd want to replace it with something capable of working as we need, when customized (not editing their code). At a first look, boost.regex is the library who can do it. It's templatized and works with iterators and traits for customization. It has integration with ICU and MFC strings, so this is useful to understand how to customize it. Understand and customize it will require time, however. And I do not want to lost time if it's going to be objected because: - It's external code - It's boost - It's a library which uses C++ as C++, i.e., uses the std library - Not only it's internal (like ICU) uses C++, but its interface too So what do you think on include such type of library in Firebird? Adriano -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel -- Egor Pugin -- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel-- Firebird-Devel mailing list, web interface at https://lists.sourceforge.net/lists/listinfo/firebird-devel