4.1.8 Released
4.1.7 Released
Thanks for finding that issue. I'll try to come up with something that fixes that issue in the next release.I'm pretty sure I didn't test for that scenario when building out the test routines. A couple of things you might find handy: 1) There is a test program that attempts to test all the ndx scenarios - xb_test_ndx. This bug find will be added to that program when resolved. 2) On the locking issues, I had stepped away from the project for a few years and am not sure what happened to the locking...
Post Scriptum : I send you the file removed from its sensitive fields which had several deleted records including only one which had been recreated with the same IDU (main key) (record that is deleted and not deleted is 378PC1710027)
hi gary, I found the time to look in more detail at the anomaly in question by diving back into the code that I debugged more than 12 years ago. I saw in passing that you removed the "RealDelete" option and that's a good thing: it was buggy. From memory I had also deleted the locking system which was also buggy (but perhaps you have cleaned it up since) and which I did not need. I was able to reproduce and found the system flaw: 1) you take a DBF table that you index on a unique index 2) you position...
If you can duplicate the problem with test data, I would be happy to take a look at it. On Thursday, November 30, 2023 at 10:03:57 AM CST, pc <panel_c@users.sourceforge.net> wrote: Hi, Unfortunately, this file contains sensitive data that I do not have the right to distribute. But I can rid it of these and do more in-depth tests. The NDX comes from version 3 of xdb. The versions of dBase are also capricious (DBF is recognized with version 5 but not with version 3, NDX is recognized neither with version...
Hi, Unfortunately, this file contains sensitive data that I do not have the right to distribute. But I can rid it of these and do more in-depth tests. The NDX comes from version 3 of xdb. The versions of dBase are also capricious (DBF is recognized with version 5 but not with version 3, NDX is recognized neither with version 3 nor version 5 - NB: by providing the short names of the files obviously) . The reconstruction of the index under dBase is also imperfect: I thought that the reconstruction...
Are you in a position where you could send me the dbf and ndx file for troubleshooting? On Thursday, November 30, 2023 at 06:59:13 AM CST, pc <panel_c@users.sourceforge.net> wrote: hello garry, And thank you for this response which indicates that the project is in good hands :) I would like to take this opportunity to thank you for your past work: many years ago, I was looking for a way (quickly) to manage dBase files and their indexes for basic and single-user functionality and found your project...
hello garry, And thank you for this response which indicates that the project is in good hands :) I would like to take this opportunity to thank you for your past work: many years ago, I was looking for a way (quickly) to manage dBase files and their indexes for basic and single-user functionality and found your project ( version 3) which I integrated into my project by simplifying it and correcting a few bugs found here and there (without going into the depth of the project). I am attaching what...
Hi From the example output you sent, there are two errors: 1) The ndx file doesn't match the dbf file. In theory, it probably needs to be reindexed. 2) The segfault at the end is probably related to trying to free up the same memory twice, I'll look into it and see if I can figure that one out and get a patch for it in the next version. Can you send me the output from doing a reindex (=8.14)? Also, fyi V 1.4.6 was just released, but it doesn't have any code in it that would impact the issues you...
Good morning, I am a former user of version 3 and was happy to see that the project was not dead. Out of curiosity, I loaded version 4.1.5, compiled it on win64 and linux64(debian) and tested some basic functions with xb_dbfutil. Under Linux or Windows, it is impossible to read an existing NDX index or one created from scratch (reindex), the first and last index also gives anything Under Linux, the output ends with a worrying “Segmentation fault” below the whole session panel@panel-UM350:~$ '/home/panel/Bureau/xbase64-4.1.5/build/debian/bin/xb_dbfutil'...
4.1.6 Released
Thanks for finding this. Your update will be included in the next release. Gary On Friday, June 2, 2023 at 01:09:22 AM CDT, Brecht Sanders <brechtsanders@users.sourceforge.net> wrote: [patches:#8] undefined reference to `xb::operator<<(std::ostream&, xb::xbString const&) Status: open Group: Created: Fri Jun 02, 2023 06:09 AM UTC by Brecht Sanders Last Updated: Fri Jun 02, 2023 06:09 AM UTC Owner: nobody I got the following error when building version 4.1.4 with MinGW-w64: D:/Prog/winlibs64-13.1.0msvcrt/mingw64/bin/../lib/gcc/x86_64-w64-mingw32/13.1.0/../../../../x86_64-w64-mingw32/bin/ld.exe:...
undefined reference to `xb::operator<<(std::ostream&, xb::xbString const&)
Sure, my build script is published here: https://github.com/brechtsanders/winlibs_recipes/blob/main/recipes/xbase64.winlib
Brecht In addition to the Cmakefile.txt updates, can you document the steps for me that you used to build xbase64 with mingw-64? I'm going to add support for it in the library and was hoping you could save me a little time and share how you did it. Thanks Gary On Friday, March 3, 2023 at 04:39:21 AM CST, Brecht Sanders <brechtsanders@users.sourceforge.net> wrote: I figured out how to build 4.1.0 with MinGW-w64. I just used the build/debian/CMakeLists.txt after applying the following changes: patch...
I figured out how to build 4.1.0 with MinGW-w64. I just used the build/debian/CMakeLists.txt after applying the following changes: patch -ulbf build/debian/CMakeLists.txt << EOF @@ -370,6 +370,6 @@ # Microsoft Windows settings -IF( WIN32 ) +IF( MSVC ) # add_definitions( /D_CRT_SECURE_NO_WARNINGS) add_definitions( /EHsc ) -ENDIF( WIN32 ) +ENDIF( MSVC ) @@ -707,3 +707,3 @@ -install (FILES include/xbconfig.h +install (FILES \${PROJECT_BINARY_DIR}/include/xbconfig.h \${PROJECT_SOURCE_DIR}/include/xbase.h...
Correction, I see some CMakeLists.txt files under build/* but which one is for MinGW-w64?
Missing CMake build files in release 4.1.0
Hi Jorg There is an updated version of the xbase64 library on Sourceforge. Version 4.1.0 has a new build setup for the Debian platform. There is a file called README.debian in the build/debian sub folder and it has listed the steps needed to build the library. It should correctly compile the library for what ever platform debian is running on. It supports 32 or 64 bit builds and also supports big or little endian machines. I set up the latest version of Debain on one of my machines and it built out...
Hi Jorg The next release will have a Debian build in it. Thanks Gary
Hi Gary, I am the maintainer of xbase at Debian. Debian will use different architectures[1]. Therefore it is desirable to be able to compile everything with a cmake file. Thanks in advance. CU Jörg [1] https://www.debian.org/ports/index.html
Can you please tell me the exact commands you used to attempt build the library on your machine? Also, is your machine a 32 bit or 64 bit machine? Thanks Gary
Build system
4.0.2 Uploaded
4.0.0 Uploaded
4.0.0 Uploaded
Hi I was one of the original authors many years ago and the library got away from me and was taken over by others. About a year ago, I started to rework the library as time permits and have a mostly newer version of the library that does not yet have any index support. The goal was to get the index logic working, then re-release the library. The version named Xbase64 was was an effort to bring 64 bit file support into the library. Hope this helps On Tuesday, April 17, 2018, 8:28:33 AM CDT, pc <panel_c@users.sourceforge.net>...
this project seems to be abandoned. many questions remains without answer since many years. many bugs reported without corrections. what a pitty ! It had started well ... I tested v.3.1.2 -remove lock system : it doesn't work. -real delete doesn't work fine (subsequent deletes causes hang) -other bugs too -documentation doesn't reflect the reality of the code (return values in particular) ....
Hello, I experienced the same using Ubuntu 16.04.1 LTS. Is there any solution? Regards.
Website still states that 2.0.0 is last stable