Tags: AmigaLabs/clib4
Tags
Merge development into master (#253) * Removed double /0 on buffer in vsprintf. This should close issue #208 * Added clib4.library.debug on make release * Merge master to development (#227) * Release v1.4.1 changes (#225) * Code improvements based on compiler warnings and cppcheck reports * Removed some commented code * Reverted the clib4.h changes and changed the initial value in common_pathconf.c * Initial commit of splitting stubs.c into separate units and modifying the libc.gmk file to consider them * All stubs file are prepended with "stubs_". libc.gmk should only be preparing a list of stubs_ files. * Further * Fix problem with lingering pipes (spawnvpe). Fix problem with undertermined size of pthread_barrier_t (common.h/pthread.h). * Remove lint in spawnvpe.c. * Further link removal and clarification before PR finalization (spawnvpe). * Free command string. Remove unnecessary strdup. Tidy up comments. (spawvnvpe) * Fix spawnv. * Fixes for test programs. asin.cpp most probably needs a fix outside clib4 * Remove LOG.txt * Added a script to check offset values for stubs * Update check_offset.sh * Small functions definitions. Added some TCP flags. Merged latest wmem changes * Added u_* types variants in sys/types.h file needed by some configure scripts * Fix spawnv. Fix problem with fstat and freeing of examinedata struct (causing NIL: streams to fail). Fix popen so it doesn't write to stderr (posix). Adding better popen test. * Add extra stack to child in spawnvpe. * puts can cause a deadlock because __funlockfile_r was called also inside __putc * Fix memory allocator. * Further fixes for memory allocator. Emulated realloc in all four memory allocators (SIMPLE, BLOCK, BLOCK_FAST and STRICT). Since we don't have a global realloc on amigaos, there is no generic way to do this, so the solution has been to reimplement all of these with size tables, reallocation using AllocVec and finaly memcpy(). * Suggested patch to include argv[0] in unix path semantics (qmake). * Change debug output to use clib4 standard notation in wireshark allocator. * Fix for broken I/O on clib4 caused by old clib2 code was causing pthreads to print on bad streams * Wrong clib4 mismatch version in crtbegin.c * Added pthread_join_io.c to test correct STDOUT behaviour * Better remove minor version from requirements since this can cause misunderstandings * Changed 0x80000000 to 0x10000000 to handle signed int status variable correctly * Replaced AllocVecTags/FreeVec with malloc/free where possible * pthread_create now use correct file streams (and added some tests before creating the thread) * Added some debug on setitimer * Replaced AllocVecTags with with malloc in environment creation * Fixed small error * Added MEMORY_DEBUG define to enable memory debug in memory allocator * -Update debug output in several modules. -Add FDF_PIPE flag to result pipe for proper closing in the fdhook function (poopen). -Move free(command) to the bottom of the function (spawnvpe). -Fix the order of calls in waitpid, so it doesn't miss the return code of the child process, when doing an explicit WaitForChildExit. -Changing the exit flag byte definition for WIFEXITED(). * Fixed waitpid completely and utterly. It really got a good banging. I am quite satisfied. * Fix broken merge * Make utimensat not fail on qmake install. * If getservbyname, getservbyport or gethostbyname has no results try to fall back to Roadshow functions. This will fix a problem if DHCP is used and no configuration files are on DEVS:Internet folder. * Small changes to gethostbyname example * Fixed a BAD problem on AllocVecTags when allocating memory for IOB * Added test for hints->ai_socktype == 0 in getaddrinfo.c (qt6). * OLD Constructor call was missing PTHREAD_CONSTRUCTOR definition. Added some debug messages on growfdtable and growiobtable * makeEnvironment now use a memory pool to allocate items so memory is correctly freed when memory pool is deleted * Added some debug on some memory functions. If MEMORY_DEBUG is set to 1 (with DEBUG ON) at the end of program we can see if there are unfreed memory allocations * Fixed IO memory leak caused by unfreed buffer memory in files. File buffer memory now is allocated via a memory pool and freed on exit * Some file buffers were allocated via AllocVec instead of AllocVecPooled. * README changes * Fix stack size code in pthread_create.c * Use sigmask to access signal bit, instead of << * Improved/Commented test programs to better udnerstand expected behavior * Added Task name in debug message * Added sigwaitinfo & sigtimedwait * Rearrange properties in clib4 to keep backwards compatibility * Converted all line endings to LF (Linux) Signed-off-by: Josef Wegner <josef.wegner@outlook.com> * Changed the mergeMaster action to include discord notification * added a dummy file * disabled builds and updated the announcement message * Some change in dummy * did a fix in the compare url * disabled builds and updated the announcement message * did a fix in the compare url * 2nd dummy change * Added the compare url as a second announcement * Added the compare url as a second announcement * 3rd change * Added own compare url * 4th change * change the comparison url * 5th change * merge messages into one notification * 6th change * The slice didn't work * 7th change * Discord announcement on PR merge * Removed uneeded call to flush * More line endings changes (#247) * Latest stdio destructor changes was creating a bad problem subtracting index in a loop.. Some additional tests on flush * Duplicate IO handles on CreateNewProc * More WINDOWS to LINUX file ending conversion * Wrong Printf version requirement printing (#248) * Modify the initial test of parameter validity in flock to embrace the actual span according to documentation and practice. (#249) * Fixed iconv problem on dcngettext. Added Amiga-1251 charset on iconv (#250) * Latest stdio destructor changes was creating a bad problem subtracting index in a loop.. Some additional tests on flush * Duplicate IO handles on CreateNewProc * More WINDOWS to LINUX file ending conversion * Wrong Printf version requirement printing * Get DOSIFace in pthread_create from extern _IDOS * Added %b format type in printf * Fix on dcngettext was crashing on bindtextdomain and on destructor * Added Amiga-1251 charset to iconv * Added dlopen example to test shared objects along rpath and soname (#251) * Latest stdio destructor changes was creating a bad problem subtracting index in a loop.. Some additional tests on flush * Duplicate IO handles on CreateNewProc * More WINDOWS to LINUX file ending conversion * Wrong Printf version requirement printing * Get DOSIFace in pthread_create from extern _IDOS * Added %b format type in printf * Fix on dcngettext was crashing on bindtextdomain and on destructor * Added Amiga-1251 charset to iconv * Removed __set_errno from librt since it isn't exported and undefined if you use the library * Added dlopen example to test shared objects alog rpath and soname * Merge Afxgroup branch into development * Latest stdio destructor changes was creating a bad problem subtracting index in a loop.. Some additional tests on flush * Duplicate IO handles on CreateNewProc * More WINDOWS to LINUX file ending conversion * Wrong Printf version requirement printing * Get DOSIFace in pthread_create from extern _IDOS * Added %b format type in printf * Fix on dcngettext was crashing on bindtextdomain and on destructor * Added Amiga-1251 charset to iconv * Removed __set_errno from librt since it isn't exported and undefined if you use the library * Added dlopen example to test shared objects alog rpath and soname * Removed some quirks from dlopen examples --------- Signed-off-by: Josef Wegner <josef.wegner@outlook.com> Co-authored-by: George Sokianos <walkero@gmail.com> Co-authored-by: rjd <3246251196ryan@gmail.com> Co-authored-by: Alfkil Thorbjørn Wennermark <alfkil@gmail.com> Co-authored-by: MigthyMax <migthymax1977@gmail.com> Co-authored-by: Josef Wegner <josef.wegner@outlook.com>
PreviousNext