5267b71995-08-09Fredrik Hübinette (Hubbe) * Det st}r inte var man skall skicka buggrapporter. fixed * .../modules/math/math.c f_ceil har felaktigt stavade felmeddelanden. fixed * toupper / tolower anv{nder inte LOCALE fixed * inherit funkar inte fixed * fel i felmedelandena i explode fixed * indexering p} mixed fixed * sp-> should be sp[-1]. in expand define fixed * handle include forgot to add ref to a string fixed * end_program freed inherit_names twice if there had been an error fixed * implode allocated too little memory fixed * MEMMEM locked up fixed * include didn't set current_file back fixed * inherit offset was defined wrong, caused driver crashes fixed * ! "" returned 1 fixed * addition of arrays, lists and mappings dumped core hopefully fixed * ulpc -t hung fixed * replace didn't work fixed * subtract on mappings, lists and arrays didn't work hopefully fixed * runtime typechecking in multiply, divide and modulo didn't work fixed * equal bugged on compact arrays hopefully fixed * lots of refcount-bugs in operators.c fixed * ... didn't work if too few arguments were given fixed * %{ %} didn't work in sprint fixed * many of the math functions didn't work fixed * indexing outside of string could crash ulpc fixed * #if efun() always returned true fixed * combine_path doesn't work fixed * multiple replace didn't work fixed * search() didn't work fixed * ctime(time()) give strange results fixed * function returning void doesn't match 'function' ? fixed * constant functionpointers pointed wrong when inherited fixed * range on compact arrays didn't work fixed * implode on empty arrays didn't work fixed * @ on arrays with only one ref didn't work fixed * call_outs didn't work fixed * find call_out didn't work fixed * type_field wrong after check_array_for_destruct fixed * range on arrays dumped core on some machines fixed * "foo"[0..-9] didn't work fixed * string comparisment (< <= > >=) could fail when comparing with "" fixed * array_remove bugged sometimes fixed * f_ltosval2 bugged fixed * compile_file forgot to add a reference to the program fixed * free_short_svalue forgot to free destination fixed * destruct didn't really work (didn't free global variables correctly) fixed * destructing current object (or any object in the history) bugged hopefully fixed * file->open() couldn't create files now it can * set_read_callback & set_write_callback sometimes accessed fd -1 fixed * throw() made fatal() when popping all error contexts fixed * trace mode doesn't work very well when setting stdin/out/err nonblocking. fixed * resize array didn't work fixed * implode didn't work right with an empty firtst element fixed * MEMMEM didn't on a x86 byteorder fixed * count_args didn't know what foreach returned fixed * lost objects were not destructed fixed * stack was corrupt at close_callback fixed * destructing an object twice led to core dumps fixed * spawed process weren't waited upon fixed * bug in merge_array (subtract) fixed * while ( -- i) dumps core? fixed * search didn't work on compact arrays (4 bytes / value) fixed * sorting an empty mapping caused malloc(0) fixed * open trashed memory if returned fd was larger than MAX_FD fixed * time() was wrong in setup code fixed * there was a bug with destructed objects in check_array_for_destruct fixed * casting an int to a float didn't work fixed * check_short_destructed crashed sometimes fixed * check_short_svalue crashed sometimes fixed * is_eq didn't work for programsd fixed * sscanf("-","%*d") returned 1 fixed * sys/select.h was missing in backend.c fixed * assigning list indexes removed when it should add and vice versa fixed * merge_mapping/merge_list had memory leaks. fixed * mkdir() could do strange things with it's second argument fixed * shrink_array didn't free svalues it should sometimes fixed (thanks per) * backtrace() generated illegal arrays if describing destructed objects fixed * setrlimit in main() didn't work fixed * line numbers are bugged after include sometimes fixed * replace bugged on 8-bit characters fixed * now you can use __LINE__ and __FILE__ and #error * sscanf("1a","%d%c",a,b) -> 2 (a=1, b="1a") ? fixed * allocate(10,"object") didn't work fixed * allocate(-10) worked fixed * #elif and #elseif behaved incorrectly fixed * switch didn't exactly work as specified fixed * calling prototypes -> hang (should be error) fixed * sscanf("foo-%-bar"%s-%%-%s",a,b); didn't work fixed
f90e541995-08-17Fredrik Hübinette (Hubbe) * machines with strange byteorders that didn't handle unaligned reads/write jumped into space because EXTRACT_INT didn't work fixed * Didn't work on 64-bit-machines (alpha for instance) fixed * the macro ALIGN interfered with some standard include files on some systems fixed
cb22561995-10-11Fredrik Hübinette (Hubbe) * destruct and this_object() dumped core if called directly from driver fixed * test for working memmove in configure.in was broken fixed * there was some casts used as lvalues in memory.c fixed * modules/files/efun.c didn't compile on some systems that doesn't have MAXPATHLEN defined, in this case we just define it to 32768 fixed * signals only worked once fixed * configure didn't find alloca if it was hideing in libPW hopefully fixed * calling void functions directly from driver callbacks causes coredumps fixed * signal handlers are not called while file->read is blocking fixed
bf4a8d1995-10-20Fredrik Hübinette (Hubbe) Version: 1.0E-15: * typechecking could dump core in lambda functions fixed * I should use a changelog instead of this file fixed Still to do:
cb22561995-10-11Fredrik Hübinette (Hubbe) * really large constant arrays crashes uLPC * need some way of using shared and non-shared module dirs..
bf4a8d1995-10-20Fredrik Hübinette (Hubbe) * Also modules need to be able to add libraries to linker stage
5267b71995-08-09Fredrik Hübinette (Hubbe) * setup_fake_program is called too often, I should keep fake_program updated instead
bf4a8d1995-10-20Fredrik Hübinette (Hubbe) * foreach somtimes says 'bad argument 1 to foreach' after a completed loop (according to Per)
5267b71995-08-09Fredrik Hübinette (Hubbe) 
bf4a8d1995-10-20Fredrik Hübinette (Hubbe) * there are still memory access bugs (it seems) (maybe) (or maybe not)
5267b71995-08-09Fredrik Hübinette (Hubbe)  * #define foo bar, and then foo(1) does not work.. * all variables aren't initialized