1New in alpha3.4: The complex bug alluded to below has been fixed (in a 2slightly kludgey temporary way that may hurt efficiency a bit; this is 3another "get it out the door for 4.4" release). The tests at the end of 4the tests file have accordingly been uncommented. The primary sign of 5the bug was that something like a?b matching ab matched b rather than ab. 6(The bug was essentially specific to this exact situation, else it would 7have shown up earlier.) 8 9New in alpha3.3: The definition of word boundaries has been altered 10slightly, to more closely match the usual programming notion that "_" 11is an alphabetic. Stuff used for pre-ANSI systems is now in a subdir, 12and the makefile no longer alludes to it in mysterious ways. The 13makefile has generally been cleaned up some. Fixes have been made 14(again!) so that the regression test will run without -DREDEBUG, at 15the cost of weaker checking. A workaround for a bug in some folks' 16<assert.h> has been added. And some more things have been added to 17tests, including a couple right at the end which are commented out 18because the code currently flunks them (complex bug; fix coming). 19Plus the usual minor cleanup. 20 21New in alpha3.2: Assorted bits of cleanup and portability improvement 22(the development base is now a BSDI system using GCC instead of an ancient 23Sun system, and the newer compiler exposed some glitches). Fix for a 24serious bug that affected REs using many [] (including REG_ICASE REs 25because of the way they are implemented), *sometimes*, depending on 26memory-allocation patterns. The header-file prototypes no longer name 27the parameters, avoiding possible name conflicts. The possibility that 28some clot has defined CHAR_MIN as (say) `-128' instead of `(-128)' is 29now handled gracefully. "uchar" is no longer used as an internal type 30name (too many people have the same idea). Still the same old lousy 31performance, alas. 32 33New in alpha3.1: Basically nothing, this release is just a bookkeeping 34convenience. Stay tuned. 35 36New in alpha3.0: Performance is no better, alas, but some fixes have been 37made and some functionality has been added. (This is basically the "get 38it out the door in time for 4.4" release.) One bug fix: regfree() didn't 39free the main internal structure (how embarrassing). It is now possible 40to put NULs in either the RE or the target string, using (resp.) a new 41REG_PEND flag and the old REG_STARTEND flag. The REG_NOSPEC flag to 42regcomp() makes all characters ordinary, so you can match a literal 43string easily (this will become more useful when performance improves!). 44There are now primitives to match beginnings and ends of words, although 45the syntax is disgusting and so is the implementation. The REG_ATOI 46debugging interface has changed a bit. And there has been considerable 47internal cleanup of various kinds. 48 49New in alpha2.3: Split change list out of README, and moved flags notes 50into Makefile. Macro-ized the name of regex(7) in regex(3), since it has 51to change for 4.4BSD. Cleanup work in engine.c, and some new regression 52tests to catch tricky cases thereof. 53 54New in alpha2.2: Out-of-date manpages updated. Regerror() acquires two 55small extensions -- REG_ITOA and REG_ATOI -- which avoid debugging kludges 56in my own test program and might be useful to others for similar purposes. 57The regression test will now compile (and run) without REDEBUG. The 58BRE \$ bug is fixed. Most uses of "uchar" are gone; it's all chars now. 59Char/uchar parameters are now written int/unsigned, to avoid possible 60portability problems with unpromoted parameters. Some unsigned casts have 61been introduced to minimize portability problems with shifting into sign 62bits. 63 64New in alpha2.1: Lots of little stuff, cleanup and fixes. The one big 65thing is that regex.h is now generated, using mkh, rather than being 66supplied in the distribution; due to circularities in dependencies, 67you have to build regex.h explicitly by "make h". The two known bugs 68have been fixed (and the regression test now checks for them), as has a 69problem with assertions not being suppressed in the absence of REDEBUG. 70No performance work yet. 71 72New in alpha2: Backslash-anything is an ordinary character, not an 73error (except, of course, for the handful of backslashed metacharacters 74in BREs), which should reduce script breakage. The regression test 75checks *where* null strings are supposed to match, and has generally 76been tightened up somewhat. Small bug fixes in parameter passing (not 77harmful, but technically errors) and some other areas. Debugging 78invoked by defining REDEBUG rather than not defining NDEBUG. 79 80New in alpha+3: full prototyping for internal routines, using a little 81helper program, mkh, which extracts prototypes given in stylized comments. 82More minor cleanup. Buglet fix: it's CHAR_BIT, not CHAR_BITS. Simple 83pre-screening of input when a literal string is known to be part of the 84RE; this does wonders for performance. 85 86New in alpha+2: minor bits of cleanup. Notably, the number "32" for the 87word width isn't hardwired into regexec.c any more, the public header 88file prototypes the functions if __STDC__ is defined, and some small typos 89in the manpages have been fixed. 90 91New in alpha+1: improvements to the manual pages, and an important 92extension, the REG_STARTEND option to regexec(). 93
