1*66361Sbostic# @(#)WHATSNEW 8.2 (Berkeley) 03/16/94 260200Sbostic 3*66361SbosticNew in alpha3.3: The definition of word boundaries has been altered 4*66361Sbosticslightly, to more closely match the usual programming notion that "_" 5*66361Sbosticis an alphabetic. Stuff used for pre-ANSI systems is now in a subdir, 6*66361Sbosticand the makefile no longer alludes to it in mysterious ways. The 7*66361Sbosticmakefile has generally been cleaned up some. Fixes have been made 8*66361Sbostic(again!) so that the regression test will run without -DREDEBUG, at 9*66361Sbosticthe cost of weaker checking. A workaround for a bug in some folks' 10*66361Sbostic<assert.h> has been added. And some more things have been added to 11*66361Sbostictests, including a couple right at the end which are commented out 12*66361Sbosticbecause the code currently flunks them (complex bug; fix coming). 13*66361SbosticPlus the usual minor cleanup. 14*66361Sbostic 15*66361SbosticNew in alpha3.2: Assorted bits of cleanup and portability improvement 16*66361Sbostic(the development base is now a BSDI system using GCC instead of an ancient 17*66361SbosticSun system, and the newer compiler exposed some glitches). Fix for a 18*66361Sbosticserious bug that affected REs using many [] (including REG_ICASE REs 19*66361Sbosticbecause of the way they are implemented), *sometimes*, depending on 20*66361Sbosticmemory-allocation patterns. The header-file prototypes no longer name 21*66361Sbosticthe parameters, avoiding possible name conflicts. The possibility that 22*66361Sbosticsome clot has defined CHAR_MIN as (say) `-128' instead of `(-128)' is 23*66361Sbosticnow handled gracefully. "uchar" is no longer used as an internal type 24*66361Sbosticname (too many people have the same idea). Still the same old lousy 25*66361Sbosticperformance, alas. 26*66361Sbostic 27*66361SbosticNew in alpha3.1: Basically nothing, this release is just a bookkeeping 28*66361Sbosticconvenience. Stay tuned. 29*66361Sbostic 3060200SbosticNew in alpha3.0: Performance is no better, alas, but some fixes have been 3160200Sbosticmade and some functionality has been added. (This is basically the "get 3260200Sbosticit out the door in time for 4.4" release.) One bug fix: regfree() didn't 3360200Sbosticfree the main internal structure (how embarrassing). It is now possible 3460200Sbosticto put NULs in either the RE or the target string, using (resp.) a new 3560200SbosticREG_PEND flag and the old REG_STARTEND flag. The REG_NOSPEC flag to 3660200Sbosticregcomp() makes all characters ordinary, so you can match a literal 3760200Sbosticstring easily (this will become more useful when performance improves!). 3860200SbosticThere are now primitives to match beginnings and ends of words, although 3960200Sbosticthe syntax is disgusting and so is the implementation. The REG_ATOI 4060200Sbosticdebugging interface has changed a bit. And there has been considerable 4160200Sbosticinternal cleanup of various kinds. 4260200Sbostic 4360200SbosticNew in alpha2.3: Split change list out of README, and moved flags notes 4460200Sbosticinto Makefile. Macro-ized the name of regex(7) in regex(3), since it has 4560200Sbosticto change for 4.4BSD. Cleanup work in engine.c, and some new regression 4660200Sbostictests to catch tricky cases thereof. 4760200Sbostic 4860200SbosticNew in alpha2.2: Out-of-date manpages updated. Regerror() acquires two 4960200Sbosticsmall extensions -- REG_ITOA and REG_ATOI -- which avoid debugging kludges 5060200Sbosticin my own test program and might be useful to others for similar purposes. 5160200SbosticThe regression test will now compile (and run) without REDEBUG. The 5260200SbosticBRE \$ bug is fixed. Most uses of "uchar" are gone; it's all chars now. 5360200SbosticChar/uchar parameters are now written int/unsigned, to avoid possible 5460200Sbosticportability problems with unpromoted parameters. Some unsigned casts have 5560200Sbosticbeen introduced to minimize portability problems with shifting into sign 5660200Sbosticbits. 5760200Sbostic 5860200SbosticNew in alpha2.1: Lots of little stuff, cleanup and fixes. The one big 5960200Sbosticthing is that regex.h is now generated, using mkh, rather than being 6060200Sbosticsupplied in the distribution; due to circularities in dependencies, 6160200Sbosticyou have to build regex.h explicitly by "make h". The two known bugs 6260200Sbostichave been fixed (and the regression test now checks for them), as has a 6360200Sbosticproblem with assertions not being suppressed in the absence of REDEBUG. 6460200SbosticNo performance work yet. 6560200Sbostic 6660200SbosticNew in alpha2: Backslash-anything is an ordinary character, not an 6760200Sbosticerror (except, of course, for the handful of backslashed metacharacters 6860200Sbosticin BREs), which should reduce script breakage. The regression test 6960200Sbosticchecks *where* null strings are supposed to match, and has generally 7060200Sbosticbeen tightened up somewhat. Small bug fixes in parameter passing (not 7160200Sbosticharmful, but technically errors) and some other areas. Debugging 7260200Sbosticinvoked by defining REDEBUG rather than not defining NDEBUG. 7360200Sbostic 7460200SbosticNew in alpha+3: full prototyping for internal routines, using a little 7560200Sbostichelper program, mkh, which extracts prototypes given in stylized comments. 7660200SbosticMore minor cleanup. Buglet fix: it's CHAR_BIT, not CHAR_BITS. Simple 7760200Sbosticpre-screening of input when a literal string is known to be part of the 7860200SbosticRE; this does wonders for performance. 7960200Sbostic 8060200SbosticNew in alpha+2: minor bits of cleanup. Notably, the number "32" for the 8160200Sbosticword width isn't hardwired into regexec.c any more, the public header 8260200Sbosticfile prototypes the functions if __STDC__ is defined, and some small typos 8360200Sbosticin the manpages have been fixed. 8460200Sbostic 8560200SbosticNew in alpha+1: improvements to the manual pages, and an important 8660200Sbosticextension, the REG_STARTEND option to regexec(). 87