xref: /netbsd-src/lib/libc/regex/WHATSNEW (revision 3ed83140b4bcc633de82f87e84ec44852660f9dd)
1*3ed83140SjtcNew in alpha3.3:  The definition of word boundaries has been altered
2*3ed83140Sjtcslightly, to more closely match the usual programming notion that "_"
3*3ed83140Sjtcis an alphabetic.  Stuff used for pre-ANSI systems is now in a subdir,
4*3ed83140Sjtcand the makefile no longer alludes to it in mysterious ways.  The
5*3ed83140Sjtcmakefile has generally been cleaned up some.  Fixes have been made
6*3ed83140Sjtc(again!) so that the regression test will run without -DREDEBUG, at
7*3ed83140Sjtcthe cost of weaker checking.  A workaround for a bug in some folks'
8*3ed83140Sjtc<assert.h> has been added.  And some more things have been added to
9*3ed83140Sjtctests, including a couple right at the end which are commented out
10*3ed83140Sjtcbecause the code currently flunks them (complex bug; fix coming).
11*3ed83140SjtcPlus the usual minor cleanup.
12*3ed83140Sjtc
136931099eSjtcNew in alpha3.2:  Assorted bits of cleanup and portability improvement
146931099eSjtc(the development base is now a BSDI system using GCC instead of an ancient
156931099eSjtcSun system, and the newer compiler exposed some glitches).  Fix for a
166931099eSjtcserious bug that affected REs using many [] (including REG_ICASE REs
176931099eSjtcbecause of the way they are implemented), *sometimes*, depending on
186931099eSjtcmemory-allocation patterns.  The header-file prototypes no longer name
196931099eSjtcthe parameters, avoiding possible name conflicts.  The possibility that
206931099eSjtcsome clot has defined CHAR_MIN as (say) `-128' instead of `(-128)' is
216931099eSjtcnow handled gracefully.  "uchar" is no longer used as an internal type
226931099eSjtcname (too many people have the same idea).  Still the same old lousy
236931099eSjtcperformance, alas.
246931099eSjtc
25b90ff831SjtcNew in alpha3.1:  Basically nothing, this release is just a bookkeeping
26b90ff831Sjtcconvenience.  Stay tuned.
27b90ff831Sjtc
28b90ff831SjtcNew in alpha3.0:  Performance is no better, alas, but some fixes have been
29b90ff831Sjtcmade and some functionality has been added.  (This is basically the "get
30b90ff831Sjtcit out the door in time for 4.4" release.)  One bug fix:  regfree() didn't
31b90ff831Sjtcfree the main internal structure (how embarrassing).  It is now possible
32b90ff831Sjtcto put NULs in either the RE or the target string, using (resp.) a new
33b90ff831SjtcREG_PEND flag and the old REG_STARTEND flag.  The REG_NOSPEC flag to
34b90ff831Sjtcregcomp() makes all characters ordinary, so you can match a literal
35b90ff831Sjtcstring easily (this will become more useful when performance improves!).
36b90ff831SjtcThere are now primitives to match beginnings and ends of words, although
37b90ff831Sjtcthe syntax is disgusting and so is the implementation.  The REG_ATOI
38b90ff831Sjtcdebugging interface has changed a bit.  And there has been considerable
39b90ff831Sjtcinternal cleanup of various kinds.
40b90ff831Sjtc
41b90ff831SjtcNew in alpha2.3:  Split change list out of README, and moved flags notes
42b90ff831Sjtcinto Makefile.  Macro-ized the name of regex(7) in regex(3), since it has
43b90ff831Sjtcto change for 4.4BSD.  Cleanup work in engine.c, and some new regression
44b90ff831Sjtctests to catch tricky cases thereof.
45b90ff831Sjtc
46b90ff831SjtcNew in alpha2.2:  Out-of-date manpages updated.  Regerror() acquires two
47b90ff831Sjtcsmall extensions -- REG_ITOA and REG_ATOI -- which avoid debugging kludges
48b90ff831Sjtcin my own test program and might be useful to others for similar purposes.
49b90ff831SjtcThe regression test will now compile (and run) without REDEBUG.  The
50b90ff831SjtcBRE \$ bug is fixed.  Most uses of "uchar" are gone; it's all chars now.
51b90ff831SjtcChar/uchar parameters are now written int/unsigned, to avoid possible
52b90ff831Sjtcportability problems with unpromoted parameters.  Some unsigned casts have
53b90ff831Sjtcbeen introduced to minimize portability problems with shifting into sign
54b90ff831Sjtcbits.
55b90ff831Sjtc
56b90ff831SjtcNew in alpha2.1:  Lots of little stuff, cleanup and fixes.  The one big
57b90ff831Sjtcthing is that regex.h is now generated, using mkh, rather than being
58b90ff831Sjtcsupplied in the distribution; due to circularities in dependencies,
59b90ff831Sjtcyou have to build regex.h explicitly by "make h".  The two known bugs
60b90ff831Sjtchave been fixed (and the regression test now checks for them), as has a
61b90ff831Sjtcproblem with assertions not being suppressed in the absence of REDEBUG.
62b90ff831SjtcNo performance work yet.
63b90ff831Sjtc
64b90ff831SjtcNew in alpha2:  Backslash-anything is an ordinary character, not an
65b90ff831Sjtcerror (except, of course, for the handful of backslashed metacharacters
66b90ff831Sjtcin BREs), which should reduce script breakage.  The regression test
67b90ff831Sjtcchecks *where* null strings are supposed to match, and has generally
68b90ff831Sjtcbeen tightened up somewhat.  Small bug fixes in parameter passing (not
69b90ff831Sjtcharmful, but technically errors) and some other areas.  Debugging
70b90ff831Sjtcinvoked by defining REDEBUG rather than not defining NDEBUG.
71b90ff831Sjtc
72b90ff831SjtcNew in alpha+3:  full prototyping for internal routines, using a little
73b90ff831Sjtchelper program, mkh, which extracts prototypes given in stylized comments.
74b90ff831SjtcMore minor cleanup.  Buglet fix:  it's CHAR_BIT, not CHAR_BITS.  Simple
75b90ff831Sjtcpre-screening of input when a literal string is known to be part of the
76b90ff831SjtcRE; this does wonders for performance.
77b90ff831Sjtc
78b90ff831SjtcNew in alpha+2:  minor bits of cleanup.  Notably, the number "32" for the
79b90ff831Sjtcword width isn't hardwired into regexec.c any more, the public header
80b90ff831Sjtcfile prototypes the functions if __STDC__ is defined, and some small typos
81b90ff831Sjtcin the manpages have been fixed.
82b90ff831Sjtc
83b90ff831SjtcNew in alpha+1:  improvements to the manual pages, and an important
84b90ff831Sjtcextension, the REG_STARTEND option to regexec().
85