2 finish byte <-> utf8 and localencoding <-> utf8 conversions
3 make substr($bytestr,0,0,$charstr) do the right conversion
4 add Unicode::Map equivivalent to core
5 add support for I/O disciplines
6 - a way to specify disciplines when opening things:
7 open(F, "<:crlf :utf16", $file)
8 - a way to specify disciplines for an already opened handle:
9 binmode(STDIN, ":slurp :raw")
10 - a way to set default disciplines for all handle constructors:
11 use open IN => ":any", OUT => ":utf8", SYS => ":utf16"
12 eliminate need for "use utf8;"
13 autoload byte.pm when byte:: is seen by the parser
14 check uv_to_utf8() calls for buffer overflow
15 make \uXXXX (and \u{XXXX}?) where XXXX are hex digits
16 to work similarly to Unicode tech reports and Java
17 notation \uXXXX (and already existing \x{XXXX))?
18 more than four hexdigits? make also \U+XXXX work?
19 See also "Locales", "Regexen", and "Miscellaneous".
22 support "use Thread;" under useithreads
24 - create new interpreter in a different thread
25 - exchange data between interpreters/threads
26 - share namespaces between interpreters/threads
27 work out consistent semantics for exit/die in threads
28 support for externally created threads?
32 auto-produce executable
33 typed lexicals should affect B::CC::load_pad
34 workarounds to help Win32
35 END blocks need saving in compiled output
37 fix comppadlist (names in comppad_name can have fake SvCUR
38 from where newASSIGNOP steals the field)
41 CPP-space: restrict what we export from headers when !PERL_CORE
42 header-space: move into CORE/perl/?
43 API-space: complete the list of things that constitute public api
46 make configuring+building away from source directory work (VPATH et al)
47 this is related to: cross-compilation configuring (see Todo)
48 _r support (see Todo for mode detailed description)
49 POSIX 1003.1 1996 Edition support--realtime stuff:
50 POSIX semaphores, message queues, shared memory, realtime clocks,
51 timers, signals (the metaconfig units mostly already exist for these)
52 PREFERABLY AS AN EXTENSION
53 UNIX98 support: reader-writer locks, realtime/asynchronous IO
54 PREFERABLY AS AN EXTENSION
55 IPv6 support: see RFC2292, RFC2553
56 PREFERABLY AS AN EXTENSION
57 there already is Socket6 in CPAN
60 figure out where the PV->NV->PV conversion gets it wrong at least
61 in AIX and Tru64 (V5.0 and onwards) when using long doubles: see the
62 regexp tricks we had to insert to t/comp/use.t and t/lib/bigfltpm.t,
63 (?:9|8999\d+) and the like.
66 Configure probe for quad_t, uquad_t, and (argh) u_quad_t, they might
67 be in some systems the only thing working as quadtype and uquadtype.
68 more pain: long_long, u_long_long.
71 deprecate traditional/legacy locales?
72 How do locales work across packages?
73 figure out how to support Unicode locales
74 suggestion: integrate the IBM Classes for Unicode (ICU)
75 http://oss.software.ibm.com/developerworks/opensource/icu/project/
76 ICU is "portable, open-source Unicode library with:
77 charset-independent locales (with multiple locales
78 simultaneously supported in same thread; character
79 conversions; formatting/parsing for numbers, currencies,
80 date/time and messages; message catalogs (resources);
81 transliteration, collation, normalization, and text
82 boundaries (grapheme, word, line-break))".
83 Check out also the Locale Converter:
84 http://alphaworks.ibm.com/tech/localeconverter
85 There is also the iconv interface, either from XPG4 or GNU (glibc).
86 iconv is about character set conversions.
87 Either ICU or iconv would be valuable to get integrated
88 into Perl, Configure already probes for libiconv and <iconv.h>.
91 make RE engine thread-safe
92 a way to do full character set arithmetics: now one can do
93 addition, negate a whole class, and negate certain subclasses
94 (e.g. \D, [:^digit:]), but a more generic way to add/subtract/
95 intersect characters/classes, like described in the Unicode technical
96 report on Regular Expression Guidelines,
97 http://www.unicode.org/unicode/reports/tr18/
98 (amusingly, the TR notes that difference and intersection
99 can be done using "Perl-style look-ahead")
100 difference syntax? maybe [[:alpha:][^abc]] meaning
101 "all alphabetic expect a, b, and c"? or [[:alpha:]-[abc]]?
102 (maybe bad, as we explicitly disallow such 'ranges')
103 intersection syntax? maybe [[..]&[...]]?
104 POSIX [=bar=] and [.zap.] would nice too but there's no API for them
105 =bar= could be done with Unicode, though, see the Unicode TR #15 about
107 http://www.unicode.org/unicode/reports/tr15/
108 this is also a part of the Unicode 3.0:
109 http://www.unicode.org/unicode/uni2book/u2.html
110 executive summary: there are several different levels of 'equivalence'
111 trie optimization: factor out common suffixes (and prefixes?)
112 from |-alternating groups (both for exact strings and character
113 classes, use lookaheads?)
117 use fchown, fchmod (and futimes?) internally when possible
118 use fchdir(how portable?)
119 create secure reliable portable temporary file modules
120 audit the standard utilities for security problems and fix them
124 alternate runops() for signal despatch
125 figure out how to die() in delayed sighandler
126 make Thread::Signal work under useithreads
129 sort out the spawnvp() mess for system('a','b','c') compatibility
130 work out DLL versioning
133 introduce @( and @) because group names can have spaces
134 add new modules (Archive::Tar, Compress::Zlib, CPAN::FTP?)
135 sub-second sleep()? alarm()? time()? (integrate Time::HiRes?
136 Configure doesn't yet probe for usleep/nanosleep/ualarm but
138 floating point handling: nans, infinities, fp exception masks, etc.
139 At least the following interfaces exist: fp_classify(), fp_class(),
140 class(), isinf(), isfinite(), finite(), isnormal(), unordered(),
141 <ieeefp.h>, <fp_class.h> (there are metaconfig units for all these),
142 fp_setmask(), fp_getmask(), fp_setround(), fp_getround()
143 (no metaconfig units yet for these).
144 Don't forget finitel(), fp_classl(), fp_class_l(), (yes, both do,
145 unfortunately, exist), and unorderedl().
146 PREFERABLY AS AN EXTENSION.
147 As of 5.6.1 there is cpp macro Perl_isnan().
148 fix the basic arithmetics (+ - * / %) to preserve IVness/UVness if
149 both arguments are IVs/UVs: it sucks that one cannot see
150 the 'carry flag' (or equivalent) of the CPU from C,
151 C is too high-level...
152 replace pod2html with new PodtoHtml? (requires other modules from CPAN)
153 automate testing with large parts of CPAN
154 turn Cwd into an XS module? (Configure already probes for getcwd())
155 mmap for speeding up input? (Configure already probes for the mmap family)
156 sendmsg, recvmsg? (Configure doesn't probe for these but the units exist)
157 setitimer, getitimer? (the metaconfig units exist)
160 SCSU? http://www.unicode.org/unicode/reports/tr6/
161 Collation? http://www.unicode.org/unicode/reports/tr10/
162 Normalization? http://www.unicode.org/unicode/reports/tr15/
163 EBCDIC? http://www.unicode.org/unicode/reports/tr16/
164 Regexes? http://www.unicode.org/unicode/reports/tr18/
165 Case Mappings? http://www.unicode.org/unicode/reports/tr21/
168 keep filenames 8.3 friendly, where feasible
169 upgrade to newer versions of all independently maintained modules
170 comprehensive perldelta.pod
173 describe new age patterns
174 update perl{guts,call,embed,xs} with additions, changes to API
175 convert more examples to use autovivified filehandles
176 document Win32 choices
177 spot-check all new modules for completeness
178 better docs for pack()/unpack()
179 reorg tutorials vs. reference sections
180 make roffitall to be dynamical about its pods and libs