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