1 GNU m4 TODO - Tasks that need implementing. -*- outline -*-
2 Copyright (C) 2000, 2001, 2006, 2007 Free Software Foundation, Inc.
4 Tell the maintainers at <bug-m4@gnu.org> if you feel like volunteering
5 for any of these ideas or if you have others to add.
9 + m4wrap.1.test fail on Solaris 2.6 using egcs 1.1a compiler (Erick B).
11 + The test case `other-tests/stackovf.test' does not work.
13 + stack overflow is basically broken
15 The routines to detect stack overflow throuh segv are basically
16 broken. The idea may be fine, but it ends up calling a sigv handler
17 that uses gettext, printf, stdout... all of which are definitely NOT
18 sig-safe. Pity, because the hardcode routine is basically very careful
19 to use write(2), to avoid this.
21 I haven't checked that translation magic is signal-safe.
23 - stdio is not signal-safe in any kind of portable setting anyways.
26 espie@schutzenberger.liafa.jussieu.fr
29 * FEATURES OR PROBLEMS
31 + m4 should keep an ``execution stack'' of macros, which applications could
32 use in their error messages.
34 + Implement discarding comment delimiters with the syntax table.
36 + Implement qindir. Like indir, except that the result of the macro call
37 is not expanded. Because the input stack might contain a file or a
38 string, it is probably best achieved by making note that the TOS input
39 should be copied rather than rescanned.
46 indir(`include', f)dnl
48 qindir(`include', f)dnl
51 patsubst(qindir(`include', f), `b', x)
54 + Use the TOS input quoting for qindir to fix this bug:
61 Stepan Kasal <kasal@ucw.cz>
63 + If configured --with-gmp for multiple precision arithmetic there are
64 some warnings, but it passes the tests.
66 + Make m4 show include dependencies like gcc so Makefile targets are
67 updated when their (included) input files are updated (Erick B).
69 + Add support for wide character sets.
72 * OPTIMIZATION AND CLEAN UP
74 + Have NULs go really undisturbed through GNU m4
75 GNU m4 is lousy regarding NULs in streams (this would require
76 maintaining the string lengths, and avoiding strlen, strcpy,
79 + The argument count limits are handled for all tokens passed around by
80 the internals: we should enable attaching these values to text macros
83 + The context parameter is just a placeholder for formerly global state.
84 We should be making the library reentrant so that multiple instances
85 of m4 can be run in the same process at the same time.
87 + The path management stuff (in path.c/m4private.h) is reinventing the
88 wheel. There are a bunch of fast path management and search functions
89 in ltdl.c: These need to be sanitized, exported through ltdl.h, and
90 then wrapped by the m4module.h path api. path.c can probably be removed
91 entirely at that point.
93 * MODULE SPECIFIC ISSUES
95 + Some way of linking a module statically is needed, for systems
96 without support for dynamic loading.
98 + Some sort of module interface versioning system needs to be implemented
99 in the module loader and the freezer so that m4 can tell if it is being
100 asked to load a frozen file that requires versions of modules with
101 interface versions unsupported by the current release.
103 + Setting of the module search path within m4 scripts:
105 append(__modulepath__, `/some/modules/live/here')
107 + Module autoloader. This would allow an m4 core with no builtins except
108 for loadmodule(). A default startup script would mark the recognised
109 set of builtins for autoload from the installed module directory on first
110 use. A new cli parameter would inhibit initialisation from this script,
111 so that customised m4 interpreters could be built on the fly!
113 + The module loader needs to differentiate between modules that are in
114 memory and modules that are loaded (i.e. visible) from various context
117 + The perl module should only be built if a suitable perl interpreter
118 is found on the build machine.
122 + Automake 1.5 doesn't seem to like pkglibexec_LTLIBRARIES +=
123 See modules/Makefile.am.