1 Priorities for release:
2 * copyrights on m4 files, aclocal output
4 * should not put texiname_TEXINFOS into distribution
5 should rename this macro anyway, to foo_texi_DEPENDENCIES
7 For now I guess I'll just have automake give an error if it encounters
8 non-C source in a libtool library specification.
10 * must split $obj into two parts: one for libtool and one for
11 deansification. Otherwise .S files will be deansified!
13 * if program has the same name as a target, do something sensible:
14 - if the target is internal, rename it
15 - if the target is mandated (eg, "info"), tell the user
16 consider auto-modifying the program name to work around this
18 * should separate actual options from strictness levels
19 strictness should only cover requirements
20 You should be able to pick and choose options
22 * cygwin32 support: add .exe to all executables; discover cygwin32
23 without using ac_canonical_*.
25 should clean up texinfos.am; one rule is repeated 3 times, but
28 should always use perl -w
34 * put parser.h into distribution if "yacc -d" is used
36 * only remove libtool at top level?
38 * clean up source directory by moving stuff into subdirs
40 * consider adding pkglibexecdir, maybe others?
41 requests for pkg-dirs with version included
43 Fix ansi2knr so that knr-style function decls aren't incorrectly
44 rewritten. Or, fix automake so that not all sources in a directory
45 must be ansi. Should also fix things so that a separate ._o file is
46 not needed; instead use a wrapper script.
52 Avoid loops when installing; instead unroll them in automake
55 * completely handle multi-":" mode for AC_CONFIG_HEADER
56 * Scan multiple input files when Makefile is generated?
57 This would provide flexibility for large projects; subsumes
58 the "Makefile.tmpl" idea
60 [ can't do this. must explain why in manual.
61 basically, solving all the problems is too hard
62 like: how to remove redundancies between generated .in files
63 instead should implement `include' directive for Makefile.am ]
64 * for multi-":" mode and AC_OUTPUT, it might be good to pick the
65 first input file that has a corresponding .am file.
67 Some long-term projects:
68 * if $(FOO) is used somewhere, ensure FOO is defined, either by
69 user or by automake if possible
70 * Don't rearrange order of `include' lines relative to += assignments.
71 * Handle += assignments at all.
73 consider putting all check-* targets onto @check?
74 To support --help/--version checking?
76 take diff-n-query code from libit
79 Per> 1) Being able to build a set of non-source programs
80 Per> from source programs, without necessarily linking them together.
81 Per> I.e. one should be able to say something like:
82 Per> dummy_SOURCES=foo.c bar.c
83 Per> and automake should realize that it needs to build foo.o and bar.o.
84 Per> 2) Being intelligent about new kinds of suffixes.
86 Per> SUFFIXES = .class .java
87 Per> and a suffix rule of the form:
89 Per> then it should be able to realize it can build .class files from
90 Per> .java files, and thus be able to generate a list of
91 Per> .class files from a list of .java source files.
94 * allow a way to use "gzip --best"
95 [ Try GZIP=--best make dist ]
96 * don't assume GNU tar is "tar" (eg in distcheck)
98 !! Must fix require_file stuff. It is really gross, and I don't
99 understand it any more.
101 * error messages should print ``[info blah blah]'' command when a
102 certain part of the standards apply. saw idea in message from
103 Craig Burley. wouldn't it be really cool if compile-mode in Emacs
104 understood this convention, and you could click on such text to
105 go to the appropriate info page?
107 !! should write autoconf-style doc entries for each m4 macro
109 Jim's idea: should look for @setfilename and warn if filenames too long
111 * allow ".info" to be missing
113 * must update GNU Hello
115 ** many requests for a way to omit a file from the distribution.
116 Should be done like `!foo' or `~foo' in _SOURCES, etc.
117 Such files should be removed explicitly after the copy step!
118 Doing this requires rewriting macros before generating Makefile.in.
120 from joerg-martin schwarz:
121 -- If Makefile.am contains $(CC), $(COMPILE), $(YLWRAP), ....
122 in an explicitly written rule, you should emit the corresponding
123 Makefile variables automatically.
125 Configuring in the large:
126 * allow hierarchy of dirs to share one aclocal.m4
129 consider printing full file name of Makefile.am or configure.in when
130 giving error. This would help for very large trees with many
133 From the GNU Standards. These things could be checked, and probably
135 * Make sure that the directory into which the distribution unpacks (as
136 well as any subdirectories) are all world-writable (octal mode 777).
137 * Make sure that no file name in the distribution is more than 14
139 * Don't include any symbolic links in the distribution itself.
141 * Make sure that all the files in the distribution are world-readable.
142 ** also, check --help output and --version output. Idea from François
143 * standards no longer prohibit ANSI C. What does this imply
144 for the de-ansi-fication feature?
146 consider supporting "var+= stuff" syntax. rewrite to just var=... on
147 output. This is sometimes convenient when you want to write a
148 Makefile.am in more-or-less modular parts
150 should be able to determine what is built by looking at rules (and
151 configure.in). Then built man pages (eg) could automatically be
152 omitted from the distribution.
154 Idea from Joerg-Martin Schwarz: allow passing different -D flags to
155 different compiles. This can be done, but with the restriction that a
156 .c cannot appear in 2 different "objects" (programs/libraries)
157 compiled with different -D options (because -c and -o do not always
158 work together and parallel makes must work). This could be
159 implemented by noticing whenever a ".o" target with no rules is being
160 emitted, and adding the appropriate compilation rule as appropriate.
161 This should work with targets from Makefile.am as well as from .P
162 files, which means rewriting so that the Makefile.am contents aren't
163 copied into the output immediately.
164 [ this could be probably done more directly by examining the sources
165 as we scan Makefile.am ]
167 Henrik Frystyk Nielsen says:
168 Henrik> 4) Flags like --include-deps are lost when you make changes to
169 Henrik> Makefile.am files and automake is run automatically. It would
170 Henrik> be nice to keep these flags as I now have to redo everything
172 ... what about other options here too?
174 Think about: maybe "make check" should just bomb if error occurs?
175 Then user must use "make -k check". This is probably more natural.
177 Consider: "cvs" option adds some cvs-specific rules?
179 Right now, targets generated internally (eg "install") are not
180 overridable by user code. This should probably be possible, even
181 though it isn't very important. This could be done by generating all
182 internal rules via a function call instead of just appending to
184 [ this will be harder to implement when scanning a rule like all-recursive
187 * Should be a way to have "nobuild_PROGRAMS" which aren't even built,
188 but which could be by running the magic make command.
191 * Must rewrite am_install_var. Should break into multiple functions.
192 This will allow the callers to be a little smarter.
193 * Rewrite clean targets.
194 * Must rewrite error handling code. Right now it is a real mess
195 Should fix up require_file junk at the same time
197 djm wants ``LINKS'' variable; list of things to link together after
198 install. In BSD environment, use:
199 LINKS = from1 to1 from2 to2 ...
201 Need way to say there are no suffixes in a Makefile (Franc,ois'
202 "override" idea suffices here)
204 Check to make sure various scripts are executable (IE when looking for
207 Use recode in dist target when MAINT_CHARSET specified. Read caveats
208 in automake.in before doing this. Note the same problem used to apply
209 to the no-dependencies option; maybe it still should? Note also that
210 each Makefile.am must be rewritten at "make dist" time if
211 MAINT_CHARSET and DIST_CHARSET are not identical. NOTE: gettext must
212 arrange for all .po files not to be recoded. In the long term this
213 might be a problem (consider when some systems use Unicode but the
215 MAINT_CHARSET *must* be local to each Makefile.am, to enable
216 merged distributions.
217 DIST_CHARSET must be passed down to subdir makes during a "make dist"
219 Handle dist-zoo. Generally add more DOS support. Maybe run "doschk"
220 (why isn't this merged with "pathchk"?) when doing a dist. Do
221 whatever else François says here...
223 Add support for html via an option. Use texi2html. Use
224 "html_TEXINFOS", and htmldir = .../html. Include html files in
225 distribution. Also allow "html_DATA", for raw .html files.
226 [ when will texinfo directly support html? ]
228 uninstall and pkg-dirs should rm -rf the dir.
230 a potential bug: configure puts "blah.o" into LIBOBJS, thus implying
231 these files can't be de-ansified. Not a problem?
232 [ fix by using ansi2knr wrapper program ]
234 In general most .am files should be merged into automake. For
235 instance all the "clean" targets could be merged by keeping lists of
236 things to be removed. This would be a lot nicer looking. Note that
237 the install targets probably should not be merged; it is sometimes
238 useful to only install a small part.
241 * Order rules sensibly
242 * Ensure every line has a purpose. Omit unused stuff
243 * Eliminate extraneous rules when possible (eg 'install-am' stuff)
244 * Make sure vertical spacing is correct
245 Omit program transform vars from header if no program installed. This
246 is currently pretty hard to do. (But with beautification code it
247 would probably be easy)
250 * It would be nice to automatically support using bison's better features
251 to rename the output files. This requires autoconf support
252 * Consider supporting syntax from autoconf "derived:source", eg:
254 for yacc and lex source
255 * what if you use flex and the option to avoid -lfl?
258 Multi-language support:
259 * should have mapping of file extensions to languages
260 * should automatically handle the linking issue (special-case C++)
261 * must get compile rules for various languages; FORTRAN probably
262 most important unimplemented language
263 This should be integrated in some way with Per's idea.
264 Eg .f.o rules should be recognized & auto-handled in _SOURCES
265 That way any random language can be treated with C/C++ on a first-class
268 It might be cool to generate .texi dependencies by grepping for
269 @include. (If done, it should be done the same way C dependencies are
272 It would be good to check some parts of GNU standards. Already check
273 for install-sh and mkinstalldirs. What else is required to be in
274 package by GNU standards or by automake?
275 Some things for --strictness=gnits:
276 * "cd $(foo); something" is an error in a rule. Should be:
277 "cd $(foo) && something"
278 * Look for 'ln -s' and warn about using $(LN) and AC_PROG_LN_S
279 * Look for $(LN) and require AC_PROG_LN_S
281 Auto-distribute "ChangeLog.[0-9]+"? "ChangeLog.[a-z]+"?
283 Internationalize. [ gettext doesn't have the necessary machinery yet ]
284 am_error should use printf-style arguments (for eventual gettext scheme)
286 François says the ordering of files in a distribution should be as follows:
290 I agree, but I don't see how to implement this yet.
291 It might be easier if "derived files" is limited to those that
292 Automake itself knows about, eg output of yacc.
294 Check all source files to make sure that FSF address is up-to-date.
295 --gnits or --gnu only.
297 Merge each -vars.am file with corresponding ".am" file. Can do this
298 because of changes to &file_contents.
300 Should libexec programs have the name transform done on them?
302 Order the output rules sensibly, so FOO_SOURCES and FOO_OBJECTS are
303 together and rules are in the usual order.
305 Make the output minimal: only output definitions for variables that
308 Look at dist's jmake for ideas. dist is the name of the distribution
309 including Metaconfig. Perl uses it.
311 Should handle directory hierarchies deeper than 2. Right now there is
312 some support for this. Here are some of the issues:
313 * Should handle AC_CONFIG_SUBDIRS, ie must handle configure.in in subdirs
314 * can do this by looking at subdirs, seeing configure.in
315 and auto-running Automake there
318 David> To avoid comments like the one about subdirs getting buried in
319 David> the middle of a Makefile.in, how about pushing comments that
320 David> start with ### to the top of the Makefile.in (in order)? Sort
321 David> of like how Autoconf uses diversions to force initialization
322 David> code to the top of configure.
325 Karl> 2) Your Makefile variable names are generally uppercase, but GNU
326 Karl> generally uses lowercase. Not that it matters :-).
328 ================================================================
332 probably should put each group of m4 files into a subdir owned by the
333 containing application.
335 ================================================================
339 SUBDIR entry must be direct subdir of this dir
343 multi-":" mode in AC_OUTPUT -- automake only looks at the first file
344 also a note on how a .am file is found in this case
346 rationale for avoiding
347 make CFLAGS="$CFLAGS" ...
350 a package that installs its own aclocal macros
352 write example of using automake with dejagnu
353 follow calc example in dejagnu docs
355 document which variables are actually scanned and which are not.
357 Document customary ordering of Makefile.am. From François.
359 Should include extended version of diagram from Autoconf (suggested by
362 Make a definition of the term "source"
364 document how to use Automake with CVS. Idea from Mark Galassi. Also
365 include Greg Woods' more sophisticated "cvs-dist" target.
367 document rebuilding configure. CONFIGURE_DEPENDENCIES
369 -- must document all variables that are supposed
370 to be public knowledge
372 automake must be run in each directory with a configure.in
373 This is insufficiently clear
375 must document the targets required for integration with
376 non-automake-using subdirs
378 use of (eg) EXTRA_PROGRAMS is not very clear right now
379 document EXTRA_foo_SOURCES
380 document why EXTRA_* vars must be statically knowable
382 document the "make SHELL='/bin/sh -x'" trick for debugging
384 section on relationship to GNU make
388 ================================================================
390 Things to do for autoconf:
392 * patch autoreconf to run automake and aclocal. I've done this but it is
393 not really available. It can't be made available until automake
394 is officially released
396 ================================================================
400 * Should support standalone library along with subdir library in same
401 Makefile.am. Maybe: turn off "standalone" mode if library's Makefile.am
402 is not only one specd? [ add an option for this ]
404 ================================================================
408 Would it be useful to integrate in some way with the Debian package
409 building utility? Must check. maybe it would be possible to deal
410 with all the different package utilities somehow. Lately I've been
411 hearing good things about the RedHat packaging utilities. Why are
412 there so many of these? Are they fun to write or something?
413 The RedHat package utility is called RPM; see
414 ftp://ftp.redhat.com/pub/code/rpm
415 It actually has problems, like no configure script and no documentation.
417 For Cygnus it would probably be good to be able to handle the native
418 package utility on each platform. There are probably 3 or 4 of these
419 (sysv, solaris?, aix?)
421 tcl/unix/Makefile.in has some code to generate a Solaris package.
423 Automake probably can't do all of this on its own. A new tool might
426 ================================================================
428 A tool to guess what the local Makefile.am should look like:
429 (see Gord's Maint program!)
431 * Probably integrate with autoscan
432 * Use various simple rules to determine what to do:
433 * get name of top directory, sans version info
434 * search for .c files with 'main' in them
435 * if in main.c, use directory name for program
436 * if in more than one, generate multiple programs
437 * if not found, generate a library named after directory
438 * order subdir searches correctly: lib first, src last
439 * assume 'testsuite' dir means we are using dejagnu
440 * maybe be smart about reading existing Makefile.am, so tool
441 can be run for incremental changes? You could imagine:
444 autoproject --incremental
446 ================================================================
448 Stuff NOT to do, and why:
450 consider auto-including any file that matches "*.in".
451 [ no: po/Makefile.in shouldn't be included ]
453 must look at mkid to see how it works (for subdir usage)
454 [ right now, it doesn't. i don't see a simple fix right now ]
456 if configure.in not found, move up a directory and try again? This
457 could eliminate a common source of problems.
458 [ this is just a bad idea ]
460 * scripts are installed in $exec_prefix/bin, not $prefix/bin
462 [ the consensus on Gnits is that this isn't required.
463 doubters can work around it anyway ]
465 * make the auto-dep code crash if GNU make not in use?
466 (doesn't it already?)
468 Looked at a program called 'ezmake', which seems to do something
469 similar. The only idea there that is possibly worth stealing is using
470 globs in definitions. Also has negations. Eg in a directory with
471 files a.c, b.c and c.c, the line:
472 foo_SOURCES = *.c ~c.c
473 would be equivalent to:
474 foo_SOURCES = a.c b.c
475 Is this worth implementing?
476 [ No... it is more reliable to spell everything out. ]
478 Scan source directories and warn about missing files, eg .c/.h files
479 that aren't mentioned?
480 [ distcheck makes this less useful ]