1 Priorities for release:
2 !! MUST finish new deansification scheme...
3 !! documentation (eg new macros)
7 * copyrights on m4 files, aclocal output
9 * BUILT_SOURCES should be examined by automake, %dep_files should be
13 - texinfo/info changes
16 - consider adding "echo"s to installs again.
17 users can use make SHELL='sh -x' to get the full dirt
19 Avoid loops when installing; instead unroll them in automake
21 Franc,ois> * ansi2knr was uselessly compiled, and it might be an
22 Franc,ois> Automake problem:
24 Some long-term projects:
25 * if $(FOO) is used somewhere, ensure FOO is defined, either by
26 user or by automake if possible
27 * Don't rearrange order of `include' lines relative to += assignments.
28 * Handle += assignments at all.
29 * From David A Swierczek: automake should notice (eg) AC_CHECK_PROG and
30 auto-generate Makefile substitution:
31 AC_CHECK_PROG(AR, ...)
32 => AR = @AR@ in Makefile.in
33 Ditto AC_SUBST and other things that make sense. Furthermore, other
34 macros (eg AC_PROG_CC) should be handled via a similar mechanism; many
35 special cases could be eliminated.
37 consider putting all check-* targets onto @check?
38 To support --help/--version checking?
40 must move CONFIG_HEADER from tags.am... allow it to work in subdir.
42 take diff-n-query code from libit
44 must at least partially rewrite dist system (to handle distributing
45 info files, which is currently somewhat broken).
48 Per> 1) Being able to build a set of non-source programs
49 Per> from source programs, without necessarily linking them together.
50 Per> I.e. one should be able to say something like:
51 Per> dummy_SOURCES=foo.c bar.c
52 Per> and automake should realize that it needs to build foo.o and bar.o.
53 Per> 2) Being intelligent about new kinds of suffixes.
55 Per> SUFFIXES = .class .java
56 Per> and a suffix rule of the form:
58 Per> then it should be able to realize it can build .class files from
59 Per> .java files, and thus be able to generate a list of
60 Per> .class files from a list of .java source files.
62 * actually use acinstall program
64 !! Must fix require_file stuff. It is really gross, and I don't
65 understand it any more.
67 * error messages should print ``[info blah blah]'' command when a
68 certain part of the standards apply. saw idea in message from
71 !! should write autoconf-style doc entries for each m4 macro
73 Jim's idea: should look for @setfilename and warn if filenames too long
75 * allow ".info" to be missing
77 * must update GNU Hello
79 ** many requests for a way to omit a file from the distribution.
80 Should be done like `!foo' or `~foo' in _SOURCES, etc.
81 Such files should be removed explicitly after the copy step!
82 Doing this requires rewriting macros before generating Makefile.in.
84 from joerg-martin schwarz:
85 -- If Makefile.am contains $(CC), $(COMPILE), $(YLWRAP), ....
86 in an explicitly written rule, you should emit the corresponding
87 Makefile variables automatically.
89 add support for Makefile.tmpl that is auto-included in every
90 Makefile.am. That makes it easier to do some non-std thing in every
93 consider printing full file name of Makefile.am or configure.in when
94 giving error. This would help for very large trees with many
97 From the GNU Standards. These things could be checked, and probably
99 * Make sure that the directory into which the distribution unpacks (as
100 well as any subdirectories) are all world-writable (octal mode 777).
101 * Make sure that no file name in the distribution is more than 14
103 * Don't include any symbolic links in the distribution itself.
105 * Make sure that all the files in the distribution are world-readable.
106 ** also, check --help output and --version output. Idea from François
108 consider supporting "var+= stuff" syntax. rewrite to just var=... on
109 output. This is sometimes convenient when you want to write a
110 Makefile.am in more-or-less modular parts
112 should be able to determine what is built by looking at rules (and
113 configure.in). Then built man pages (eg) could automatically be
114 omitted from the distribution.
116 Consider using libfoo_SOURCES, etc, for libraries. From Gord
117 Matzigkeit. There is a patch.
119 Idea from Joerg-Martin Schwarz: allow passing different -D flags to
120 different compiles. This can be done, but with the restriction that a
121 .c cannot appear in 2 different "objects" (programs/libraries)
122 compiled with different -D options (because -c and -o do not always
123 work together and parallel makes must work). This could be
124 implemented by noticing whenever a ".o" target with no rules is being
125 emitted, and adding the appropriate compilation rule as appropriate.
126 This should work with targets from Makefile.am as well as from .P
127 files, which means rewriting so that the Makefile.am contents aren't
128 copied into the output immediately. This feature is probably required
129 to fully support libtool ("grody compilation issue")
131 Henrik Frystyk Nielsen says:
132 Henrik> 4) Flags like --include-deps are lost when you make changes to
133 Henrik> Makefile.am files and automake is run automatically. It would
134 Henrik> be nice to keep these flags as I now have to redo everything
136 ... what about other options here too?
138 Think about: maybe "make check" should just bomb if error occurs?
139 Then user must use "make -k check". This is probably more natural.
141 Consider: "cvs" option adds some cvs-specific rules?
143 "Cygnus"-specific features:
144 * Should look for certain tools in the build tree:
145 expect, dejagnu, makeinfo
146 * `no-installinfo' is the default
147 * always generate `info' and `install-info' targets
148 * Allow `texinfo.tex' to be missing
150 Automake: devo/inet/Makefile.am has "all-local". "install" depends on
151 "all", but the local installs get run before the stuff in "all". Gross.
153 Right now, targets generated internally (eg "install") are not
154 overridable by user code. This should probably be possible, even
155 though it isn't very important. This could be done by generating all
156 internal rules via a function call instead of just appending to
159 * Should be a way to have "nobuild_PROGRAMS" which aren't even built,
160 but which could be by running the magic make command.
163 * Must rewrite am_install_var. Should break into multiple functions.
164 This will allow the callers to be a little smarter.
165 * Rewrite clean targets.
166 * Must rewrite error handling code. Right now it is a real mess
167 Should fix up require_file junk at the same time
169 Things to finish libtool support:
170 * Handle grody compilation issue
171 * Handle install changes
172 * Handle clean changes
173 * New definition for LINK
175 Need way to say there are no suffixes in a Makefile (Franc,ois'
176 "override" idea suffices here)
178 Check to make sure various scripts are executable (IE when looking for
181 Use recode in dist target when MAINT_CHARSET specified. Read caveats
182 in automake.in before doing this. Note the same problem used to apply
183 to the no-dependencies option; maybe it still should? Note also that
184 each Makefile.am must be rewritten at "make dist" time if
185 MAINT_CHARSET and DIST_CHARSET are not identical. NOTE: gettext must
186 arrange for all .po files not to be recoded. In the long term this
187 might be a problem (consider when some systems use Unicode but the
189 MAINT_CHARSET *must* be local to each Makefile.am, to enable
190 merged distributions.
191 DIST_CHARSET must be passed down to subdir makes during a "make dist"
193 Handle dist-zoo. Generally add more DOS support. Maybe run "doschk"
194 (why isn't this merged with "pathchk"?) when doing a dist. Do
195 whatever else François says here...
197 Add support for html via an option. Use texi2html. Use
198 "html_TEXINFOS", and htmldir = .../html. Include html files in
199 distribution. Also allow "html_DATA", for raw .html files.
200 [ when will texinfo directly support html? ]
202 uninstall and pkg-dirs should rm -rf the dir.
204 a potential bug: configure puts "blah.o" into LIBOBJS, thus implying
205 these files can't be de-ansified. Not a problem?
206 [ fix by using ansi2knr wrapper program ]
208 In general most .am files should be merged into automake. For
209 instance all the "clean" targets could be merged by keeping lists of
210 things to be removed. This would be a lot nicer looking. Note that
211 the install targets probably should not be merged; it is sometimes
212 useful to only install a small part.
215 * Order rules sensibly
216 * Ensure every line has a purpose. Omit unused stuff
217 * Eliminate extraneous rules when possible (eg 'install-am' stuff)
218 * Make sure vertical spacing is correct
219 * pretty-print targets
220 * regularize how backslash-newline is done. Just one space between text
221 and backslash should be the rule. Update makefile-mode to allow this.
222 (set column to 0, probably)
223 Omit program transform vars from header if no program installed. This
224 is currently pretty hard to do. (But with beautification code it
225 would probably be easy)
228 * It would be nice to automatically support using bison's better features
229 to rename the output files. This requires autoconf support
230 * Consider supporting syntax from autoconf "derived:source", eg:
232 for yacc and lex source
234 Multi-language support:
235 * should have mapping of file extensions to languages
236 * should automatically handle the linking issue (special-case C++)
237 * must get compile rules for various languages; FORTRAN probably
238 most important unimplemented language
239 This should be integrated in some way with Per's idea.
240 Eg .f.o rules should be recognized & auto-handled in _SOURCES
241 That way any random language can be treated with C/C++ on a first-class
244 It might be cool to generate .texi dependencies by grepping for
245 @include. (If done, it should be done the same way C dependencies are
248 It would be good to check some parts of GNU standards. Already check
249 for install-sh and mkinstalldirs. What else is required to be in
250 package by GNU standards or by automake?
251 Some things for --strictness=gnits:
252 * "cd $(foo); something" is an error in a rule. Should be:
253 "cd $(foo) && something"
254 * Look for 'ln -s' and warn about using $(LN) and AC_PROG_LN_S
255 * Look for $(LN) and require AC_PROG_LN_S
257 Auto-distribute "ChangeLog.[0-9]+"? "ChangeLog.[a-z]+"?
259 Internationalize. [ gettext doesn't have the necessary machinery yet ]
260 am_error should use printf-style arguments (for eventual gettext scheme)
262 François says the ordering of files in a distribution should be as follows:
266 I agree, but I don't see how to implement this yet.
267 It might be easier if "derived files" is limited to those that
268 Automake itself knows about, eg output of yacc.
270 Check all source files to make sure that FSF address is up-to-date.
271 --gnits or --gnu only.
273 Merge each -vars.am file with corresponding ".am" file. Can do this
274 because of changes to &file_contents.
276 Should libexec programs have the name transform done on them?
278 Order the output rules sensibly, so FOO_SOURCES and FOO_OBJECTS are
279 together and rules are in the usual order.
281 Make the output minimal: only output definitions for variables that
284 Look at dist's jmake for ideas. dist is the name of the distribution
285 including Metaconfig. Perl uses it.
287 Should handle directory hierarchies deeper than 2. Right now there is
288 some support for this. Here are some of the issues:
289 * Should handle AC_CONFIG_SUBDIRS, ie must handle configure.in in subdirs
290 * can do this by looking at subdirs, seeing configure.in
291 and auto-running Automake there
293 . Consider supporting guile-style PLUGIN directories automatically?
296 David> To avoid comments like the one about subdirs getting buried in
297 David> the middle of a Makefile.in, how about pushing comments that
298 David> start with ### to the top of the Makefile.in (in order)? Sort
299 David> of like how Autoconf uses diversions to force initialization
300 David> code to the top of configure.
303 Karl> 2) Your Makefile variable names are generally uppercase, but GNU
304 Karl> generally uses lowercase. Not that it matters :-).
306 ================================================================
310 probably should put each group of m4 files into a subdir owned by the
311 containing application.
313 ================================================================
317 document which variables are actually scanned and which are not.
321 Document customary ordering of Makefile.am. From François.
323 Should include extended version of diagram from Autoconf (suggested by
326 Make a definition of the term "source"
328 need xref to libtool in docs
330 document how to use Automake with CVS. Idea from Mark Galassi. Also
331 include Greg Woods' more sophisticated "cvs-dist" target.
333 document rebuilding configure. CONFIGURE_DEPENDENCIES
335 -- must document all variables that are supposed
336 to be public knowledge
338 automake must be run in each directory with a configure.in
339 This is insufficiently clear
341 must document the targets required for integration with
342 non-automake-using subdirs
344 use of (eg) EXTRA_PROGRAMS is not very clear right now
346 document EXTRA_DIST_DIRS, distributing things in subdirs
349 ================================================================
351 Things to do for autoconf:
353 * allow random code to be inserted a la the 2nd arg to AC_OUTPUT:
354 Joel> I know that the following is needed at the end of configure.in:
355 Joel> [test -z "$CONFIG_HEADERS" || echo timestamp >stamp-h])
356 Joel> However, if automake checked that this line is present, it would
357 Joel> help...this bit me for a while.
359 * patch autoreconf to run automake and aclocal (this is done but
360 not really available)
362 ================================================================
366 * Should support standalone library along with subdir library in same
367 Makefile.am. Maybe: turn off "standalone" mode if library's Makefile.am
368 is not only one specd? [ add an option for this ]
370 ================================================================
374 Would it be useful to integrate in some way with the Debian package
375 building utility? Must check. maybe it would be possible to deal
376 with all the different package utilities somehow. Lately I've been
377 hearing good things about the RedHat packaging utilities. Why are
378 there so many of these? Are they fun to write or something?
379 The RedHat package utility is called RPM; see
380 ftp://ftp.redhat.com/pub/code/rpm
381 It actually has problems, like no configure script and no documentation.
383 For Cygnus it would probably be good to be able to handle the native
384 package utility on each platform. There are probably 3 or 4 of these
385 (sysv, solaris?, aix?)
387 ================================================================
389 A tool to guess what the local Makefile.am should look like:
390 (see Gord's Maint program!)
392 * Probably integrate with autoscan
393 * Use various simple rules to determine what to do:
394 * get name of top directory, sans version info
395 * search for .c files with 'main' in them
396 * if in main.c, use directory name for program
397 * if in more than one, generate multiple programs
398 * if not found, generate a library named after directory
399 * order subdir searches correctly: lib first, src last
400 * assume 'testsuite' dir means we are using dejagnu
401 * maybe be smart about reading existing Makefile.am, so tool
402 can be run for incremental changes? You could imagine:
405 autoproject --incremental
407 ================================================================
409 Stuff NOT to do, and why:
411 consider auto-including any file that matches "*.in".
412 [ no: po/Makefile.in shouldn't be included ]
414 must look at mkid to see how it works (for subdir usage)
415 [ right now, it doesn't. i don't see a simple fix right now ]
417 if configure.in not found, move up a directory and try again? This
418 could eliminate a common source of problems.
419 [ this is just a bad idea ]
421 * scripts are installed in $exec_prefix/bin, not $prefix/bin
423 [ the consensus on Gnits is that this isn't required.
424 doubters can work around it anyway ]
426 * make the auto-dep code crash if GNU make not in use?
427 (doesn't it already?)
429 Looked at a program called 'ezmake', which seems to do something
430 similar. The only idea there that is possibly worth stealing is using
431 globs in definitions. Also has negations. Eg in a directory with
432 files a.c, b.c and c.c, the line:
433 foo_SOURCES = *.c ~c.c
434 would be equivalent to:
435 foo_SOURCES = a.c b.c
436 Is this worth implementing?
437 [ No... it is more reliable to spell everything out. ]
439 Scan source directories and warn about missing files, eg .c/.h files
440 that aren't mentioned?
441 [ distcheck makes this less useful ]