1 Priorities for release:
11 must move CONFIG_HEADER from tags.am... allow it to work in subdir.
13 take diff-n-query code from libit
15 must at least partially rewrite dist system (to handle distributing
16 info files, which is currently somewhat broken).
19 Per> 1) Being able to build a set of non-source programs
20 Per> from source porgrams, without necessarily linking them together.
21 Per> I.e. one should be able to say something like:
22 Per> dummy_SOURCES=foo.c bar.c
23 Per> and automake should realize that it needs to build foo.o and bar.o.
24 Per> 2) Being intelligent about new kinds of suffixes.
26 Per> SUFFIXES = .class .java
27 Per> and a suffix rule of the form:
29 Per> then it should be able to realize it can build .class files from
30 Per> .java files, and thus be able to generate a list of
31 Per> .class files from a list of .java source files.
33 * fix this distribution to be fully standards compliant
34 * pathchk says we run over the limit
35 * must change aclocal's implementation somehow -- gross
37 * actually use acinstall program
39 !! Must fix require_file stuff. It is really gross, and I don't
40 understand it any more.
42 * error messages should print ``[info blah blah]'' command when a
43 certain part of the standards apply. saw idea in message from
46 * patch from Joel Weber about fixing yacc; in particular generating .h file
48 !! should write autoconf-style doc entries for each m4 macro
50 [ this should really go into autoconf in some automatic way ]
51 Joel> I know that the following is needed at the end of configure.in:
52 Joel> [test -z "$CONFIG_HEADERS" || echo timestamp >stamp-h])
53 Joel> However, if automake checked that this line is present, it would
54 Joel> help...this bit me for a while.
56 Jim's idea: should look for @setfilename and warn if filenames too long
58 * allow ".info" to be missing
60 should put inverse of @MAINT@ before `.PHONY: configure'. This means
61 fixing configure target name (no $srcdir)
63 * must update GNU Hello
65 ** many requests for a way to omit a file from the distribution.
66 Should be done like `!foo' or `~foo' in _SOURCES, etc.
67 Such files should be removed explicitly after the copy step!
68 Doing this requires rewriting macros before generating Makefile.in.
70 add support for Makefile.tmpl that is auto-included in every
71 Makefile.am. That makes it easier to do some non-std thing in every
74 consider printing full file name of Makefile.am or configure.in when
75 giving error. This would help for very large trees with many
78 From the GNU Standards. These things could be checked, and probably
80 * Make sure that the directory into which the distribution unpacks (as
81 well as any subdirectories) are all world-writable (octal mode 777).
82 * Make sure that no file name in the distribution is more than 14
84 * Don't include any symbolic links in the distribution itself.
86 * Make sure that all the files in the distribution are world-readable.
87 ** also, check --help output and --version output. Idea from François
89 consider supporting "var+= stuff" syntax. rewrite to just var=... on
90 output. This is sometimes convenient when you want to write a
91 Makefile.am in more-or-less modular parts
93 should be able to determine what is built by looking at rules (and
94 configure.in). Then built man pages (eg) could automatically be
95 omitted from the distribution.
97 Consider using libfoo_SOURCES, etc, for libraries. From Gord
98 Matzigkeit. There is a patch.
100 Idea from Joerg-Martin Schwarz: allow passing different -D flags to
101 different compiles. This can be done, but with the restriction that a
102 .c cannot appear in 2 different "objects" (programs/libraries)
103 compiled with different -D options (because -c and -o do not always
104 work together and parallel makes must work). This could be
105 implemented by noticing whenever a ".o" target with no rules is being
106 emitted, and adding the appropriate compilation rule as appropriate.
107 This should work with targets from Makefile.am as well as from .P
108 files, which means rewriting so that the Makefile.am contents aren't
109 copied into the output immediately. This feature is probably required
110 to fully support libtool ("grody compilation issue")
112 Henrik Frystyk Nielsen says:
113 Henrik> 4) Flags like --include-deps are lost when you make changes to
114 Henrik> Makefile.am files and automake is run automatically. It would
115 Henrik> be nice to keep these flags as I now have to redo everything
117 ... what about other options here too?
119 Think about: maybe "make check" should just bomb if error occurs?
120 Then user must use "make -k check". This is probably more natural.
122 Consider: "cvs" option adds some cvs-specific rules?
124 "Cygnus"-specific features:
125 * An option that statically rewrites @MAINT@ to "#M#".
126 * Should look for certain tools in the build tree:
127 expect, dejagnu, makeinfo
128 * `no-installinfo' is the default
129 * always generate `info' and `install-info' targets
130 * Allow `texinfo.tex' to be missing
132 Automake: devo/inet/Makefile.am has "all-local". "install" depends on
133 "all", but the local installs get run before the stuff in "all". Gross.
135 Right now, targets generated internally (eg "install") are not
136 overridable by user code. This should probably be possible, even
137 though it isn't very important. This could be done by generating all
138 internal rules via a function call instead of just appending to
141 * Should be a way to have "nobuild_PROGRAMS" which aren't even built,
142 but which could be by running the magic make command.
144 * Should have tool like "autoreconf" that only remakes Makefiles that
145 need it. Probably autoreconf should be modified to handle automake
148 * Must rewrite am_install_var. Should break into multiple functions.
149 This will allow the callers to be a little smarter.
150 * Rewrite clean targets.
151 * Must rewrite error handling code. Right now it is a real mess
152 Should fix up require_file junk at the same time
154 Things to finish libtool support:
155 * Handle grody compilation issue
156 * Handle install changes
157 * Handle clean changes
158 * New definition for LINK
160 Scan source directories and warn about missing files, eg .c/.h files
161 that aren't mentioned?
163 Need way to say there are no suffixes in a Makefile (Franc,ois'
164 "override" idea suffices here)
166 Check to make sure various scripts are executable (IE when looking for
169 Use recode in dist target when MAINT_CHARSET specified. Read caveats
170 in automake.in before doing this. Note the same problem used to apply
171 to the no-dependencies option; maybe it still should? Note also that
172 each Makefile.am must be rewritten at "make dist" time if
173 MAINT_CHARSET and DIST_CHARSET are not identical. NOTE: gettext must
174 arrange for all .po files not to be recoded. In the long term this
175 might be a problem (consider when some systems use Unicode but the
177 MAINT_CHARSET *must* be local to each Makefile.am, to enable
178 merged distributions.
179 DIST_CHARSET must be passed down to subdir makes during a "make dist"
181 Handle dist-zoo. Generally add more DOS support. Maybe run "doschk"
182 (why isn't this merged with "pathchk"?) when doing a dist. Do
183 whatever else François says here...
185 Add support for html via an option. Use texi2html. Use
186 "html_TEXINFOS", and htmldir = .../html. Include html files in
187 distribution. Also allow "html_DATA", for raw .html files.
188 [ when will texinfo support html? ]
189 [ is there a texinfo.tex that supports texi2html extensions? ]
191 uninstall and pkg-dirs should rm -rf the dir.
193 a potential bug: configure puts "blah.o" into LIBOBJS, thus implying
194 these files can't be de-ansified. Not a problem?
196 In general most .am files should be merged into automake. For
197 instance all the "clean" targets could be merged by keeping lists of
198 things to be removed. This would be a lot nicer looking. Note that
199 the install targets probably should not be merged; it is sometimes
200 useful to only install a small part.
203 * Order rules sensibly
204 * Ensure every line has a purpose. Omit unused stuff
205 * Eliminate extraneous rules when possible (eg 'install-am' stuff)
206 * Make sure vertical spacing is correct
207 * pretty-print targets
208 * regularize how backslash-newline is done. Just one space between text
209 and backslash should be the rule. Update makefile-mode to allow this.
210 (set column to 0, probably)
211 Omit program transform vars from header if no program installed. This
212 is currently pretty hard to do. (But with beautification code it
213 would probably be easy)
216 * It would be nice to automatically support using bison's better features
217 to rename the output files. This requires autoconf support
218 * Consider supporting syntax from autoconf "derived:source", eg:
220 for yacc and lex source
221 * if AC_PROG_LEX used, ensure (no, *PUT*) LEXLIB in foo_LDADD
223 Multi-language support:
224 * should have mapping of file extensions to languages
225 * should automatically handle the linking issue (special-case C++)
226 * must get compile rules for various languages; FORTRAN probably
227 most important unimplemented language
229 Should 'distclean' "rm -rf .deps"?
231 It might be cool to generate .texi dependencies by grepping for
232 @include. (If done, it should be done the same way C dependencies are
235 It would be good to check some parts of GNU standards. Already check
236 for install-sh and mkinstalldirs. What else is required to be in
237 package by GNU standards or by automake?
238 Some things for --strictness=gnits:
239 * "cd $(foo); something" is an error in a rule. Should be:
240 "cd $(foo) && something"
241 * Look for 'ln -s' and warn about using $(LN) and AC_PROG_LN_S
242 * Look for $(LN) and require AC_PROG_LN_S
244 automake.in: should ".cc" really -> "$o"? This doesn't really seem
245 right, but maybe it is so names can be rewritten uniformly? Must
248 Auto-distribute "ChangeLog.[0-9]+"? "ChangeLog.[a-z]+"?
250 Internationalize. [ gettext doesn't have the necessary machinery yet ]
251 am_error should use printf-style arguments (for eventual gettext scheme)
253 François says the ordering of files in a distribution should be as follows:
257 I agree, but I don't see how to implement this yet.
258 It might be easier if "derived files" is limited to those that
259 Automake itself knows about, eg output of yacc.
261 Check all source files to make sure that FSF address is up-to-date.
262 --gnits or --gnu only.
264 Merge each -vars.am file with corresponding ".am" file. Can do this
265 because of changes to &file_contents.
267 Looked at a program called 'ezmake', which seems to do something
268 similar. The only idea there that is possibly worth stealing is using
269 globs in definitions. Also has negations. Eg in a directory with
270 files a.c, b.c and c.c, the line:
271 foo_SOURCES = *.c ~c.c
272 would be equivalent to:
273 foo_SOURCES = a.c b.c
274 Is this worth implementing?
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 These can both be handled via dist-hook:
294 . Consider supporting guile-style PLUGIN directories automatically?
295 . Consider allowing eg "foo/bar" to appear in EXTRA_DIST, and generating
296 code to make directory foo at dist time
299 David> To avoid comments like the one about subdirs getting buried in
300 David> the middle of a Makefile.in, how about pushing comments that
301 David> start with ### to the top of the Makefile.in (in order)? Sort
302 David> of like how Autoconf uses diversions to force initialization
303 David> code to the top of configure.
306 Karl> 2) Your Makefile variable names are generally uppercase, but GNU
307 Karl> generally uses lowercase. Not that it matters :-).
309 ================================================================
313 probably should put each group of m4 files into a subdir owned by the
314 containing application.
316 must fill in definitions for some of the AC_FEATURE macros
318 consider including autosystem; that is the only way the AC_FEATURE
319 macros even make sense. Or move the AC_FEATURE macros back into
322 ================================================================
326 document which variables are actually scanned and which are not.
330 Document customary ordering of Makefile.am. From François.
332 Should include extended version of diagram from Autoconf (suggested by
335 Make a definition of the term "source"
337 need xref to libtool in docs
339 document how to use Automake with CVS. Idea from Mark Galassi. Also
340 include Greg Woods' more sophisticated "cvs-dist" target.
342 document rebuilding configure. CONFIGURE_DEPENDENCIES
344 -- must document all variables that are supposed
345 to be public knowledge
347 automake must be run in each directory with a configure.in
348 This is insufficiently clear
350 must document the targets required for integration with
351 non-automake-using subdirs
353 use of (eg) EXTRA_PROGRAMS is not very clear right now
355 ================================================================
359 * Should support standalone library along with subdir library in same
360 Makefile.am. Maybe: turn off "standalone" mode if library's Makefile.am
361 is not only one specd? [ add an option for this ]
363 ================================================================
367 Would it be useful to integrate in some way with the Debian package
368 building utility? Must check. maybe it would be possible to deal
369 with all the different package utilities somehow. Lately I've been
370 hearing good things about the RedHat packaging utilities. Why are
371 there so many of these? Are they fun to write or something?
372 The RedHat package utility is called RPM; see
373 ftp://ftp.redhat.com/pub/code/rpm
374 It actually has problems, like no configure script and no documentation.
376 For Cygnus it would probably be good to be able to handle the native
377 package utility on each platform. There are probably 3 or 4 of these
378 (sysv, solaris?, aix?)
380 ================================================================
382 A tool to guess what the local Makefile.am should look like:
383 (see Gord's Maint program!)
385 * Probably integrate with autoscan
386 * Use various simple rules to determine what to do:
387 * get name of top directory, sans version info
388 * search for .c files with 'main' in them
389 * if in main.c, use directory name for program
390 * if in more than one, generate multiple programs
391 * if not found, generate a library named after directory
392 * order subdir searches correctly: lib first, src last
393 * assume 'testsuite' dir means we are using dejagnu
394 * maybe be smart about reading existing Makefile.am, so tool
395 can be run for incremental changes? You could imagine:
398 autoproject --incremental
400 ================================================================
402 Stuff NOT to do, and why:
404 consider auto-including any file that matches "*.in".
405 [ no: po/Makefile.in shouldn't be included ]
407 must look at mkid to see how it works (for subdir usage)
408 [ right now, it doesn't. i don't see a simple fix right now ]
410 if configure.in not found, move up a directory and try again? This
411 could eliminate a common source of problems.
412 [ this is just a bad idea ]
414 * scripts are installed in $exec_prefix/bin, not $prefix/bin
416 [ the consensus on Gnits is that this isn't required.
417 doubters can work around it anyway ]
419 * make the auto-dep code crash if GNU make not in use?
420 (doesn't it already?)