In the past, there had been some debate and confusion about
whether "aclocal --install" should copy third-party .m4 files
in the first directory passed to the `-I' option even when
such directory was given as an absolute path, or whether it
was better to do so only for directories specified with a
relative path.
The rationale for this latter behaviour was that, before the
existence of the `ACLOCAL_PATH' variable, the only way (a poor
way, I might add) for a common user to extend the search path
of a system-wide installation of aclocal was to export something
like ACLOCAL="aclocal -I /my/extra/macros" in the environment.
Today, the correct way to proceed is undoubtedly through the
use of ACLOCAL_PATH, so we can settle the question once and for
all, and start verifying the correct behaviour of `-I' with a
new test.
* tests/aclocal-install-absdir.test: New test.
* tests/Makefile.am (TESTS): Add it.
* doc/automake.texi (aclocal Options): Be more explicit about
this part of `--install' semantics.
docs: fix node names for automake and aclocal invocations
With the older node names, an "info automake" command issued
from the command line would have opened the node about the
invocation of the automake program, rather than the Top node
of the automake documentation. To invoke the Top node, one
had to issue the command "info Automake" instead (note the
different capitalization). This was suboptimal, and certainly
confusing.
With this change, "info automake" will open the Top node of the
automake documentation; to access the nodes about the invocation
of the automake and aclocal program, one has now to issue "info
automake-invocation" and "info aclocal-invocation" respectively.
gitignore: use only one .gitignore file, in the top-level directory
* doc/.gitignore, doc/amhello/.gitignore, lib/Automake/.gitignore,
lib/Automake/tests/.gitignore, tests/.gitignore: Deleted, their
contents merged into ...
* .gitignore: ... this top-level file. Since we are at it, improve
it by anchoring files where it makes sense.
tests: one more fixlet for gettext macros requirement
* tests/gettext-macros.test (configure.in): Add calls to
AC_INIT and AC_PROG_CC, to avoid a spurious failure in the
later aclocal invocation.
Luckily, this wart wasn't causing any spurious SKIP or FAIL,
but only the redundant addition to the `-Wno-syntax' option
to some aclocal invocations in the gettext tests.
The code introduced in the earlier change `v1.11-581-gb7d67d5'
and aimed at automatically fetching all the .m4 files provided by
gettext has proven inadequate, since it hasn't managed to truly
and always get *all* the required .m4 files. For example, it has
failed to fetch the file `intldir.m4' (present in gettext 0.18.1,
and where the macro `AM_GNU_GETTEXT_INTL_SUBDIR' is defined),
which in turn has caused spurious SKIPs of at least the test
`gettext3.test'.
The reason for this debacle is that autopoint looks at the version
specified in AM_GNU_GETTEXT_VERSION to decide which version of the
gettext infrastructure to bring in; since we were unconditionally
specifying the older 0.10.35 version, the newer `.m4' files weren't
brought in.
* tests/gettext-macros.test: In AM_GNU_GETTEXT_VERSION, instead
of unconditionally specifying the older 0.10.35 version, specify
the version of the available `gettextize' or `autopoint' program,
dynamically extracted with the help of ...
(extract_program_version):... this new function, and saved ...
($autopoint_version): ... in this new variable.
tests: fix failure due to debugging code forgotten into a test
* tests/missing-tar.test: Don't ever call the `missing' script
with `sh -x'; this was used for debugging, but an instance of
it slipped into the committed test case. Bug revealed by a
failure on a Solaris 10 system with GNU tar installed as `gtar'.
tests: avoid spurious failure of libtool and gettext tests
On Solaris 10 (and presumably earlier), /bin/sh trips up on
here-documents that contains a command substitution *and* are
fed to a shell function:
# All as expected.
$ cat <<END
`pwd`
END
/home/stefano
$ echo status = $?
status = 0
# An apparently innocuous function ...
$ kitty () { cat; }
# ... but hilarity ensues!
$ kitty <<END
`pwd`
END
/tmp/sh137723: cannot open
$ echo status = $?
status = 1
We need to work around this misbehaviour in a couple of our
tests (whose failures where causing cascading failures in a
lot of other tests).
* tests/gettext-macros.test: Avoid the use of command substitution
in a here-document passed to the `indent' function, by using the
`echo' builtin instead.
* tests/libtool-macros.test: Likewise.
See also:
<http://lists.gnu.org/archive/html/bug-autoconf/2011-12/msg00001.html>
hacking: distribute it, and mention it in the ChangeLog
Not distributing the HACKING file might make it more difficult,
for some random curious user, to get informed about or interested
in the Automake development process, or to send us patches.
* Makefile.am (EXTRA_DIST): Add HACKING.
* HACKING: It's OK to distribute this file, and to mention it in
the ChangeLog.
Today, practically nobody uses the GNU rx library, which, according
to its own website <http://www.gnu.org/software/rx/rx.html>, has
been "decommissioned". Consequently, the automake-provided macro
AM_WITH_REGEX is not used nor required anymore. Deprecate it, so
that it will be possible to safely remove it in the next major
automake version.
* m4/regex.m4 (AM_WITH_REGEX): Give a warning of the class
`obsolete' when this macro is used.
* doc/automake.texi (Public Macros): Move description of
`AM_WITH_REGEX' from here ...
(Obsolete Macros): ... to here, and declare it as obsolete
and "to be removed in a future version".
* tests/regex-obsolete.test: New test.
* tests/Makefile.am (TESTS): Add it.
* NEWS: Update.
See also:
<http://lists.gnu.org/archive/html/automake-patches/2010-11/msg00131.html>
Many GNU packages are moving towards xz-compressed tarballs, so
let's follow suit, by dropping the creation and distribution of
a bzip2-compressed tarball and switching to xz instead.
For compatibility and safeness, we will continue to create and
distribute a gzip-compressed tarball as well.
* configure.ac (AM_INIT_AUTOMAKE): Drop `dist-bzip2', add
`dist-xz'.
* NEWS: Update.
* configure.ac: Now that Automake requires autoconf 2.68 for its
own bootstrapping and build system, we can assume that PACKAGE_URL
gets automatically AC_SUBT'd.
* automake.in (handle_configure): When processing `configure.am',
also expand `HAVE-MAKEFILE-IN-DEPS' to a boolean telling whether
`MAKEFILE-IN-DEPS' is empty or not.
* lib/am/configure.am [?HAVE-MAKEFILE-IN-DEPS?]
(%MAKEFILE-IN-DEPS%): New target without dependencies, to
avoid the "deleted .am file" problem. Emit this only when
`?HAVE-MAKEFILE-IN-DEPS?' is true, to avoid generating an
"empty" dependency declaration.
* tests/deleted-am.test: Make grepping of error message stricter.
* tests/dist-missing-am.test: Likewise.
* tests/remake-deleted-am.test: New test.
* tests/remake-deleted-am-2.test: Likewise.
* tests/remake-deleted-am-subdir.test: Likewise.
* tests/remake-renamed-am.test: Likewise.
* tests/makefile-deps.test: Likewise.
* tests/Makefile.am (TESTS): Add the new tests.
* NEWS: Update.
Fixes automake bug#9768.
Report by Peter Johansson.
See also commit `Release-1-10-40-gd0ebf71', which fixed a similar
problem for .m4 files included by configure.ac.
tests: better handling of gettext and libtool requirements
This change fixes automake bug#9807.
Before this change, the automake testsuite only looked for the
`.m4' files containing libtool and gettext macros definitions in
the directory `${prefix}/share/aclocal' (and in the directories
specified by the `dirlist' file in there, if any), where ${prefix}
was the configure-time automake installation prefix (defaulting
to `/usr/local').
This approach had various shortcomings and disadvantages. Let's
briefly describe the three major ones.
First, on most GNU/Linux systems, a libtool or gettext installed
from distro-provided packages (e.g., by dpkg on Debian/Ubuntu, or
by rmp on RedHat/Fedora) would have `/usr', not `/usr/local', as
its ${prefix}; so, trying to run the automake testsuite with a
simple "./configure && make && make check" would have failed to
execute the libtool and gettext tests on most GNU/Linux distros.
It's true that it was quite easy to work around this issue, by
creating a proper `/usr/local/share/aclocal/dirlist' file with
an entry pointing to `/usr/share/aclocal' (a workaround in fact
used by most automake developers); but the typical user wasn't
aware of the necessity of this trick, so the libtool and gettext
tests was usually skipped on testsuite runs "in the wild", thus
needlessly reducing coverage.
Second, the older testsuite behaviour made more difficult for
the developers to run the testsuite with non-default libtool or
gettext. For example, assume the developer is working on a system
that has a default libtool version 1.5 installed in the /usr/local
hierarchy; to improve coverage, the developer installs also a more
modern libtool version, say 2.4, in its home directory, let's say
in ~/libtool-2.4; he then tries to run the automake testsuite with
this more modern libtool by doing an (apparently) simple:
$ PATH=$HOME/libtool-2.4:$PATH make check
But the automake testsuite would still look for libtool macros in
/usr/local/share/aclocal, not in ~/libtool-2.4/share/aclocal, so
the wrong version of the macros would be picked up, and the tests
would either fail spuriously or (which would be worse) pass without
truly covering the libtool version the developers was thinking to
be testing with.
Worse again, the automake testsuite would *unconditionally* look
for libtool macros in /usr/local/share/aclocal, so even something
like:
$ export ACLOCAL_PATH=$HOME/libtool-2.4/share/aclocal
$ PATH=$HOME/libtool-2.4:$PATH make check
wouldn't work.
Third and last, during a "make distcheck", automake is configured
with a ${prefix} pointing to a proper subdirectory of the build
directory (usually `pwd`/_inst), which gets created on-the-fly;
in this case, with the old approach, the automake testsuite never
found the libtool and gettext macro files, ans so the libtool and
gettext tests was *always* skipped in a "make distcheck".
* tests/libtool-macros.test: New helper test, looking (with the
help of the `libtoolize' script) for libtool macro files required
by most libtool tests, and making them easily accessible.
* tests/gettext-macros.test: New helper test, looking (with the
help of the `libtoolize' script) for libtool macro files required
by most libtool tests, and making them easily accessible.
* tests/defs.in: Update to make it rely on the results and setups
of `libtool-macros.test' and `gettext-macros.test'.
* tests/Makefile.am: Declare dependency of all the logs of libtool
tests from `libtool-macros.log', and all the logs of gettext tests
from `gettext-macros.log'.
(TESTS): Add the new tests.
tests: report useful system information in 'test-suite.log'
It has already happened various times that a user has run the
automake testsuite, experienced a failure, read the messages
telling him "See tests/test-suite.log" and "Please report to
bug-automake@gnu.org", and done exactly that -- sending us only
the contents of `tests/test-suite.log', which are usually not
enough to start debugging the reported failure. So we have to
ask him for more details, and usually also for the `config.log'
file generated by configure. It's time to fix this recurring
feedback inefficiency. We do so by creating a dummy test case
that takes care of copying the contents of `config.log', plus
other useful system information, in the final `test-suite.log'.
* tests/get-sysconf.test: New test, gathering system information
and then always terminating with a SKIP, so that its output gets
copied in `test-suite.log'.
* tests/Makefile.am (TESTS): Add it.
Reuben Thomas [Mon, 5 Dec 2011 22:40:48 +0000 (23:40 +0100)]
python: remove relics for Python 1.5 support
* m4/python.m4: The comments in here claim to support only
Python >= 2.0, yet this file still has specific support for
Python 1.5. Just remove it, python 1.5 is 12 years old now,
and practically defunct.
* NEWS: Update.
configure: remove extraneous 'eval's from AM_RUN_LOG invocations
* configure.ac: Remove extra 'eval's from AM_RUN_LOG invocations;
for example, instead of "AM_RUN_LOG([eval $PERL --version])",
simply use "AM_RUN_LOG([$PERL --version])"
maint: snapshots from `maint' are still development snapshots
The maintenance-oriented development line in the `maint' branch,
while being usually pretty stable and 99% backward-compatible,
is not always right off production-quality; but until now, the
Automake package version declared in configure.ac hid this fact,
since it appeared to be the version of a stable release (e.g.,
11.1). Fix this.
* configure.ac (AC_INIT): Bump version to "1.11.0a"
* configure.ac: If the current release is detected to be a test
release or a development snapshot, print a proper warning for
the user.
* README-alpha: Delete, it's obsolete now (and in fact this file
hasn't been touched in eleven years, since release 1.4b or so).
* HACKING (Release procedure): Don't say to update README-alpha.
devel: help in comparing Makefile.in from different commits
Now that the generated Makefile.in, configure and aclocal.m4 files
are no longer committed in Automake's git repository, a simple
"git diff" or "git log" no longer shows if and how a change in
Automake results in changes to the Makefile.in files and/or
configure script of its own build system. Still, the ability to
peek so easily at such differences has proved itself quite useful
in the past, often revealing inconsistencies and blunders, and
sometimes even bugs; so it would be a pity to lose that altogether.
With this change, we add a new maintainer recipe that re-introduces
much of that capability, by generating and comparing on the fly the
Makefile.in, configure and aclocal.m4 derived from two arbitrary
commits of the Automake repository.
* Makefile.am (autodiffs, compare-autodiffs): New phony targets.
repo: don't commit generated files in the git repository anymore
It has been quite some time since autoconf and libtool have stopped
committing the generated autotools files in their git repositories,
with no significant ill effects we're aware of. It's true that the
autoconf bootstrap process has now the minor annoyance that a
pre-installed autoconf is required to complete it; but luckily
automake will not have a similar annoyance, since our bootstrap
script take care, through some hoops, to use the very automake and
aclocal versions from the current git checkout to generate the
required aclocal.m4 and Makefile.in files. In fact, this has been
a necessity also in the past, because automake has been known to
use in its own build system new development features that hadn't
been present in any previously released automake distribution.
* aclocal.m4: Delete.
* configure: Likewise.
* Makefile.in: Likewise.
* doc/Makefile.in: Likewise.
* lib/Automake/Makefile.in: Likewise.
* lib/Automake/tests/Makefile.in: Likewise.
* lib/Makefile.in: Likewise.
* lib/am/Makefile.in: Likewise.
* m4/Makefile.in: Likewise.
* tests/Makefile.in: Likewise.
* .gitignore: Ignore configure, aclocal.m4, and all the
Makefile.in files.
* configure.ac (AC_PREREQ): New macro call, to require the
latest autoconf (2.68 for the moment).
* missing-fixes:
missing: don't try to re-run tar with a munged command line
missing: inform the user if GNU tar is called
missing: if GNU tar exists but fails when called, give up
missing: miscellaneous fixlets
missing: don't try to re-run tar with a munged command line
* lib/missing: If the default `tar' program fails with the given
arguments, and GNU tar is not available, don't try to re-run the
default `tar' with a munged command line (e.g., ditching possibly
unportable options), as that could be subtly alter the intended
semantics (and maybe even create a somewhat corrupted tarball).
Also, it's worth noting that the main purpose of the `missing'
script is to allow a non-developer to build the package in the
face of slightly-skewed timestamps, not to provide wrappers for
all the maintainer tools -- so we don't have to try too hard
when `missing' is just called to wrap `tar'.
* tests/missing-tar.test: New test.
* tests/Makefile.am (TESTS): Add it.
* lib/missing: If the code trying to run GNU tar is reached, it
means that the previous attempt to run the default tar program
has failed, very likely producing some error message. At this
point, just running GNU tar without further comments might be
confusing.
missing: if GNU tar exists but fails when called, give up
* lib/missing: If the code trying to run GNU tar is reached,
don't continue if the invoked GNU tar program fails, as there
is little point in doing so (and can even be confusing and
counter-productive).
* lib/missing: Some shells, such as Solaris or FreeBSD /bin/sh,
warn about missing programs before performing redirections.
Therefore, where we have to silently check whether a program
exists, perform redirections on a subshell.
Remove redundant uses of double-quotes in variable definitions.
Delete an extra blank line.
Peter Rosin [Tue, 20 Dec 2011 20:42:54 +0000 (21:42 +0100)]
tests: fix spurious failure on systems lacking unistd.h
This is for automake bug#10324.
* tests/silent-lex-generic.test (foo.l): Add a dummy #define of
YY_NO_UNISTD_H, so that the generated foo.c file won't require
unistd.h to be present (it is not present when compiling with,
e.g., MSVC 9).
test defs: hack to support autoconf-wrapper programs
* tests/defs.in ($AUTOCONF): Add a dummy `-B' option to the
autoconf invocation, so that, when the Debian autoconf wrapper
is involved, it will correctly dispatch an autoconf >= 2.50
instead of defaulting to autoconf 2.13.
($AUTOHEADER, $AUTORECONF): Likewise, but for autoheader and
autoreconf respectively.
Reported by Bruno Haible:
<http://lists.gnu.org/archive/html/automake/2011-12/msg00039.html>
tests: fix a minor spurious failure with FreeBSD make
* tests/remake-am-pr10111.test: Avoid using `#' comments in
makefile recipes, as these have been confusing FreeBSD make.
The failure was masked by the fact that this test is currently
expected to fail.
readme: reference webpages for automake mailing lists
* README: Rationalize and reorganize the (brief) description of
the automake mailing lists. In particular, instead of suggesting
the reader to write to the `-request' addresses to subscribe to
mailing lists, point him to the relevant webpages, where he can
also subscribe via a web form.
* README: Now that the automake testsuite uses the parallel-tests
driver, there is no need for the user to capture the stdout of
"make check" to determine which tests have failed: a detailed log
is automatically saved into the `tests/test-suite.log' file.
Since we are at it, improve the wording by dropping an extra
"please".
readme: don't reference the old homepage at sources.redhat.com
* README: Don't reference the old homepage at sources.redhat.com,
which is no longer active; reference the homepage on www.gnu.org
instead. See also automake bug#10157 and bug#10248.
* tests/README: Likewise, and remove related extra-pedantic advice
about copyright papers for test cases (we'll ask for those papers
explicitly when we think they are warranted).
* HACKING (Release Procedure): Place the list of pre-release
bootstrap-and-test commands on a line of its own, so it's easy to
select and then paste it into a terminal window.
Fix the explanation of "make git-release", as, since the previous
change, "make git-release" would simply run "make dist" rather
than "make distcheck".
* Makefile.am (git-dist): The developers should test extensively
before finally creating the release tarball; so don't run "make
distcheck" on their behalf here; instead ...
* HACKING (Release procedure): ... state here that "make check"
and "make distcheck" should be run before calling "make git-dist".
* HACKING (Release procedure): Announces for alpha and beta releases
should be sent to the `platform-testers@gnu.org' mailing list. Some
related rewording.
* HACKING (Administrivia): If a commit fixes a bug registered at GNU
debbugs, its bug number be reported in the ChangeLog entry. Re-order
the entries to give more visibility to the advice on how to verify
that a commit really fixes a bug.
(Working with git): Improve advice about which pre-existing branch
a topic branch should be based on.
Before the present change, automake-generated `dist-xz' rule used
a hard-coded `xz -9'. That was a problem because on this front,
xz differs from gzip and bzip2. While the latter two don't incur
any run-time decompression penalty for using a higher compression
level, specifying -9 with xz imposes a potentially fatal virtual
memory requirement on any client that wants to decompress your
tar.xz file.
People have complained that a tarball compressed with -9 cannot
be uncompressed in a low-memory environment (wrt-based embedded).
Hence, instead of defaulting to -9, which is useful only for very
large tarballs, it defaults to -e (equivalent to -6e). This
limits the default memory requirements imposed on decompressors,
yet still gives very good compression ratios.
* lib/am/distdir.am (dist-xz): Do not hard-code xz's -9: that made
it impossible to override. Actually don't default to -9, either,
since that induced inordinately large virtual memory usage when
merely decompressing. Instead, use its XZ_OPT envvar, defaulting
to -e if not defined. Suggested by Lasse Collin.
(dist, dist-all) [?XZ?]: Likewise
(dist-bzip2): Similarly, do not hard-code -9, but do continue to
use -9 by default. Honor the BZIP2 envvar.
(dist, dist-all) [?BZIP2?]: Likewise
* NEWS: Update.
* doc/automake.texi (The Types of Distributions): Describe the
newly enabled environment variables.
This is inspired to commit v1.11-389-g6da46f3, with additional
changes to reflect that the xz compression level should default
to -e, not -9.
tests: fix spurious failures due to missing 'yywrap()' function
The AC_PROG_LEX Autoconf macro does not diagnose a failure to find
the "lex library" expected to provide a `yywrap' function (function
which is required to link most lex-generated programs). On the
contrary, when all the link attempts (i.e., with `-ll' and `-lfl')
fail, configure declares that no lex library is needed, and simply
proceeds with the configuration process -- only for the build to
possibly fail later, at make time.
This behaviour is intended; the Autoconf manual reads:
``You are encouraged to use Flex in your sources, since it is
both more pleasant to use than plain Lex and the C source it
produces is portable. In order to ensure portability, however,
you must either provide a function `yywrap' or, if you don't use
it (e.g., your scanner has no `#include'-like feature), simply
include a `%noyywrap' statement in the scanner's source.''
This AC_PROG_LEX behaviour is causing some spurious failures of
the Automake testsuite in environments which lack a proper library
providing `yywrap' (this happens for example on Fedora-based
systems). The proper workaround is to simply provide a fall-back
implementation of `yywrap' in our lexers.
Peter Rosin [Sat, 26 Nov 2011 13:54:00 +0000 (14:54 +0100)]
tests: fix 'distcheck-override-infodir.test' on Cygwin
* tests/distcheck-override-infodir.test (Makefile.am): Do not add
any `/' between $(DESTDIR) and the following paths. Otherwise,
when $(DESTDIR) is empty, the recipes will try to access files
with a leading double slash, which have an implementation-defined
interpretation (e.g., for Cygwin, they mean UNC paths).
coverage: undistributed '.am' and '.m4' files are diagnosed
The stub rules emitted to work around the "deleted header problem"
for `.m4' files (included by autoconf in e.g., configure.ac) and
for `.am' files (included by automake in e.g., Makefile.am) should
not prevent "make" from correctly complaining when such a required
file is missing from a distribution tarball.
* tests/dist-missing-am.test: New test.
* tests/dist-missing-m4.test: Likewise.
* tests/dist-missing-included-m4.test: Likewise.
* tests/Makefile.am (TESTS): Add them.
coverage: required but missing '.am' and '.m4' files are diagnosed
The stub rules emitted to work around the "deleted header problem"
for `.m4' files (included by autoconf in e.g., configure.ac) and
for `.am' files (included by automake in e.g., Makefile.am) should
not prevent the remake rules from correctly erroring out when a
still-required file is missing.
texinfo: work around Solaris 10 xpg4 shell bug in install rules
* lib/am/texinfos.am (install-html-am): Use an extra variable
indirection to work around a bug in Solaris 10 /usr/xpg4/bin/sh.
Bug revealed by a failure of `txinfo21.test'. See also:
<http://lists.gnu.org/archive/html/bug-autoconf/2011-11/msg00005.html>
<http://debbugs.gnu.org/cgi/bugreport.cgi?bug=10026#23>
tests: fix spurious failure with older install-info
* tests/install-info-dir.test (foo.texi): Also add proper
`@settitle', `@dircategory' and `@direntry' directive, otherwise
older versions of `install-info' (e.g., 4.8) will fail to create
the `dir' index file in ${infodir}.
tests: fix spurious error in 'uninstall-fail.test' on Solaris
* tests/uninstall-fail.test: Solaris 10 /usr/xpg4/bin/sh can add
a line number before the `:' in the error messages issued by shell
builtins. Account for that in our grepping of make output.
tests: avoid a spurious failure of 'ltinit.test' MinGW
* tests/ltinit.test: Be laxer in grepping configure output, to
avoid spurious failures on systems which lack POSIX dynamic
linking (e.g., MinGW), or when cross-compiling for such systems.
tests: testsuite is now safe to run with dmake in parallel mode
* tests/defs.in: Unset variables DMAKE_CHILD, DMAKE_DEF_PRINTED
and DMAKE_MAX_JOBS, which are exported by Solaris dmake when run
in parallel mode, and which might confuse make processes spawned
by our testsuite.
tests: fix spurious failures w.r.t. parallel make and colorization
* tests/color2.test: Skip the test if the $MAKE program fails to
consider the standard output as a tty when spawned by `expect'.
This is required for make implementations, like FreeBSD make and
Solaris dmake, that redirect the output of recipes to temporary
files or pipes when run in parallel mode. Since we are at it,
simplify the detection of a working `expect' program, and throw
in other minor simplifications.
tests: avoid another failure of 'uninstall-fail.test' on Solaris
* tests/uninstall-fail.test: On Solaris 10, if `/bin/rm' is run
with the `-f' option, it doesn't print any error message when
failing to remove a file (due to e.g., "Permission denied").
Yikes. Cater to this incompatibility, by relaxing the test when
a faulty `rm' is detected.
* m4/maintainer.m4 (AM_MAINTAINER_MODE): Fix the "checking ..."
configure message related to the enabling/disabling of maintainer
mode.
* tests/help-maintainer.test: Renamed ...
* tests/maintmode-configure-msg.test: ... to this, and extended
to cover the fixed bug. Also, since we are at it, make grepping
of configure help screens stricter, and throw in some few other
changes to reduce code duplication and enhance readability.
* tests/Makefile.am (TESTS): Update.
* THANKS: Update with Zack's new e-mail address.
Report and initial patch by Zack Weinberg, test cases added by
Stefano Lattarini.
info: allow user to inhibit creation/update of '${infodir}/dir'
With this change, we allow the user to request the install-info
rules not to update the `${infodir}/dir' file, by exporting the
environment variable `AM_UPDATE_INFO_DIR' to the value "no".
This is especially useful to distro packagers, and is a definite
improvement over our previous hack of looking whether the
`install-info' program was the Debian or GNU version -- hack
which had been silently broken with recent versions of debian
install-info BTW (probably since dpkg 1.15.4, 2009-09-06).
This change fixes automake bug#9773. See also Debian Bug#543992.
* lib/am/texinfos.am: Don't look anymore at the output of
`install-info --version' to decide whether to use it to update
the `${infodir}/dir' or not; instead, honour the environment
variable `AM_UPDATE_INFO_DIR'.
* tests/install-info-dir.test: New test.
* tests/Makefile.am (TESTS): Add it.
* tests/defs: Also unset `AM_UPDATE_INFO_DIR', to avoid unwanted
interferences from the environment.
* doc/automake.texi (Texinfo): Update.
* NEWS: Likewise.
* THANKS: Likewise.
Report by Jonathan Nieder.
dejagnu: allow the package developer to extend site.exp
Fixes automake bug#7873.
* lib/am/dejagnu.am (site.exp): Depend from the files listed in
$(EXTRA_DEJAGNU_SITE_CONFIG), if any. Append their contents to
the generated site.exp (still preserving user edits).
* doc/automake.texi (Dejagnu Tests): Update.
* tests/dejagnu-siteexp-append.test: New test.
* tests/dejagnu-siteexp-extend.test: Likewise.
* tests/dejagnu-siteexp-useredit.test: Likewise.ù
* NEWS: Update.
* tests/Makefile.am (TESTS): Update.
Merge branch 'dejagnu-srcdir-discrepancy' into maint
* dejagnu-srcdir-discrepancy:
dejagnu: ensure 'srcdir' is defined as a relative directory
Fix testsuite failure of check12.test without DejaGNU.
Extend and improve tests on DejaGnu support.
This change partly reverts commit "Create subdirs for generated
sources even when not dep tracking", of 2011-04-02.
That commit had caused the bugs #8485 and #8526. Since we are
nearing the bug-fixing automake release 1.11.2, the safest policy
at the moment is to just revert the problematic hunks: an older,
known bug is better than a regression.
* automake.in (handle_single_transform): Don't add a dirstamp
dependency, even when $object is derived and lands in a subdir.
* tests/Makefile.am (XFAIL_TESTS): Add lex-subobj-nodep.test,
remove yacc-dist-nobuild-subdir.test.
tests: fix spurious failures with "chatty" make implementations
* tests/distcheck-missing-m4.test: On failure, some make
implementations (such as Solaris make) print the whole failed
recipe on standard output. This was causing a spurious failure
in the checks grepping the output from make. Work around this.
* tests/distcheck-outdated-m4.test: Likewise.
* tests/aclocal-path-precedence.test: Also remove the `configure'
script between different test runs, to ensure it is always remade
by autoconf. Add proper explicative comments.
tests: avoid spurious failure of 'uninstall-fail.test' on Solaris
* tests/uninstall-fail.test: All the Solaris 10 shells (/bin/sh,
/bin/ksh, and /usr/xpg4/bin/sh), upon failing to chdir to a
directory with the `cd' builtin, print a message like:
"sh: /root: permission denied"
which doesn't report the `cd' builtin anywhere. Relax the grepping
of the error message accordingly.
tests: avoid spurious failure of 'uninstall-fail.test' on Cygwin
* tests/uninstall-fail.test: Be sure to really skip this test
on systems that allows files to be removed from unwritable
directories. Motivated by a spurious failure on Cygwin 1.5.
dejagnu: ensure 'srcdir' is defined as a relative directory
This change fixes automake bug#7833.
* lib/am/dejagnu.am (check-DEJAGNU): Prefer using plain $(srcdir)
over calculating and using the absolute path of $(srcdir).
* tests/dejagnu-relative-srcdir.test: New test.
* tests/dejagnu-absolute-builddir.test: Likewise.
* tests/Makefile.am (TESTS): Update.
Report by Ian Lance Taylor. Suggestions by Ralf Wildenhues.
tests: avoid spurious failure in 'parallel-tests3.test'
This fixes automake bug#8788.
* tests/parallel-tests3.test: To ensure that the serial run of
the dummy testsuite is still ongoing when the parallel run has
terminated, use `kill -0', not a bare `kill'. This will prevent
a testsuite crash on NetBSD 5.1, and a testsuite hang on FreeBSD
8.2. Also, since we are at it, try harder to avoid possible
hangs of the script in other unusual situations.
tests: fix spurious failure with FreeBSD make and Yacc in VPATH
* tests/subpkg.test: Some cosmetic adjustments. Move the tests
checking that $(YLWRAP) is defined and installed properly when
ylwrap is in a default auxdir found in a parent package ...
* tests/subpkg-yacc.test: ... into this new test, which carefully
avoids to trigger the known bug#7884 (combo FreeBSD make plus Yacc
plus VPATH build).
* tests/Makefile.am (TESTS): Update.
* tests/aclocal-path-precedence.test: Rewrite configure.in,
rather than appending to it, to avoid spurious failures (at
least with autoconf 2.62) due to repeated calls to AC_INIT.
Also, add package name and version arguments to AC_INIT, to
avoid spurious errors from automake.
docs: avoid using colon character inside arguments of @pxref
Fixes automake bug#9753
* doc/automake.texi (VPATH Builds): Avoid using colon character
`:' inside arguments of @pxref, as this can cause problems in
the generated `.info' files, and such an usage will be explicitly
forbidden by future texinfo documentation.
* THANKS: Update.
projects / thirdparty / automake.git / log
