check: Make 'check-TESTS' target depend on test dependencies
The 'check' and 'recheck' targets are already depending on the progams
and scripts required to run the test suite. The 'check-TESTS' target
while not in the public API might be used by uninformed users, so it
makes sense to add the same dependency to it. This fixes bug#27186.
Thomas Martitz [Mon, 13 Mar 2017 11:41:59 +0000 (12:41 +0100)]
automake: Shorter object file names under subdir-objects
Combining the 'subdir-objects' option with target-specific flags had
the consequence of producing long object file names. This was done to
preventively ensure the uniqueness of object file names. We are now
using shorter names by default, and handle long names when an actual
conflict is detected. This will hopefully reduce the necessity of
using the 'prog_SHORTNAME' facility.
This is workaround to a regression introduced by 48107579abadbe857c8299c38d2ca0a3f8f354c8. With current
"lib/texinfo.tex", 'texi2dvi' is not able to produce a valid output
for a Texinfo document containing a @synindex command. As a
consequence the test "t/txinfo-no-clutter.sh" was failing.
* t/txinfo-no-clutter.sh: Remove use of @synindex command.
Mathieu Lirzin [Sun, 5 Mar 2017 09:57:29 +0000 (10:57 +0100)]
maint: Update "lib/texinfo.tex".
This fixes a regression in 48107579abadbe857c8299c38d2ca0a3f8f354c8
where 'texi2dvi' was not able to produce a valid output for a Texinfo
document without text. As a consequence the test "t/silent-texi.sh"
was failing.
Mathieu Lirzin [Thu, 2 Mar 2017 16:50:02 +0000 (17:50 +0100)]
maint: Update fetch URLs.
* maintainer/maint.mk (SV_CVS, SV_GIT_AC): Remove.
(SV_GIT_CF, SV_GIT_GL): Use https for Savannah Git repositories.
(FETCHFILES): Use Gnulib repository as the source of "texinfo.tex",
"gendocs.sh", and "gendocs_template" files.
Jim Meyering [Fri, 9 Sep 2016 01:58:43 +0000 (18:58 -0700)]
automake: do not emit rule with two dependents
On a system using our replacement alloca, make would emit
this warning:
target '.deps/alloca.Po' given more than once in the same rule
That arose because automake would emit a rule depending on both
./.deps/alloca.Po and .deps/alloca.Po. Normally, duplicate
dependents are avoided by virtue of their names being keys in
the %dep_files hash, but in this case, that particular file
was specified in two different ways.
* bin/automake.in (handle_ALLOCA): When the $dir prefix is empty,
make it './', to ensure that when we add ./.deps/alloca.Po
it is deduped.
See: http://bugs/gnu/org/22702
Paul Eggert [Thu, 31 Mar 2016 23:35:29 +0000 (16:35 -0700)]
automake: port to Perl 5.22 and later
Without this change, Perl 5.22 complains "Unescaped left brace in
regex is deprecated" and this is planned to become a hard error in
Perl 5.26. See:
http://search.cpan.org/dist/perl-5.22.0/pod/perldelta.pod#A_literal_%22{%22_should_now_be_escaped_in_a_pattern
* bin/automake.in (substitute_ac_subst_variables): Escape left brace.
Paul Eggert [Tue, 29 Mar 2016 02:44:19 +0000 (19:44 -0700)]
automake: port better to future gzip
* lib/am/distdir.am (dist-gzip, dist-shar, distcheck):
Port better to future versions of gzip, which are planned to
deprecate the GZIP environment variable (Bug#20132).
* lib/compile: Have icl be treated similarly to cl
(scriptversion): Update.
* t/ax/am-test-lib.sh (require_tool): Handle icl.
* t/compile7.sh: Add new test file for icl...
* t/list-of-tests.mk (handwritten_TESTS): ...and use it.
* NEWS: Update.
* THANKS: Update.
Copyright-paperwork-exempt: yes Signed-off-by: Peter Rosin <peda@lysator.liu.se>
* deps-pr13928:
deps: fix corner-case "make distclean" bug
compile: don't place built object files in $(srcdir), ever ...
tests: fix some bugs in an XFAILing test
deps: 'subdir-object' option now works when foo_SOURCES contains $(var)
Assume we have package satisfying the following conditions:
(1) automatic dependency tracking is enabled;
(2) the 'subdir-objects' Automake option is enabled;
(3) the package uses a recursive make setup.
Also assume that:
(a) a subdir Makefile declares a foo_SOURCES variable containing
a source file in the parent directory;
(b) that parent Makefile declare a compiled program itself.
Then BSD and Solaris make used to fail when running "make distclean",
because the 'distclean' target of the subdir Makefile removed the
whole '.deps' directory before the parent Makefile was done with the
included '.Po' makefile fragments in that directory. This issue was
revealed by failures in the 'subobj-vpath-pr13928.sh' test when those
make implementations were used.
We fix the issue by ensuring the 'distclean' target of any Makefile
only removed the '.Po' makefile fragments included by it, rather than
the whole '.deps' directory where such files resides.
This change should be the last step in fixing automake bug#13928
for good.
* bin/automake.in (handle_languages), lib/am/depend.am: Adjust
to implement the new 'distclean' logic.
* t/pr224.sh: Adjust to avoid a spurious failure.
* PLANS/subdir-objects.txt: Update.
compile: don't place built object files in $(srcdir), ever ...
... even when a source file is specified as '$(srdir)/foo.c' or
'$(top_srcdir)/bar.c'. And ditto for dependency-tracking makefile
fragments (those under '.deps' directories).
Such issues used to occur when the 'subdir-objects' option was given.
This change should fix the second and last part of automake bug#13928.
See also bug#16375 and bug#15293.
* NEWS: Update.
* bin/automake.in (handle_single_transform): Make sure object files
and dependency-tracking makefile fragments coming from source like
'$(srcdir)/foo.c' and '$(top_srcdir)/bar.c' are placed respectively
under $(builddir) and $(top_builddir).
* t/subobj-vpath-pr13928.sh: Enhance to expose even more aspects
of the bug we've just fixed.
* t/subobj-pr13928-more-langs.sh: New test, similar to the one above,
but with non-C languages as well.
* t/list-of-tests.mk (XFAIL_TESTS): Remove 'subobj-vpath-pr13928.sh',
it's now supposed to pass.
(handwritten_TESTS): Add 'subobj-pr13928-more-langs.sh'.
deps: 'subdir-object' option now works when foo_SOURCES contains $(var)
Following a suggestions of Johan Kristensen, we have config.status use
'make' invocations rather than Makefile-parsing 'sed' hacks to bootstrap
the dependency-tracking '.Po' and '.Plo' makefile fragments. To handle
the inclusion of such files that are still missing when make is first
we basically generate a temporary Makefile without these includes, and
call 'make' on that Makefile.
This fixes the serious bug bug#13928, which was an hard blocker to make the
behavior mandated by the 'subdir-object' active by default (which we want
to do in Automake 2.0).
The issue has also been reported in bug#15919.
* NEWS, THANKS: Update.
* bin/automake.in (handle_languages): Add a trailing "marking" comment
("# am--include-marker") to the generated Makefile lines issuing 'include'
directives for the dependency-tracking '.Po' and '.Plo' makefile fragments.
Also rename the generated Makefile variable 'am__depfiles_maybe' to the
clearer 'am__maybe_remake_depfiles'.
Minor unrelated refactoring.
* lib/am/configure.am: Adjust to account for the 'am__depfiles_maybe' ->
'am__maybe_remake_depfiles' renaming.
* lib/am/depend.am: Add rules to generate a dummy version of all the
dependency-tracking '.Po' and '.Plo' makefile fragments.
* m4/depout.m4: Use make invocations rather than Makefile-parsing sed hacks
to bootstrap the dependency-tracking '.Po' and '.Plo' makefile fragments.
We still use some sed trickery in order to remove the inclusion of the
still non existing .Po and .Plo files from the Makefile we invoke make
upon; this is done stripping lines that contain the magic string
"# am--include-marker".
* m4/make.m4 (AM_MAKE_INCLUDE): Given that now automake generates Makefiles
containing include statements with trailing comment, adjust the checks done
here to make sure $MAKE support that; e.g., "include foo.mk # comment"
rather than just "include foo.mk".
Also refactor and adjust to leave better debugging info in config.log.
* t/postproc.sh: Rename ...
* t/depend-postproc.sh: ... to this, and adjust and enhance.
* t/list-of-tests.mk (handwritten_TESTS): Adjust.
(XFAIL_TESTS): Remove 't/subobj-indir-pr13928.sh', which is now succeeding.
* t/subobj-indir-pr13928.sh: Simplify slightly, now that we expect it to
pass.
* t/depcomp8a.sh: Adjust grepping check to account for the changes in
the generated Makefile, and tp be somewhat more robust in light of possible
future modifications.
* t/depcomp8b.sh: Likewise.
* t/subobj11b.sh: Likewise.
* t/subobj11c.sh: Likewise.
* t/extra-sources.sh: Likewise.
* t/lex-depend-grep.sh: Likewise.
* t/lex-depend-cxx.sh: Add a command to help debugging in case of test
failure.
Helped-by: Johan Kristensen <johankristensen@gmail.com> Signed-off-by: Stefano Lattarini <stefano.lattarini@gmail.com>
* lib/config.sub: This.
* lib/INSTALL, lib/config.guess, lib/gitlog-to-changelog, lib/gnupload,
lib/update-copyright: And this (but only for copyright year update, no
real semantic change)
Merge branch 'minor' into micro after 1.15 release
And bump version number: 1.15 -> 1.15.0a
* minor:
release: stable minor release 1.15
NEWS: minor improvements and fixed some typos and grammaros
docs: "make distcheck" implementation details are not to be abused
NEWS: improve and adjust in light of the oncoming 1.15 release
Fix dumb logic error preventing $install_sh from being be overridden
Expose automake bug#19311
build: fix race in parallel builds
init: ensure $ac_aux_dir is defined before being used
plans: enabling subdir-object by default is blocked on bug#13928
maint: update copyright years
maint: sync files from upstream ("make fetch")
Typofixes in warning messages and manual
NEWS: a typofix, and better word wrapping
parallel-tests: avoid possible implicit "make all" in test-suite.log rule
Allow user to extend .PRECIOUS target
cosmetics: remove a couple of extra trailing white spaces
tests: fix a spurious failure on Mac OS X
docs: make clear the JAVA primary is frozen
install-sh: a slightly better diagnostic, and tests enhancements
install-sh: be stricter in catching invalid usages
tests: more significant names for some tests
tests: some cosmetic fixes
tests: more significant names for a test
docs: drop a few obsolescent FIXME/TODO comments, and associated text
testsuite harness: report test exit status in log file
TAP driver: no need to invoke AC_PROG_AWK directly
TAP driver: remove perl implementation (move it into contrib/)
NEWS: stop reporting "new" Automake versioning scheme
cosmetics: untabify the install-sh script
install-sh: assume that "set -f" and "set +f" work...
install-sh: assume ${var:-value} works as expected
install-sh: assume 'dirname' is available and working correctly
post-release: micro version bump (1.14a)
docs: "make distcheck" implementation details are not to be abused
* doc/automake.texi: State explicitly and in detail that the exact location
and the exact structure of the subdirectory used by "make distcheck" is to
be considered an implementation detail, which can change at any time.
shell-no-trail-bslash: improve diagnostic in case of failure
* t/ax/shell-no-trail-bslash.in: Here, by fixing a typo in a
variable name and a logic error.
* t/self-check-shell-no-trail-bslash.sh: Enhance to catch the
issue.
tests: make script 'shell-no-trail-bslash' simpler and more robust
This solves spurious failure in the 'check-no-trailing-backslash-in-recipes'
target for Automake-NG.
This is basically a backport of Automake-NG commit v1.14.1-1010-g85aae58;
the point is to minimize the amount of spurious diffs between the mainline
Automake and the Automake-NG source trees.
* t/ax/shell-no-trail-bslash.in: Simplify and fortify.
* t/self-check-shell-no-trail-bslash.sh: Enhance.
tests: fix spurious failure in test on TEXINFO_TEX overriding
* t/txinfo-override-texinfo-tex.sh: Here. The issue was pre-existing, but
has been only recently exposed by the fix for automake bug#18286 "distcheck
fails to detect missing files" (see commit v1.14.1-4-g01a7a4a) and by the
BSD make semantics. To convince yourself this change actually makes sense
semantically, see https://sourceware.org/ml/binutils/2012-06/msg00004.html
* t/lex-noyywrap.sh: Here.
* t/instmany-mans.sh: And here.
* t/instmany-python.sh: And here.
* t/instmany.sh: And here.
* t/parallel-tests-concurrency.sh: And here.
dist: fix bug#18286 "distcheck fails to detect missing files"
BTW, this issue had been already reported in the past:
http://lists.gnu.org/archive/html/automake/2006-09/msg00008.html
http://lists.gnu.org/archive/html/automake/2013-01/msg00049.html
"make distcheck" could sometimes fail to detect missing files in the
distribution tarball, especially in those cases where both the generated
files and their dependencies are explicitly in $(srcdir). An important
example of this are *generated* makefile fragments included at Automake
time in Makefile.am. A basic example:
# -*- Makefile.am -*-
$(srcdir)/fragment.am: $(srcdir)/data.txt $(srcdir)/preproc.sh
cd $(srcdir) && $(SHELL) preproc.sh <data.txt >fragment.am
include $(srcdir)/fragment.am
...
If the use forgot to add data.txt and/or preproc.sh in the distribution
tarball, "make distcheck" would have erroneously succeeded!
The reason is that, while $(srcdir)/data.txt does not exist, make also
looks in $(srcdir)/$(srcdir)/data.txt, and in the distcheck-issued
VPATH build where $(srcdir) is '..', that file exists, as it is
part of the original development directory.
* t/distdir.am (distcheck): Adjust to have the build directory be
'$(distdir)/_build/sub' rather than just '$(distdir)/_build'. Thanks
Nicola Fontana for the suggestion.
* t/distcheck-pr18286.sh: Enhance and tighten a little.
* t/list-of-tests.mk (XFAIL_TESTS): Remove 't/distcheck-pr18286.sh',
as it's now passing.
* t/subdir-am-cond.sh: Adjust to avoid a fully spurious failure due
to the new distcheck semantics.
* t/subdir-ac-subst.sh: Likewise.
* t/dejagnu-relative-srcdir.sh: Likewise.
* t/txinfo-builddir.sh: Likewise.
* NEWS: Update.
Helped-by: Nicola Fontana <ntd@entidi.it> Helped-by: Peter Johansson <trojkan@gmail.com> Signed-off-by: Stefano Lattarini <stefano.lattarini@gmail.com>
include: fix bug in handling of user-defined makefile fragments generation
If a user defined one single Makefile fragment to be included (via Automake
includes) in his main Makefile.am, and gave a rule to generate that file
from other data, Automake used to spuriously complain about with something
like "overrides Automake target '$(srcdir)/foo.am". This change remove that
spurious error (via a simple hack rather than a systematic change, but oh
well).
* lib/am/configure.am (%MAKEFILE-IN-DEPS%) [?HAVE-MAKEFILE-IN-DEPS?]: Add
a trailing "$(am__empty)" to the list of targets, which is enough to trick
Automake into not complaining about "duplicated targets" in case the
'%MAKEFILE-IN-DEPS%' list expands to a single target that is also declared
in some user-defined rule.
* t/list-of-tests.mk (XFAIL_TESTS): Remove now-passing test
't/am-include-only-one-generated-fragment.sh'.
* NEWS: Update.
tests: expose bug in handling of user-defined makefile fragments generation
If a user defines one single Makefile fragment to be included (via Automake
includes) in his main Makefile.am, and givse a rule to generate that file
from other data, Automake will spuriously complain about with something
like "overrides Automake target '$(srcdir)/foo.am".
* t/am-include-only-one-generated-fragment.sh: Expose the bug (this test
is still XFAILing).
* t/list-of-tests.mk: Add the new test.
* micro:
cleanup: refactor code to initialize DIST_COMMON
dist: ordering of files in DIST_COMMON is deterministic now
tests: refactor some tests on DIST_COMMON
maint: make output of 'gen-testsuite-part' deterministic
When computing lispdir, don't load emacs site wide init file.
PATH: quote $(PATH_SEPARATOR) as well
Improve detection of GNU make, avoiding "Arg list too long" errors.
There is not need to make that an Automake variable early,
only to later get and munge its contents, and use the new
content to redefine the variable.
* bin/automake.in (@dist_common): New global variable.
(push_dist_common, handle_dist): Use it.
(handle_dist): Define am__DIST_COMMON instead of DIST_COMMON
directly.
(initialize_per_input): Reset it to empty.
($configure_dist_common): Turn this scalar variable ...
(@configure_dist_common): ... into this array variable.
(handle_dist): Adjust.
(required_file_check_or_copy): Update and wrap some comments.
* lib/am/distdir.am (DIST_COMMON): Append $(am__DIST_COMMON).
* t/distcom2.sh: Tighten a little.
dist: ordering of files in DIST_COMMON is deterministic now
It had likely stopped being deterministic due to the new perl behavior
of having non-deterministic order of numerating hash keys:
<http://search.cpan.org/dist/perl-5.18.0/pod/perldelta.pod#Hash_randomization>
<http://onionstand.blogspot.ie/2012/12/are-you-relying-on-hash-keys-being.html>
See also similar commit v1.14-19-g52e6404, albeit in this case the issue
is likely coming from autom4te/autoconf, not from automake itself.
Fixes automake bug http://debbugs.gnu.org/17908
* bin/automake.in (handle_dist): Sort @dist_common.
(print_autodist_files): Swap invocations of 'sort' and 'uniq', for
consistency with the new code in 'handle_dist' and to get rid of a
minor hack.
* NEWS: Update.
So that they prefer checking the semantics of the generated Makefiles,
rather than grepping their content. This will be useful in an upcoming
refactoring.
* t/distcom-subdir.sh: Adjust this test.
* t/distcom2.sh: And this.
* t/distcom3.sh: And this.
* t/distcom4.sh: And this.
* t/distcom5.sh: And this.
Andrew Burgess [Wed, 30 Jul 2014 16:41:15 +0000 (17:41 +0100)]
When computing lispdir, don't load emacs site wide init file.
When computing the lispdir emacs was previously invoked with the '-q'
option to avoid loading the users initialisation files, however, the
site wide initialisation file was still loaded, in some cases this can
cause emacs to hang, with the result that a configure can also hang.
The lisp code that aclocal causes to be executed reduces the load-path
list (in emacs) to empty. The load-path is used by emacs to find
packages which it wants to load. Currently, if emacs tries to auto
load a package during shut down, and the package is not found, then
emacs will hang. This does seem like an emacs bug, but protecting
against this in aclocal is simply a case of not loading the site wide
initialisation file.
In this patch then the '-q' option to emacs is replaced with '-Q',
this has the same, the '-Q' option is similar to '-q --no-site-file
--no-splash'.
* doc/automake.texi (Hard-Coded Install Paths): Update explanation of
emacs code used to get lispdir.
* m4/lispdir.m4 (AM_PATH_LISPDIR): Update emacs flags.
Improve detection of GNU make, avoiding "Arg list too long" errors.
Such errors could take place when the main makefile included too many
sub-makefiles, making $(MAKEFILE_LIST) too long and causing the
recipes $(am__is_gnu_make) to exceed the shell's command-line length
limits. This is not a theoretical issue: it could happen for projects
having lots of C/C++ sources and using automatic dependency tracking,
which created an included .Po sub-makefile for each of such sources.
Fixes http://debbugs.gnu.org/18744
* lib/am/header-vars.am (am__is_gnu_make): Fix the logic to avoid
the use of $(MAKEFILE_LIST).
* NEWS: Update.
* micro:
sync: update third-part files from upstream
Make sure AM_INIT_AUTOMAKE has a trailing newline
dist: adjust warning messages about shar and tarZ deprecation
docs: improve description of ${PACKAGE}, ${VERSION}, and similar variables
This used to be the case until Automke 1.13, but we broke it in
Automake 1.14 (see commit v1.13.1-71-gf78b0f0). This caused
issues like http://debbugs.gnu.org/16841
* m4/init.m4 (AM_INIT_AUTOMAKE): Adjust.
* t/aminit-trailing-dnl-comment-pr16841.sh: New test.
* t/list-of-tests.mk: Add it.
* NEWS, THANKS: Update.
docs: improve description of ${PACKAGE}, ${VERSION}, and similar variables
In particular, suggesting that $PACKAGE and $VERSION should be aliases of
the $PACKAGE_TARNAME and $PACKAGE_VERSION definitions coming from AC_INIT,
and not be defined via an obsolete 2-argument invocation of AM_INIT_AUTOMAKE;
and why that is the best default, given all our historical baggage.
See discussion in http://debbugs.gnu.org/16623 for more information and
background.
Paul Eggert [Sat, 23 Aug 2014 14:55:28 +0000 (07:55 -0700)]
build: fix race in parallel builds
Reported by Friedrich Beckmann in: http://bugs.gnu.org/18301
* lib/am/texi-vers.am (?DIRSTAMP?): Put the process-ID into the
temporary file name. Use a similar temporary in the source dir.
Paul Eggert [Sat, 23 Aug 2014 14:55:28 +0000 (07:55 -0700)]
build: fix race in parallel builds
Reported by Friedrich Beckmann in: http://bugs.gnu.org/18301
* lib/am/texi-vers.am (?DIRSTAMP?): Put the process-ID into the
temporary file name. Use a similar temporary in the source dir.
init: ensure $ac_aux_dir is defined before being used
Since we use '$ac_aux_dir' to define '$am_aux_dir', we need
to ensure the former has been initialized before we try to
define the latter, otherwise the definition:
am_aux_dir=`cd $ac_aux_dir && pwd`
will set '$am_aux_dir' to '$HOME', likely causing weird and
unexpected behaviours.
This change fixes automake bug#15981.
* m4/auxdir.m4 (AM_AUX_DIR_EXPAND): AC_REQUIRE expansion
of 'AC_CONFIG_AUX_DIR_DEFAULT'. Fix redundant comment and
AC_PREREQ, add extra quoting around '$ac_aux_dir'.
* t/auxdir-pr15981.sh: New test.
* t/auxdir-cc-pr15981.sh: Likewise.
* t/list-of-tests.mk (handwritten_TESTS): Add them.
* THANKS, NEWS: Update.
* bin/automake.in: Fix a harmless typo in comments, that
I happened to notice while writing this patch.
Paul Eggert [Sat, 4 Jan 2014 04:52:29 +0000 (20:52 -0800)]
doc: fix encoding error with UTF-8 characters
* doc/automake.texi: Specify @documentencoding and
@documentlanguage, to prevent encoding errors for parts of this
input file that are UTF-8. This also causes the .info output to
use curly quotes, which is easier to read though it does assume
UTF-8 support.
parallel-tests: avoid possible implicit "make all" in test-suite.log rule
This change fixes automake bug#16302.
* lib/am/check.am ($(TEST_SUITE_LOG)): Avoid running "make $redo_logs"
when $redo_logs expands to empty, since in that case we are actually
ending up invoking a full "make all". That shouldn't be required, and
can cause slowdowns for people implementing their extra "laziness
wrappers" around check-TESTS (automake bug#16302).
* NEWS: Update.
* bin/automake.in: Adjust to ensure we handle '.PRECIOUS' the same way
we do for '.PHONY' and '.MAKE'.
* lib/Automake/Rule.pm: Likewise.
* t/precious.sh: New test.
* t/list-of-tests.mk: Add it.
* t/phony.sh: Enhance a little while at it.
* NEWS: Update.
* THANKS: Likewise.
Reported-by: Holger Hans Peter Freyther <holger@freyther.de> Signed-off-by: Stefano Lattarini <stefano.lattarini@gmail.com>
install-sh: a slightly better diagnostic, and tests enhancements
* lib/install-sh: When called with no non-option arguments and the '-t'
option with an argument that is not an existing directory, have the
diagnostic output complain about the lack of required arguments rather
than about the bad argument passed to '-t'.
* t/install-sh-unittests.sh: Enhance to also check diagnostic printed
in cases of expected failure.
projects / thirdparty / automake.git / log
