`foo=bar break', but unfortunately they also choke on `foo=bar;
break' and `foo=bar && break'. Writing a sophisticated test to
catch just the intended case seems like more trouble than it's
worth, but leaving the test in causes the testsuite to fail on
valid m4sh output:
* tests/assign.test: Removed; chokes on perfectly valid shell
syntax.
* tests/Makefile.am (COMMON_TESTS): Remove assign.test.
* tests/defs.in (scripts): Don't check the m4sh inputs, go back
to checking the generated ltmain.sh script.
patch; fix them carefully. Introduce an opt_dry_run to ltmain.in
so that the implementations of func_mkdir_p can converge, and a
func_mktempdir to do a better job of temporary directory creation:
* libtoolize.in (func_mkdir_p): Don't fail if the directory wasn't
created in dry run mode.
* tests/defs (func_mkdir_p): Ditto. We don't actually have a dry
run mode for the tests, but the function is written carefully to
be kept in synch and work correctly here too.
* config/ltmain.in (func_mkdir_p): Ditto. This copy of the
function now only differs in its use of $echo over $ECHO.
(func_extract_archive): Removed first redundant mkdir call.
(func_mktempdir): New function that tries to avoid races when
making temporary directories.
(opt_dry_run): Set this if --dry-run is given at the CLI, or if
tests/mdemo-dryrun.test has forced the value of $run.
(func_mode_install): Call $MKDIR directly and error out if the
directory cannot be created.
(func_mode_link): Rather than copying func_mkdir_p into the
wrapper script as a replacement for $MKDIR_P, we know that the
script won't be called my `make -j', so write the current value of
$MKDIR.
below is horrendously broken since it makes the installed libtool
try to run the mkinstalldirs or install-sh helper scripts. This
patch fixes the `make -j' mkdir race condition internally:
* libtoolize.in (func_mkdir_p): New slice and dice algorithm
to build the directory tree one dir at a time, ignoring errors
until the end incase they are transient due to a concurrent
identical mkdir.
* tests/defs (func_mkdir_p): Ditto.
* config/ltmain.in (func_mkdir_p): Ditto, except that ltmain.in
uses $echo rather than $ECHO, and uses $show and $run to interact
correctly with the command line.
* configure.ac (AM_INIT_AUTOMAKE): 1.8 is sufficient again.
* Makefile.am (edit): Don't substitute automake's $(mkdir_p).
* NEWS: Updated.
from Gary V. Vaughan) to compile under MinGW.
* config/mkstamp: Translate input data so that it is assured to
use the Unix line terminations. This is necessary under Windows if
the files are checked out with CR/NL line terminations. A
side-effect of this change is that input must always be from
stdin. Previously a filename argument was accepted as well.
avoid forking at every call. Changed all callers.
(func_mode_link): Simplified, and removed the final non-portable
nested escaped double quotes in back quotes usage in the process.
compiler flags need an argument or not, and can disect options and
pass parts through incorrectly. This changeset reverts both
2004-09-06 Albert Chin-A-Young (topmost) and 2004-09-05 Albert
Chin-A-Young below, and recognises more options explicitly:
* config/ltmain.in (func_mode_link): Pass through options needed
to compile in 64-bit mode with gcc, and the SGI, Sun, HP and IBM
compilers.
were convenience libs, so we did not test the creation of shared
libraries on any platform. Also none of the libs used any actual
fortran library calls, adding a call to write(*,*) in each lib
causes the tests to fail on darwin (and presumably other platforms).
These new tests would likely cause many more fortran test failures,
so if there is an error during make it is reported as SKIP.
* tests/f77demo-make.test: Don't fail hard, skip on failure.
* tests/f77demo/foof2.f: New file.
* tests/f77demo/foof3.f: New file.
* tests/f77demo/fprogram.f: Call routine in new lib.
* tests/f77demo/foof.f: Call routine in another new lib.
* tests/f77demo/Makefile.am: Make a couple of new libraries, add
$(FLIBS) to cprogram LDADD.
unmolested will prove to be a good plan for the 2.0 release. If
it goes awry, we can get back to where we were by reverting this
patch and the 2004-09-05 Albert Chin-A-Young patch, aka
gary@gnu.org--2004/libtool--devo--1.0--patch-169,172:
* config/ltmain.in (func_mode_link): Because we now pass through
compiler arguments we doesn't recognize, the code to pass
through GCC's -m* arguments is not needed.
already exists warnings" for .lib directory. Build on automake
wisdom to support make -j builds where a suitable mkdir is
available:
* configure.ac (AM_INIT_AUTOMAKE): Require 1.9, which goes to some
trouble to find a $(mkdir_p) that doesn't interfere with make -j.
* Makefile.am (edit): Substitute for @mkdir_p@.
* config/ltmain.in (MKDIR_P): Take @mkdir_p@ value.
(mkdir): Removed.
(func_mkdir_p): New function to use $MKDIR_P. Factor out all
previous mkdir calls to use this function.
(RM, MV): Declared as in libtoolize.in. Changed all callers.
* libtoolize.in (MKDIR_P): Take @mkdir_p@ value from automake.
(func_mkdir_p): Use it, similarly to new config/ltmain.in.
* tests/defs: Synchronize boiler plate code with ltmain.in.
Adjust all callers.
Add missing copyright preamble.
* tests/demo-static.test: Add missing copyright preamble.
Reported by Daniel Reed <n@ml.org>
upstream. Mostly comments, formal item boxing, a sort function,
and const madness reduction.
(slist_new): Removed.
(slist_box, slist_unbox, slist_sort): New.
(SListCompare, SListCallback): Swapped!
(slist_remove, slist_find): Change order of parameters for
orthogonality with slist_foreach. Changed all callers.
* libltdl/lt_dlloader.c (loader_cmp): Renamed to...
(loader_callback): ...this. Return boxed item.
(lt_dlloader_remove): Adjust to new loader_callback semantics;
unbox each removed item before returning.
Remove unused variable.
Remove const from name parameter, since the slist API cannot
guarantee userdata const-ancy for its callback functions.
(lt_dlloader_find): Adjust to new loader_callback semantics; need
to return the contents of the boxed item.
Remove const from name parameter, since the slist API cannot
guarantee userdata const-ancy for its callback functions.
* libltdl/lt_dlloader.h (lt_dlloader_find, lt_dlloader_remove):
Adjust to new constless footprint.
* libltdl/ltdl.c (ltdl_exit): The global `loaders' list is changed
variable `loader' is invalidated. Since some loaders may be
resident modules that cannot be unloaded (though we have none
yet), we must save each `next' address before calling
`lt_dlloader_remove'.
* NEWS: Updated.
* THANKS: Added Ralf.
from slist_new.
* libltdl/ltdl.c (loader_init): Trust lt_dlloader_add(), don't
overwrite its error type, and simplify.
Reported by Ralf Wildenhues <Ralf.Wildenhues@gmx.de>
preamble.
* tests/defs: Put a full m4sh.m4 style 'Be Bourne compatible'
preamble in here too.
* HACKING: Note that tests/defs needs synching with m4sh.m4 too.
* TODO: Add new item.
warning.
(foreach_dirinpath, try_dlopen, lt_dlpath_insertdir): Fix gcc
signed vs unsigned warning.
Reported by Bob Friesenhahn <bfriesen@simple.dallas.tx.us>
compatible. Instead, revert to the old way of doing things with
an array of symbol name vs. address, and adding the originator as
the first symbol but with a NULL address:
* config/ltmain.in (func_extract_dlsyms): Store originator as a
NULL address symbol.
* libltdl/ltdl.h (lt_dlsymbol): Removed.
(lt_dlsymlist): Remove originator field.
(LTDL_SET_PRELOADED_SYMBOLS): Adjust.
* libltdl/loaders/preopen.c (vm_open, lt_dlpreload_open): Adjust
for new types.
(vm_sym): Skip the new originator symbol when scanning symbol
names.
* m4/libtool.m4 (_LT_CMD_GLOBAL_SYMBOLS): Adjust preloaded symbols
test file to match.
* NEWS: Updated.
enable_shared, apply to the host architecture rather than per-
language. Running the tests over a broken compiler can corrupt
these values, when there is no C++ compiler for example, so we
skip most of the tests in those cases.
* m4/libtool.m4 (_LT_LANG_CXX_CONFIG): If caught_CXX_compiler was
set from configure.ac, then don't run all the compiler tests.
(_LT_LANG_F77_CONFIG): Similarly for caught_F77_compiler.
from the build tree that were left behind by previous versions of
libtool.
($top_srcdir)/stamp-vcl: Since this rule is already called on
every invocation of make, hook clean-ltmain-sh in here too.
Suggested by Ralf Wildenhues <Ralf.Wildenhues@gmx.de>
value of this function is often passed to free, don't declare it
`const'. Responsibility for this memory is given back to the
caller on return.
* libltdl/lt_dlloader.h (lt_dlloader_remove): Ditto.
* libltdl/ltdl.c (lt_dlexit): Adjust.
* NEWS: Updated.
dependency libraries from dlpreopened modules to libraries:
* config/ltmain.in: Support new -weak link mode option. Adjust
help message.
* libltdl/Makefile.am (libltdlc_la_LDFLAGS): Use it.
* doc/libtool.texi (Linking with dlopened modules): Document it.
(Link mode): Mention -weak.
* NEWS: Updated.
complaint now, and recently a few more have crept in. Make a
start at refactoring the code with shell functions, and
standardising on option parsing style:
* config/ltmain.in (func_echo, func_verbose, func_error)
(func_fatal_error, func_fatal_help, func_missing_arg)
(func_version, func_usage, func_help): Boiler plate shell
functions mostly copied from libtoolize.in. Changed *lots* of
shell snippets to call these functions as appropriate, rather than
inlining the same code over and over.
(func_config, func_features, func_enable_tag, func_mode_help):
Factored out from the shared option parsing loop. Write a new
shared option parser based on the one in libtoolize.in.
(func_check_version_match): Factored out from global scope.
(func_mode_compile, func_mode_execute, func_mode_finish)
(func_mode_install, func_mode_link, func_mode_uninstall): Break
each of the top level mode handlers out into a function. Adjust
main loop.
(modename, prevopt, prev, optarg, show_help): Removed in the
course of cleaning up the shared option parser.
(opt_duplicate_deps): Use `:' and `false' instead of `yes' and `',
so we can run it directly, rather than calling test.
instructions, and provide additional useful info alongside the bug
reporting address in --help output:
* Makefile.am (edit): Substitute $host_triplet.
* libtoolize.in: Output useful debug info with --help.
* config/ltmain.in: Ditto.
* README: Rewritten. Added copyright info.
* README-alpha: Rewritten. Added copyright info.
* HACKING: New file, partly taken from old README-alpha.
distributed files (m4/ltversion.m4 and config/ltmain.in) in the
readonly source tree:
* Makefile.am (stamp-vcl): Keep only one copy of stamp-vcl; in
$(top_srcdir), so that an up-to-date version is rolled into the
dist tarball even from a VPATH build.
($(top_srcdir)/config/ltmain.sh): Don't depend on Makefile,
otherwise ltmain.sh needs to be rebuilt after every config.status
run, and then configure needs to be rebuilt, and it all goes
horribly wrong :-/ Unfortunately, we can't depend on Makefile.am
either, because that is a circular dependency. The downside of
this change is that it is now possible to edit Makefile.am to
change the way ltmain.sh is generated, and the dependency
information can't tell that ltmain.sh needs regenerating.
the many and varied m4 symbols accumulated by libtool over the
years, considering also that modern autoconf can detect unexpanded
macros even without AC in the macro name:
* bootstrap: Remove libltdl/config.h from previous releases.
* m4/libtool.m4, m4/ltdl.m4, doc/libtool.texi: Basically, run a
giant sed transformation like this:
s,AC_LIB_LTDL,LTDL_INIT,g
s,AC_LIBLTDL_CONVENIENCE,LTDL_CONVENIENCE,g
s,AC_LIBLTDL_INSTALLABLE,LTDL_INSTALLABLE,g
s,AC_WITH_LTDL,LT_WITH_LTDL,g
s,AC_LTDL_ENABLE_INSTALL,_LT_ENABLE_INSTALL,g
s,AC_DEPLIBS_CHECK_METHOD,_LT_CHECK_MAGIC_METHOD,g
s,AC_LIBTOOL_OBJDIR,_LT_CHECK_OBJDIR,g
s,AC_LTDL_OBJDIR,_LT_CHECK_OBJDIR,g
s,AC_LTDL_DLPREOPEN,_LT_CHECK_DLPREOPEN,g
s,AC_LIBTOOL_SYS_MAX_CMD_LEN,LT_CMD_MAX_LEN,g
s,AC_LIBTOOL_SYS_GLOBAL_SYMBOL_PIPE,_LT_CMD_GLOBAL_SYMBOLS,g
s,AC_LIBTOOL_SYS_OLD_ARCHIVE,_LT_CMD_OLD_ARCHIVE,g
s,AC_PROG_LD_RELOAD_FLAG,_LT_CMD_RELOAD,g
s,AC_LIBTOOL_SYS_LIB_STRIP,_LT_CMD_STRIPLIB,g
s,AC_CHECK_LIBM,LT_LIB_M,g
s,AC_LTDL_DLLIB,LT_LIB_DLLOAD,g
s,AC_LIBTOOL_COMPILER_OPTION,_LT_COMPILER_OPTION,g
s,AC_LIBTOOL_PROG_CC_C_O,_LT_COMPILER_C_O,g
s,AC_LIBTOOL_PROG_COMPILER_NO_RTTI,_LT_COMPILER_NO_RTTI,g
s,AC_LIBTOOL_PROG_COMPILER_PIC,_LT_COMPILER_PIC,g
s,AC_LIBTOOL_SYS_HARD_LINK_LOCKS,_LT_COMPILER_FILE_LOCKS,g
s,AC_LTDL_DLSYM_USCORE,LT_FUNC_DLSYM_USCORE,g
s,AC_LIBTOOL_LINKER_OPTION,_LT_LINKER_OPTION,g
s,AC_LIBTOOL_PROG_LD_SHLIBS,_LT_LINKER_SHLIBS,g
s,AC_LIBTOOL_PROG_LD_HARDCODE_LIBPATH,_LT_LINKER_HARDCODE_LIBPATH,g
s,AC_PATH_MAGIC,_LT_PATH_MAGIC,g
s,AC_PATH_TOOL_PREFIX,_LT_PATH_TOOL_PREFIX,g
s,AC_PROG_LD,LT_PATH_LD,g
s,AC_PROG_LD_GNU,_LT_PATH_LD_GNU,g
s,AC_PROG_NM,LT_PATH_NM,g
s,AC_LTDL_SYS_DLOPEN_DEPLIBS,LT_SYS_DLOPEN_DEPLIBS,g
s,AC_LIBTOOL_DLOPEN_SELF,LT_SYS_DLOPEN_SELF,g
s,AC_LIBTOOL_POSTDEP_PREDEP,_LT_SYS_HIDDEN_LIBDEPS,g
s,AC_LTDL_SYSSEARCHPATH,LT_SYS_DLSEARCH_PATH,g
s,AC_LTDL_SHLIBEXT,LT_SYS_MODULE_EXT,g
s,AC_LTDL_SHLIBPATH,LT_SYS_MODULE_PATH,g
s,AC_LTDL_SYMBOL_USCORE,LT_SYS_SYMBOL_USCORE,g
s,AC_LIBTOOL_SYS_DYNAMIC_LINKER,_LT_SYS_DYNAMIC_LINKER,g
s,_LT_AC_TAGVAR,_LT_TAGVAR,g
s,_LT_AC_SYS_COMPILER,_LT_TAG_COMPILER,g
s,_LT_AC_PROG_ECHO_BACKSLASH,_LT_PROG_ECHO_BACKSLASH,g
s,_LT_AC_SYS_LIBPATH_AIX,_LT_SYS_MODULE_PATH_AIX,g
s,_LT_AC_SHELL_INIT,_LT_SHELL_INIT,g
s,_LT_AC_LOCK,_LT_ENABLE_LOCK,g
s,_LT_AC_CHECK_DLFCN,_LT_HEADER_DLFCN,g
s,_LT_AC_TRY_DLOPEN_SELF,_LT_TRY_DLOPEN_SELF,g
s,LT_AC_PROG_EGREP,_LT_DECL_EGREP,g
s,LT_AC_PROG_SED,_LT_DECL_SED,g
* doc/libtool.texi (Autoconf macros): Document exported macros.
* libtoolize.in: Compare ltdl.m4 serial numbers for LTDL_INIT
instead of newly obsoleted AC_LIB_LTDL.
(func_scan_files): Also set seen_ltdl for LTDL_INIT and
LT_WITH_LTDL.
* libltdl/ltdl.c (lt_dlforeachfile): LT_SYS_LIBSEARCH_PATH is not
an environment variable, it is an actual path.
Reported by Noah Mish <noah@cs.caltech.edu>
Makefile.
(timestamp): Factor out timestamp setting and edit it based on the
version number, so that it is empty for release versions.
($(top_srcdir)/config/ltmain.sh): Use it.
(libtoolize): Ditto.
`cd $(top_srcdir)' for this rule, we can't mv ltversion.tmp to the
relative $(top_srcdir) again! Also, run $(MKSTAMP) before changing
directories, or it won't be found.
($(top_srcdir)/config/ltmain.sh): Ditto.
(EXTRA_DIST): We should distribute ltversion.m4 too.
not actually be set, we should use SHELL here, also make darn sure
that SHELL is set to something. Bug from Willem Jan Palenstijn
<wpalenst@math.leidenuniv.nl>.
the installed libltdl sources for libtoolize need to put
libltdl/configure.ac in the parent package so that *it* can
subconfigure libltdl. That way we don't put any constraints on
the build system used by the parent package because the
libtoolized libltdl subtree is a self-contained sub-project:
* libltdl/Makefile.am (EXTRA_DIST): Add configure.ac.
(nobase_ltdldata_DATA): Change back to libltdl's own
configure.ac.
required to write this changeset -- especially that ltmain.sh is
the same on any machine for a given release, which I hadn't
noticed for some bizarre reason: It turns out that generating
distributed files from configure causes no end of hassle, as
evidenced by the many patches I've generated over the last few
days to try and get the dist and distcheck make rules to work.
Instead of all that hair, we now simply generate our distributed
files (now including ltmain.sh) with make rules -- and since
automake creates make variables for all AC_SUBSTs, that is really
easy. The code looks a lot more like automake and autoconf
Makefile.ams now, and doesn't have all the rough edges the earlier
hacky solution suffered from. We still generate libtool from
config.status, but that is not a distributed file, and doesn't
break the golden rule. Besides, there is way more going on there
than a bunch of substitutions:
* Makefile.am (edit): New common sed substitutions for files now
generated by make instead of config.status.
(CLEANFILES): Clean new tmp files.
(EXTRA_DIST): Add ltmain.sh.
(vcl-tmp): Reinstated.
(m4/ltversion.m4, config/ltmain.sh): New rules. Generate from
here instead of config.status.
(libtoolize): Ditto.
(libtool): Call config.status to regenerate if necessary.
(dist-hook): Removed.
* config/ltmain.in: Moved here from top_srcdir.
* README-alpha: Update instructions to check AS_SHELL_SANITIZE is
up to date.
* bootstrap: Rewritten. Generate m4/ltversion.m4 and
config/ltmain.sh because configure depends on them.
* configure.ac (AC_CONFIG_SRCDIR): Use libtoolize.in now that
ltmain.in has moved.
(AC_CONFIG_FILES): Don't generate distributed files,
config/ltmain.sh and libtoolize from config.status. We have make
rules to do that now.
help `make distcheck'; not only do we have to make a copy of
ltmain.sh available in the source tree, but we have to be really
careful about permission bit twiddling when we copy it across into
the possible read-only $(top_srcdir)/config directory!
quite rightly for every project except libtool, that there will be
an ltmain.sh in the source tree.
(EXTRA_DIST): As do the dist rules it generates.
(dist-hook): Even though we definitely don't want to distribute
our local ltmain.sh.
the generated m4/ltversion.m4. Except for CVS revisions:
* ltmain.in (package_revision): Don't try to deduce it from
TIMESTAMP, since that is empty for release versions; simply
substitute it...
* configure.ac (package_revision): ...from here, having
precalculated it before resetting TIMESTAMP for release versions.
* bootstrap: Force regeneration of m4/ltversion.m4, but retain
timestamps if there is no change to the content.
CVS autoconf, otherwise pdksh chokes on some nested escaped quotes
from libtool.m4. Using CVS autoconf protects configure from
non-bourne compatible shells in the same way this changeset
protects libtool and libtoolize:
* libtoolize.in: Include latest CVS autoconf AS_SHELL_SANITIZE
expansion to make various shells fully bourne compatible.
* ltmain.in: Ditto.
* m4/libtool.m4 (AC_PATH_TOOL_PREFIX): Remove spurious quotes, the
RHS of = assignment is not IFS split.
(_LT_LANG_CXX_CONFIG): Ditto.
(AC_LIBTOOL_POSTDEP_PREDEP): Ditto.
