STLport is an an adaptation of SGI's Standard Template Library.
The goal of the STLport effort is to make SGI's STL
implementation usable with most compilers while retaining full
functionality.
The Standard Template Library is a C++ library of container
classes, algorithms, and iterators providing many of the basic
algorithms and data structures of computer science.
PR: ports/8939
Submitted by: josh@quick.net
Spin is an efficient on-the-fly verification system
(a `model checker') for asynchronous concurrent systems,
such as data communication protocols, distributed operating
systems, database systems, etc.
It can be used to prove both safety and liveness properties,
including all correctness requirements expressible in linear
time temporal logic.
PR: ports/9058
Submitted by: pangolin@home.com
libPropList for property settings. Window Maker also uses libPropList,
as does wdm. However, there seem to be some differences between the two and
nevertheless, it would not be good to have future gnome ports depend on
Window Maker. I made this port from the libPropList found on the gnome ftp
sites; the distribution file is called libPropList-0.7.1gnome, but it builds
the library by default as libPropList. This conflicts with the libPropList
installed by Window Maker, so I made this port build the library as
libPropListgnome. So any future gnome ports that need libPropList should link
against libPropListgnome instead.
PR: ports/9006
Submitted by: Glenn Johnson <gljohns@bellsouth.net>
dispite their meanings. (Sometimes we're too smart for computers. :)
Found by: sorting ports/INDEX by "sort -t '|' +1 -2"
(Note: the whole "x11" category appears at the end with the above sort
command, but I'll leave that the way it is for now -- "ls" shows it
before other x11-* entries.)
Disable native shared library support, doesn't work, use own scheme as before.
Don't use shared library version numbering on ELF for now, needs more thought.
Document shared library version numbering scheme I chose.
Pump shared library version number on aout.
Better COMMENT and DESCR (still too long).
Remove patch to regognize now outdated FreeBSD JDK.
since as Satoshi points out, "it doesn't seem very useful to
have a meta-port that you can't package". We'll let the world
judge whether it is useful to them.
Nudged by: asami
be relevant, and think I've covered everything...
This now build and a 'make package' now works...with my track record on
this particular port, its bound to be still broken for *someone*, no?
intrinsic functions to allows a greter range for FORTRAN code to
be compiled with the stock f77.
PR: ports/8689
Reviewed by: jkoshy
Submitted by: Glenn Johnson <gljohns@bellsouth.net>
Fix Makefile:
to use bsd.port.{pre.post}.mk
check for proper libstdc++.a, depending on elf or aout
run autoconf before configure
Remove old patches and add:
patch-ak
search and find tcl8.0 header file
setup shared library extension properly according to elf vs aout
patch-al
properly find gtk.h
Leaving it as marked BROKEN, while I continue to figure out thie elf vs
aout stuff ...
===> Building package for mico-2.2.3
Creating package /usr/ports/packages/All/mico-2.2.3.tgz
Registering depends: Mesa-3.0 XFree86-3.3.2 gtk-1.0.6 qt-1.40 tcl-8.0.2.
Creating gzip'd tar ball in '/usr/ports/packages/All/mico-2.2.3.tgz'
tar: can't add file lib/libmicoaux2.2.3.so.1 : No such file or directory
tar: can't add file lib/libmicoqt2.2.3.so.1 : No such file or directory
tar: can't add file lib/libmicox2.2.3.so.1 : No such file or directory
/usr/sbin/pkg_create: tar command failed with code 256
*** Error code 1
add the dependency for glibstdc++28 (shouldn't the gcc28 port automatically
install this?) ...
Please try and let me know if this fixes the build problem...
is still the same.
-------
===> Building for mico-2.2.3
for i in admin include; do gmake -C $i adm || exit 1; done
gmake[1]: Entering directory `/usr/ports/devel/mico/work/mico/admin'
g++28 -I../include -fno-exceptions -I/usr/X11R6/include/X11/qt -I/usr/local/include -c mkdepend.cc -o mkdepend.o
mkdepend.cc:27: iostream.h: No such file or directory
mkdepend.cc:28: fstream.h: No such file or directory
mkdepend.cc:29: string: No such file or directory
mkdepend.cc:30: vector: No such file or directory
mkdepend.cc:31: set: No such file or directory
gmake[1]: *** [mkdepend.o] Error 1
gmake[1]: Leaving directory `/usr/ports/devel/mico/work/mico/admin'
gmake: *** [system] Error 1
*** Error code 2
Stop.
-------
===> Building for gide-0.0.11
cd . && aclocal
aclocal: configure.in: 7: macro `AM_PATH_GTK' not found in library
gmake: *** [aclocal.m4] Error 1
*** Error code 2
Stop.
mico and gcc 2.7.2.1 ... not usable with koffice at this time, as it requires
you to disabl-mini-stl, but that will require either gcc 2.8.1 or egcs to
make that happen, and still working out bugs with the more stable version of
egcs then is in ports...
something already there (PORTOBJFORMAT, OSVERSION) or move stuff from after
.include <bsd.port.mk> to before.
(This is not by any means the complete list but just the ones I've noticed
recently.)
with the standard gcc that comes with FreeBSD 3.0 (gcc 2.7.2.1) ... this
is not usable for compiling either corba or koffice, which is still being
worked on...
Please report any bugs on this to me, as I'm trying to work with the
MICO guys as far as getting FreeBSD recognized as an "official port" instead
of their current web page, which reports "somehow it works under FreeBSD"...
current bug in the egcs port (its not a bug in the port, its a bug in egcs
itself)...
Change dependencies so that it will compile using --disable-mini-stl
Portcheckout(1) checks out a FreeBSD port and all runtime
and buildtime depending ports into the current working
directory. The output is written to stdout as an
executable shell-script.
Compiling a FreeBSD port usually require a full tree of ports
in /usr/port. A cvs checkout or a cvs update command takes
a very long time and need much free space.
With portcheckout, you checkout only the parts of the port
tree which you really need. This is a magnitude faster!
Replaced the --with-tcl=/usr/include with just --with-tcl
Left in BROKEN status...can someone test and let me know if there is
anything else I've overlooked?
is definitely wrong.
===
cd ../libs; rm -f libmicox2.2.1.a; ln -f -s ../auxdir/libmicox2.2.1.a .
/usr/X11R6/bin/moc ../include/mico/qtmico.h -o qtmico.moc
c++ -I. -I../include -I/usr/ports/devel/mico/work/mico/./include/ministl -I/usr/X11R6/include -I/usr/X11R6/lib/glib/include -I/usr/X11R6/include/gtk10 -I/usr/X11R6/include -I/usr/local/include -c qtmico.cc -o qtmico.o
In file included from qtmico.cc:25:
../include/mico/qtmico.h:32: qapp.h: No such file or directory
../include/mico/qtmico.h:35: qsocknot.h: No such file or directory
../include/mico/qtmico.h:36: qtimer.h: No such file or directory
In file included from qtmico.cc:266:
qtmico.moc:17: qmetaobject.h: No such file or directory
gmake[1]: *** [qtmico.o] Error 1
gmake[1]: Leaving directory `/usr/ports/devel/mico/work/mico/auxdir'
gmake: *** [system] Error 1
*** Error code 2
Stop.
The port is marked BROKEN anyway...updating it to what I have so far that
appears to work, so that if I update ports, I don't lose it all again.
Hope to have it unBROKEN over the weekend ... so far, builds out of the
box with the stock cc 2.7.2.1 ...
===
/bin/sh ../libtool --mode=link cc -O -pipe -o ffitest ffitest.o libffi.la
libtool: link: warning: using a buggy system linker
libtool: link: relinking will be required before `ffitest' can be installed
cc -O -pipe -o .libs/ffitest ffitest.o /usr/ports/devel/libffi/work/libffi-1.18/src/.libs/libffi.so
creating ffitest
./ffitest
ld.so failed: Undefined symbol "ffi_prep_args" in ffitest:/usr/ports/devel/libffi/work/libffi-1.18/src/.libs/libffi.so
*** Error code 1
Stop.
Creating package /usr/ports/packages/All/tvision-0.6.tgz
Registering depends:.
Creating gzip'd tar ball in '/usr/ports/packages/All/tvision-0.6.tgz'
tar: can't add file lib/libtvision.so.0 : No such file or directory
/usr/sbin/pkg_create: tar command failed with code 256
*** Error code 1
g++ -fhandle-exceptions -pipe -I. -D__OMNIORB2__ -I../../../include -DDEFAULT_LOGDIR=\"/var/omninames\" -o omniNames NamingContext_i.o log.o omniNames.o -L../../../src/lib/omniORB2 -lomniORB2 -L../../../src/lib/omnithread -lomnithread
../../../src/lib/omnithread/libomnithread.a(posix.o): In function `omni_mutex::omni_mutex(void)':
posix.o(.text+0x10): undefined reference to `pthread_mutex_init'
../../../src/lib/omnithread/libomnithread.a(posix.o): In function `omni_mutex::~omni_mutex(void)':
posix.o(.text+0xa2): undefined reference to `pthread_mutex_destroy'
../../../src/lib/omnithread/libomnithread.a(posix.o): In function `omni_mutex::lock(void)':
===> Building package for linux_devel-0.2
Creating package /usr/ports/packages/All/linux_devel-0.2.tgz
Registering depends: linux_lib-2.5.
Creating gzip'd tar ball in '/usr/ports/packages/All/linux_devel-0.2.tgz'
tar: can't add file usr/lib/libfbm.so.1 : No such file or directory
tar: can't add file usr/lib/libjpeg.so.1 : No such file or directory
tar: can't add file usr/lib/libtiff.so.1 : No such file or directory
/usr/sbin/pkg_create: tar command failed with code 256
===> Building package for dlmalloc-2.6.4
Creating package /usr/ports/packages/All/dlmalloc-2.6.4.tgz
Registering depends:.
Creating gzip'd tar ball in '/usr/ports/packages/All/dlmalloc-2.6.4.tgz'
tar: can't add file lib/libdlmalloc.so.2 : No such file or directory
/usr/sbin/pkg_create: tar command failed with code 256
*** Error code 1
===> Generating temporary packing list
===> Compressing manual pages for p5-Data-ShowTable-3.3
/usr/local/man//man1/showtable.1: No such file or directory
*** Error code 1
Use right shared library names on ELF.
Disable Java support on ELF systems, because we don't have an ELF jdk.
Disable Java support on aout systems, because ILU runtime dumps core.
New PARC patches, now as one distfile instead of patches/ collection.
ILU now works with optimization enabled -> make a package.
Thread support now works, on -current and -stable. Thanks folks!
===
:
cc -c -O2 -pipe -Idevelhdrs xc_builtins.c
xc_builtins.c:37: X11/Xlib.h: No such file or directory
xc_builtins.c:38: X11/Xutil.h: No such file or directory
xc_builtins.c:39: X11/Xresource.h: No such file or directory
xc_builtins.c:101: parse error before `*'
xc_builtins.c:148: parse error before `*'
:
For the record, pkg_delete also gave me the following errors:
: cannot delete specified file `/a/tmp/hoek/usr/i486-linuxaout/lib/ldscripts - it is a directory!
: cannot delete specified file `/a/tmp/hoek/usr/include/linux' - it is a directory!
: cannot delete specified file `/a/tmp/hoek/usr/include/asm' - it is a directory!
There wouldn't happen to be a problem with pkg/PLIST, would there?
>> Checksum OK for ORBit-0.2.1.tar.gz.
===> Extracting for ORBit-0.2.1
===> ORBit-0.2.1 depends on executable: gmake - found
===> ORBit-0.2.1 depends on shared library: glib11\.0\. - found
===> Patching for ORBit-0.2.1
===> Applying FreeBSD patches for ORBit-0.2.1
===> Applying FreeBSD patch /usr/ports/devel/ORBit/patches/patch-aa
Hmm... Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|--- src/orb/orbit_poa.c.orig Thu Aug 13 21:11:35 1998
|+++ src/orb/orbit_poa.c Mon Aug 17 00:18:29 1998
--------------------------
Patching file src/orb/orbit_poa.c using Plan A...
Hunk #1 succeeded at 29.
done
===> Applying FreeBSD patch /usr/ports/devel/ORBit/patches/patch-ab
Hmm... Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|--- Makefile.in.orig Sat Aug 8 23:10:55 1998
|+++ Makefile.in Sat Aug 8 23:11:08 1998
--------------------------
Patching file Makefile.in using Plan A...
Hunk #1 succeeded at 86 with fuzz 1.
done
===> Applying FreeBSD patch /usr/ports/devel/ORBit/patches/patch-ac
Hmm... Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|--- src/ORBitutil/machine_endian.h.orig Tue Aug 18 20:29:55 1998
|+++ src/ORBitutil/machine_endian.h Tue Aug 18 20:30:59 1998
--------------------------
Patching file src/ORBitutil/machine_endian.h using Plan A...
Hunk #1 failed at 1.
1 out of 1 hunks failed--saving rejects to src/ORBitutil/machine_endian.h.rej
done
*** Error code 1
Stop.
+= for variables defined only once.
Also, MAKE_FLAGS is defined with ?=, not += (probably a bad idea,
but hard to change now) , in bsd.port.mk, so you need the "-f" in there.
The only reason why this worked was because there is a "Makefile"
target in the source's Makefile.
going to be a little verbose here about the changes so people who fix
up the rest of the stuff can have an idea what to do.
(1) Manpages are moved to version-specific subdirectories. Since they
are all defined in MAN? macros, I just needed to add
${PERL_VERSION} to the end of MAN*PREFIX and delete them from the
PLIST.
(2) Most of the module files are moved to version-specific
subdirectories. This is usually accomplished by adding
"%%PERL_VER%%/" right after "site_perl/" in the PLIST. However,
some things didn't move, and some moved more than one level (like,
all the way into "site_perl/%%PERL_VER%%/auto/FooBar") so some
fishing, err, find'ing was necessary.
(3) Fix up dependencies. See (2) above.
(4) While I was there, I changed some "@dirrm foobar"'s to "@unexec
rmdir %D/foobar 2>/dev/null || true", so there won't be any
warnings if a directory can't be removed due to other modules
sharing it.
Now all of devel/p5-* will install and deinstall without a single
warning. I hope others will follow suit. :)
:
===> Installing for crossgo32-djgpp2-2.01
===> crossgo32-djgpp2-2.01 depends on executable: i386-go32-gcc - found
===> Moving old crossgo libs away
mv: lib: No such file or directory
*** Error code 1
Stop.
except to official homepages, but in this batch there were a few
cases where a link was desireable anyways. For these, the form
http://url/ (FAQ/whatever)
was used.
