[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[SCM] Debian GNU Hurd packaging branch, master, updated. debian/20120710
|
From: |
Samuel Thibault |
|
Subject: |
[SCM] Debian GNU Hurd packaging branch, master, updated. debian/20120710-3-79-g768f71d |
|
Date: |
Tue, 27 Nov 2012 23:31:03 +0000 |
The following commit has been merged in the master branch:
commit 768f71db6da93af39eadb8a7df9ce64704c9a1f0
Author: Samuel Thibault <address@hidden>
Date: Tue Nov 27 23:49:56 2012 +0000
New ustream snapshot
- Migrates everything to libpthread.
- Bump libc0.3-dev build dependency to get pthread_hurd_cond_wait.
- Break libthreads-based netdde.
- Remove install-msgids.diff, pflocal.patch,
runsystem_setup_pflocal.patch, makedev_keep_options.patch,
makedev_parted_store.patch, ext2fs_nowait.patch, libpager_deadlock.patch,
libdiskfs_self-reauth.patch, libdiskfs_sync.patch, merged upstream
- Refresh exec_filename_exec.patch, exec_filename_fs.patch,
ext2fs_large_stores.patch, external.patch, init_try_runsystem.gnu.patch,
libdde_workqueue.patch, libports_stability.patch, pfinet_dhcp.patch,
posix-sigcodes.patch, proxy-defpager.diff.
* New patch ext2fs_large_stores_pthread.patch to move ext2fs large store
support to pthread.
diff --git a/debian/changelog b/debian/changelog
index fe6d139..a9921b4 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,5 +1,19 @@
-hurd (20120710-4) UNRELEASED; urgency=low
-
+hurd (20121127-1~0) UNRELEASED; urgency=low
+
+ * New ustream snapshot
+ - Migrates everything to libpthread.
+ - Bump libc0.3-dev build dependency to get pthread_hurd_cond_wait.
+ - Break libthreads-based netdde.
+ - Remove install-msgids.diff, pflocal.patch,
+ runsystem_setup_pflocal.patch, makedev_keep_options.patch,
+ makedev_parted_store.patch, ext2fs_nowait.patch, libpager_deadlock.patch,
+ libdiskfs_self-reauth.patch, libdiskfs_sync.patch, merged upstream
+ - Refresh exec_filename_exec.patch, exec_filename_fs.patch,
+ ext2fs_large_stores.patch, external.patch, init_try_runsystem.gnu.patch,
+ libdde_workqueue.patch, libports_stability.patch, pfinet_dhcp.patch,
+ posix-sigcodes.patch, proxy-defpager.diff.
+ * New patch ext2fs_large_stores_pthread.patch to move ext2fs large store
+ support to pthread.
* control: Add missing libx11-dev, x11proto-core-dev, pkg-config, xkb-data,
flex, and bison build-dependencies.
diff --git a/debian/control b/debian/control
index d1ab620..ab5254c 100644
--- a/debian/control
+++ b/debian/control
@@ -4,7 +4,7 @@ Priority: required
Maintainer: GNU Hurd Maintainers <address@hidden>
Standards-Version: 3.9.3
Build-Depends: mig (>= 1.3-2), gnumach-dev (>= 2:1.3.99.dfsg.cvs20070526),
- libc0.3-dev (>= 2.9), texinfo, texi2html, libncursesw5-dev, cdbs,
+ libc0.3-dev (>= 2.13-38~), texinfo, texi2html, libncursesw5-dev, cdbs,
debhelper (>= 4.2.0), gcc-4.7, dpkg (>= 1.15.4) | install-info, autoconf,
libparted-dev, libpciaccess-dev, libx11-dev, x11proto-core-dev, pkg-config,
xkb-data, flex, bison
@@ -18,7 +18,7 @@ Vcs-Git: git://git.debian.org/pkg-hurd/hurd.git
Package: hurd
Essential: yes
Depends: ${misc:Depends}, sysv-rc, netdde (>= 0.0.20120518~), libc0.3 (>=
2.13-31~), ${shlibs:Depends}, xkb-data
-Breaks: gnumach (<< 2:1.3.99.dfsg.cvs20070526-1), libc0.3 (<< 2.11.2-12)
+Breaks: gnumach (<< 2:1.3.99.dfsg.cvs20070526-1), libc0.3 (<< 2.11.2-12),
netdde (<< 0.0.20121127~)
Suggests: hurd-doc
Recommends: bf-utf-source
Provides: makedev, login, console-driver-xkb
diff --git a/debian/patches/ext2fs_large_stores_pthread.patch
b/debian/patches/ext2fs_large_stores_pthread.patch
index d410838..71f76a3 100644
--- a/debian/patches/ext2fs_large_stores_pthread.patch
+++ b/debian/patches/ext2fs_large_stores_pthread.patch
@@ -4,6 +4,8 @@ Date: Mon Sep 3 22:19:16 2012 +0200
Move large storage patch to pthreads
+TODO: merge into ext2fs_large_stores.patch.
+
Index: hurd-debian/ext2fs/ext2fs.h
===================================================================
--- hurd-debian.orig/ext2fs/ext2fs.h 2012-11-26 00:24:49.000000000 +0000
@@ -340,21 +342,3 @@ index 24533e7..0d71db7 100644
}
}
-diff --git a/libpager/lock-object.c b/libpager/lock-object.c
-index 93c89f0..c022d0c 100644
---- a/libpager/lock-object.c
-+++ b/libpager/lock-object.c
-@@ -65,11 +65,11 @@ _pager_lock_object (struct pager *p,
- }
- }
-
-- mutex_unlock (&p->interlock);
-+ pthread_mutex_unlock (&p->interlock);
- memory_object_lock_request (p->memobjcntl, offset, size, should_return,
- should_flush, lock_value,
- sync ? p->port.port_right : MACH_PORT_NULL);
-- mutex_lock (&p->interlock);
-+ pthread_mutex_lock (&p->interlock);
-
- if (sync)
- {
diff --git a/debian/patches/ext2fs_nowait.patch
b/debian/patches/ext2fs_nowait.patch
deleted file mode 100644
index c38b325..0000000
--- a/debian/patches/ext2fs_nowait.patch
+++ /dev/null
@@ -1,156 +0,0 @@
-Avoid waiting for disk I/O completion. This improves performance quite a bit,
-and should not be less safe.
-
-Merged upstream.
-
-diff --git a/ext2fs/dir.c b/ext2fs/dir.c
-index 66d8c8a..f0f52f7 100644
---- a/ext2fs/dir.c
-+++ b/ext2fs/dir.c
-@@ -690,7 +690,7 @@ diskfs_direnter_hard (struct node *dp, const char *name,
struct node *np,
- }
- }
-
-- diskfs_file_update (dp, 1);
-+ diskfs_file_update (dp, diskfs_synchronous);
-
- return 0;
- }
-@@ -727,7 +727,7 @@ diskfs_dirremove_hard (struct node *dp, struct dirstat *ds)
- if (dp->dn->dirents && dp->dn->dirents[ds->idx] != -1)
- dp->dn->dirents[ds->idx]--;
-
-- diskfs_file_update (dp, 1);
-+ diskfs_file_update (dp, diskfs_synchronous);
-
- return 0;
- }
-@@ -753,7 +753,7 @@ diskfs_dirrewrite_hard (struct node *dp, struct node *np,
struct dirstat *ds)
-
- munmap ((caddr_t) ds->mapbuf, ds->mapextent);
-
-- diskfs_file_update (dp, 1);
-+ diskfs_file_update (dp, diskfs_synchronous);
-
- return 0;
- }
-diff --git a/ext2fs/truncate.c b/ext2fs/truncate.c
-index 077225b..2058a1c 100644
---- a/ext2fs/truncate.c
-+++ b/ext2fs/truncate.c
-@@ -294,7 +294,7 @@ diskfs_truncate (struct node *node, off_t length)
- node->dn_stat.st_size = length;
- node->dn_set_mtime = 1;
- node->dn_set_ctime = 1;
-- diskfs_node_update (node, 1);
-+ diskfs_node_update (node, diskfs_synchronous);
- return 0;
- }
-
-@@ -309,6 +309,7 @@ diskfs_truncate (struct node *node, off_t length)
- {
- diskfs_node_rdwr (node, (void *)zeroblock, length, block_size - offset,
- 1, 0, 0);
-+ /* Make sure that really happens to avoid leaks. */
- diskfs_file_update (node, 1);
- }
-
-@@ -323,7 +324,7 @@ diskfs_truncate (struct node *node, off_t length)
- node->dn_stat.st_size = length;
- node->dn_set_mtime = 1;
- node->dn_set_ctime = 1;
-- diskfs_node_update (node, 1);
-+ diskfs_node_update (node, diskfs_synchronous);
-
- err = diskfs_catch_exception ();
- if (!err)
-diff --git a/libdiskfs/dir-init.c b/libdiskfs/dir-init.c
-index 2cba3a4..4efded0 100644
---- a/libdiskfs/dir-init.c
-+++ b/libdiskfs/dir-init.c
-@@ -63,6 +63,6 @@ diskfs_init_dir (struct node *dp, struct node *pdp, struct
protid *cred)
- return err;
- }
-
-- diskfs_node_update (dp, 1);
-+ diskfs_node_update (dp, diskfs_synchronous);
- return 0;
- }
-diff --git a/libdiskfs/dir-link.c b/libdiskfs/dir-link.c
-index 7cc8863..df1bb6a 100644
---- a/libdiskfs/dir-link.c
-+++ b/libdiskfs/dir-link.c
-@@ -101,7 +101,7 @@ diskfs_S_dir_link (struct protid *dircred,
- }
- np->dn_stat.st_nlink++;
- np->dn_set_ctime = 1;
-- diskfs_node_update (np, 1);
-+ diskfs_node_update (np, diskfs_synchronous);
-
- /* Attach it */
- if (tnp)
-diff --git a/libdiskfs/dir-rename.c b/libdiskfs/dir-rename.c
-index 867e395..298669b 100644
---- a/libdiskfs/dir-rename.c
-+++ b/libdiskfs/dir-rename.c
-@@ -164,7 +164,7 @@ diskfs_S_dir_rename (struct protid *fromcred,
- }
- fnp->dn_stat.st_nlink++;
- fnp->dn_set_ctime = 1;
-- diskfs_node_update (fnp, 1);
-+ diskfs_node_update (fnp, diskfs_synchronous);
-
- if (tnp)
- {
-diff --git a/libdiskfs/dir-renamed.c b/libdiskfs/dir-renamed.c
-index ce8f415..319a41a 100644
---- a/libdiskfs/dir-renamed.c
-+++ b/libdiskfs/dir-renamed.c
-@@ -177,7 +177,7 @@ diskfs_rename_dir (struct node *fdp, struct node *fnp,
const char *fromname,
- }
- fnp->dn_stat.st_nlink++;
- fnp->dn_set_ctime = 1;
-- diskfs_node_update (fnp, 1);
-+ diskfs_node_update (fnp, diskfs_synchronous);
-
- if (tnp)
- {
-diff --git a/libdiskfs/file-set-trans.c b/libdiskfs/file-set-trans.c
-index 26a19eb..c9b2c61 100644
---- a/libdiskfs/file-set-trans.c
-+++ b/libdiskfs/file-set-trans.c
-@@ -196,7 +196,7 @@ diskfs_S_file_set_translator (struct protid *cred,
- if (!error)
- {
- np->dn_stat.st_mode = newmode;
-- diskfs_node_update (np, 1);
-+ diskfs_node_update (np, diskfs_synchronous);
- }
- mutex_unlock (&np->lock);
- return error;
-diff --git a/libdiskfs/node-create.c b/libdiskfs/node-create.c
-index 4a7d108..5b5e463 100644
---- a/libdiskfs/node-create.c
-+++ b/libdiskfs/node-create.c
-@@ -131,7 +131,7 @@ diskfs_create_node (struct node *dir,
- if (S_ISDIR (mode))
- err = diskfs_init_dir (np, dir, cred);
-
-- diskfs_node_update (np, 1);
-+ diskfs_node_update (np, diskfs_synchronous);
-
- if (err)
- {
-diff --git a/libdiskfs/node-drop.c b/libdiskfs/node-drop.c
-index f44966b..c3d32c9 100644
---- a/libdiskfs/node-drop.c
-+++ b/libdiskfs/node-drop.c
-@@ -78,7 +78,7 @@ diskfs_drop_node (struct node *np)
- np->dn_stat.st_mode = 0;
- np->dn_stat.st_rdev = 0;
- np->dn_set_ctime = np->dn_set_atime = 1;
-- diskfs_node_update (np, 1);
-+ diskfs_node_update (np, diskfs_synchronous);
- diskfs_free_node (np, savemode);
- }
- else
diff --git a/debian/patches/install-msgids.diff
b/debian/patches/install-msgids.diff
deleted file mode 100644
index 975d38e..0000000
--- a/debian/patches/install-msgids.diff
+++ /dev/null
@@ -1,38 +0,0 @@
-2005-08-25 Alfred M. Szmidt <address@hidden>
-
- * Makefile (MSGIDS): New variable.
- (all, install-msgids, $(datadir)/msgids): New targets.
- (install): Specify install-header and install-msgids as
- prerequisites.
-
-merged upstream
-
----
- hurd/Makefile | 9 +++++++--
- 1 file changed, 7 insertions(+), 2 deletions(-)
-
---- a/hurd/Makefile
-+++ b/hurd/Makefile
-@@ -22,15 +22,20 @@ DIST_FILES = subsystems $(notdir $(hdrs)
-
- INSTHDRS = hurd_types.h version.h ioctl_types.h paths.h shared.h console.h \
- $(notdir $(wildcard $(srcdir)/*.defs))
-+MSGIDS := hurd.msgids $(patsubst %.defs,%.msgids,$(filter %.defs,$(INSTHDRS)))
-
- include ../Makeconf
-
--install-headers install: $(includedir)/hurd \
-+all: $(MSGIDS)
-+
-+install: install-msgids install-headers
-+install-headers: $(includedir)/hurd \
- $(addprefix $(includedir)/hurd/,$(INSTHDRS))
-+install-msgids: $(MSGIDS) $(datadir)/msgids; $(INSTALL_DATA) $^
-
- $(includedir)/hurd/%: $(srcdir)/%; $(INSTALL_DATA) $< $@
-
--$(includedir)/hurd:;mkdir -p $@
-+$(datadir)/msgids $(includedir)/hurd:;mkdir -p $@
-
- %.msgids: $(srcdir)/%.defs
- if grep -q '^subsystem' $<; \
diff --git a/debian/patches/libdiskfs_self-reauth.patch
b/debian/patches/libdiskfs_self-reauth.patch
deleted file mode 100644
index 24de393..0000000
--- a/debian/patches/libdiskfs_self-reauth.patch
+++ /dev/null
@@ -1,34 +0,0 @@
-"RPC to self with rendez-vous leading to duplicate port destroy"
-
-http://lists.gnu.org/archive/html/bug-hurd/2011-03/msg00045.html
-
-Merged upstream
-
-diff --git a/libfshelp/fetch-root.c b/libfshelp/fetch-root.c
-index 54d3c0c..6585052 100644
---- a/libfshelp/fetch-root.c
-+++ b/libfshelp/fetch-root.c
-@@ -79,7 +79,9 @@ fshelp_fetch_root (struct transbox *box, void *cookie,
- if (err)
- ret = MACH_PORT_NULL;
-
-- mach_port_destroy (mach_task_self (), rend);
-+ /* crdir could be ourselves, and thus diskfs_S_io_reauthenticate
-+ * would get the same name, so don't _destroy_ it !! */
-+ mach_port_mod_refs (mach_task_self (), rend, MACH_PORT_RIGHT_RECEIVE,
-1);
-
- return ret;
- }
-diff --git a/libshouldbeinlibc/exec-reauth.c b/libshouldbeinlibc/exec-reauth.c
-index dd267ef..12b34fc 100644
---- a/libshouldbeinlibc/exec-reauth.c
-+++ b/libshouldbeinlibc/exec-reauth.c
-@@ -59,7 +59,7 @@ exec_reauth (auth_t auth, int secure, int must_reauth,
- if (!err)
- err = auth_user_authenticate (auth, ref, MACH_MSG_TYPE_MAKE_SEND,
- &newport);
-- mach_port_destroy (mach_task_self (), ref);
-+ mach_port_mod_refs (mach_task_self (), ref, MACH_PORT_RIGHT_RECEIVE,
-1);
- if (err)
- {
- if (must_reauth)
diff --git a/debian/patches/libdiskfs_sync.patch
b/debian/patches/libdiskfs_sync.patch
deleted file mode 100644
index 19e4517..0000000
--- a/debian/patches/libdiskfs_sync.patch
+++ /dev/null
@@ -1,16 +0,0 @@
-Set diskfs stores as readonly on shutdown before enabling RPCs again.
-
-diff --git a/libdiskfs/init-startup.c b/libdiskfs/init-startup.c
-index bf1acf2..700a26e 100644
---- a/libdiskfs/init-startup.c
-+++ b/libdiskfs/init-startup.c
-@@ -148,6 +148,9 @@ diskfs_S_startup_dosync (mach_port_t handle)
- diskfs_set_hypermetadata (1, 1);
- _diskfs_diskdirty = 0;
-
-+ diskfs_readonly = 1;
-+ diskfs_readonly_changed (1);
-+
- ports_resume_class_rpcs (diskfs_protid_class);
- }
-
diff --git a/debian/patches/libpager_deadlock.patch
b/debian/patches/libpager_deadlock.patch
deleted file mode 100644
index b9bc4e1..0000000
--- a/debian/patches/libpager_deadlock.patch
+++ /dev/null
@@ -1,22 +0,0 @@
-See http://lists.gnu.org/archive/html/bug-hurd/2010-03/msg00127.html
-for the story. This should be completely safe, but Sergio believes
-memory_object_lock_request shouldn't actually block.
-
-merged upstream
-
-diff --git a/libpager/lock-object.c b/libpager/lock-object.c
-index d108666..d8d4a1b 100644
---- a/libpager/lock-object.c
-+++ b/libpager/lock-object.c
-@@ -65,9 +65,11 @@ _pager_lock_object (struct pager *p,
- }
- }
-
-+ mutex_unlock (&p->interlock);
- memory_object_lock_request (p->memobjcntl, offset, size, should_return,
- should_flush, lock_value,
- sync ? p->port.port_right : MACH_PORT_NULL);
-+ mutex_lock (&p->interlock);
-
- if (sync)
- {
diff --git a/debian/patches/makedev_keep_options.patch
b/debian/patches/makedev_keep_options.patch
deleted file mode 100644
index f967446..0000000
--- a/debian/patches/makedev_keep_options.patch
+++ /dev/null
@@ -1,60 +0,0 @@
-Add options -k and -K to MAKEDEV.sh to chose not to overwrite existing entries.
-
-TODO: submit to bug-hurd
-
-merged upstream
-
-Index: hurd/sutils/MAKEDEV.sh
-===================================================================
---- hurd.orig/sutils/MAKEDEV.sh 2010-07-01 18:14:22.000000000 +0200
-+++ hurd/sutils/MAKEDEV.sh 2010-07-01 18:17:37.000000000 +0200
-@@ -8,6 +8,8 @@
- ECHO=: # Change to "echo" to echo commands.
- EXEC="" # Change to ":" to suppress command execution.
- DEVDIR=`pwd` # Reset below by -D/--devdir command line option.
-+STFLAGS="-g" # Set to -k if active translators are to be kept.
-+KEEP= # Set to something if existing files are to be left alone.
-
- while :; do
- case "$1" in
-@@ -18,6 +20,8 @@
-
- -D, --devdir=DIR Use DIR when a device node name must be
- embedded in a translator; default is the cwd
-+ -k, --keep-active Leave any existing active translator running
-+ -K, --keep-all Don't overwrite existing files
- -n, --dry-run Don't actually execute any commands
- -v, --verbose Show what commands are executed to make the
devices
- -?, --help Give this help list
-@@ -28,11 +32,14 @@
- --devdir=*) DEVDIR="`echo "$1" | sed 's/^--devdir=//'`"; shift 1;;
- -D) DEVDIR="$2"; shift 2;;
- -D*) DEVDIR="`echo "$1" | sed 's/^-D//'`"; shift 1;;
-+ --keep-active|-k) STFLAGS="-k"; shift;;
-+ --keep-all|-K) KEEP=1; shift;;
- --verbose|-v) ECHO=echo; shift;;
- --dry-run|-n) EXEC=:; shift;;
- -nv|-vn) ECHO=echo; EXEC=:; shift;;
- --usage)
-- echo "Usage: $0 [-V?] [-D DIR] [--help] [--usage] [--version]
[--devdir=DIR] DEVNAME..."
-+ echo "Usage: $0 [-V?] [-D DIR] [--help] [--usage] [--version]"
-+ echo " [--devdir=DIR] [--keep-active] [--keep-all]
DEVNAME..."
- exit 0;;
- --version|-V)
- echo "STANDARD_HURD_VERSION_MAKEDEV_"; exit 0;;
-@@ -61,10 +68,13 @@
- local OWNER="$2"
- local PERM="$3"
- shift 3
-- if cmd settrans -cg "$NODE"; then
-+ if [ "$KEEP" ] && showtrans "$NODE" > /dev/null 2>&1 ; then
-+ return;
-+ fi
-+ if cmd settrans $STFLAGS -c "$NODE"; then
- cmd chown "$OWNER" "$NODE"
- cmd chmod "$PERM" "$NODE"
-- cmd settrans "$NODE" "$@"
-+ cmd settrans $STFLAGS "$NODE" "$@"
- fi
- }
-
diff --git a/debian/patches/makedev_parted_store.patch
b/debian/patches/makedev_parted_store.patch
deleted file mode 100644
index f4d4529..0000000
--- a/debian/patches/makedev_parted_store.patch
+++ /dev/null
@@ -1,75 +0,0 @@
-Add option -p to MAKEDEV to use user-space parted stores for partition devices.
-
-TODO: use in installed system, test, and submit to bug-hurd. IIRC the issue
with
-it was that the swapper does not like parted-based stores.
-
-merged upstream
-
-Index: hurd/sutils/MAKEDEV.sh
-===================================================================
---- hurd.orig/sutils/MAKEDEV.sh 2010-07-08 00:48:13.000000000 +0000
-+++ hurd/sutils/MAKEDEV.sh 2010-07-08 01:18:26.000000000 +0000
-@@ -10,6 +10,7 @@
- DEVDIR=`pwd` # Reset below by -D/--devdir command line option.
- STFLAGS="-g" # Set to -k if active translators are to be kept.
- KEEP= # Set to something if existing files are to be left alone.
-+USE_PARTSTORE= # Whether to use the newer part: stores
-
- while :; do
- case "$1" in
-@@ -22,6 +23,8 @@
- embedded in a translator; default is the cwd
- -k, --keep-active Leave any existing active translator running
- -K, --keep-all Don't overwrite existing files
-+ -p, --parted Prefer user-space parted stores to kernel devices
-+ for partition devices
- -n, --dry-run Don't actually execute any commands
- -v, --verbose Show what commands are executed to make the
devices
- -?, --help Give this help list
-@@ -34,11 +37,12 @@
- -D*) DEVDIR="`echo "$1" | sed 's/^-D//'`"; shift 1;;
- --keep-active|-k) STFLAGS="-k"; shift;;
- --keep-all|-K) KEEP=1; shift;;
-+ --parted|-p) USE_PARTSTORE=1; shift;;
- --verbose|-v) ECHO=echo; shift;;
- --dry-run|-n) EXEC=:; shift;;
- -nv|-vn) ECHO=echo; EXEC=:; shift;;
- --usage)
-- echo "Usage: $0 [-V?] [-D DIR] [--help] [--usage] [--version]"
-+ echo "Usage: $0 [-V?] [-D DIR] [--help] [--usage] [--version]
[--parted]"
- echo " [--devdir=DIR] [--keep-active] [--keep-all]
DEVNAME..."
- exit 0;;
- --version|-V)
-@@ -150,6 +154,7 @@
- ;;
-
- [hrsc]d*)
-+ local sliceno=
- local n="${I#?d}"
- local major="${n%%[!0-9]*}"
- if [ -z "$major" ]; then
-@@ -170,7 +175,9 @@
- ;;
- esac
- case "$rest" in
-- '') ;; # Whole slice
-+ '') # Whole slice, can use parted stores
-+ sliceno=$slice
-+ ;;
- [a-z]) ;; # BSD partition after slice
- *)
- lose "$I: Invalid partition \`$rest'"
-@@ -183,7 +190,12 @@
- esac
-
- # The device name passed all syntax checks, so finally use it!
-- st $I root 640 /hurd/storeio $I
-+ if [ "$USE_PARTSTORE" ] && [ -z "$rest" ] && [ "$sliceno" ]; then
-+ local dev=${I%s[0-9]*}
-+ st $I root 640 /hurd/storeio -T typed part:$sliceno:device:$dev
-+ else
-+ st $I root 640 /hurd/storeio $I
-+ fi
- ;;
-
- netdde)
diff --git a/debian/patches/pflocal.patch b/debian/patches/pflocal.patch
deleted file mode 100644
index 7ebba8d..0000000
--- a/debian/patches/pflocal.patch
+++ /dev/null
@@ -1,675 +0,0 @@
-2005-05-17 Neal H. Walfield <address@hidden>
-
- * connq.h (struct connq_request): Remove forward.
- (connq_listen): Wait for a request to be queued not until there is
- a connection attempt. Remove REQ parameter. Update callers.
- (connq_request_complete): Remove declaration.
- (connq_connect): Wait for a slot to queue a request not until
- there is an acceptor. Remove SOCK parameter. Update callers.
- (connq_connect_complete): New declaration.
- (connq_connect_cancel): New declaration.
-
- * connq.c (struct connq): Remove fields noqueue, queue, length,
- head and tail. Add fields head, tail, count, max, connectors and
- num_connectors. That is, replace the circular buffer with a
- singly linked list.
- (qnext): Remove function.
- (struct connq_request): Remove field signal, lock, completed and
- err. Add field next.
- (connq_request_init): Rewrite according to new semantics.
- (connq_request_enqueue): New function.
- (connq_request_dequeue): New function.
- (connq_create): Update according to new semantics.
- (connq_destroy): Likewise.
- (connq_listen): Rewrite to not block until there is a connector
- but until there is a request in the queue.
- (connq_request_complete): Remove function.
- (connq_connect): Rewrite to not block until there is an acceptor
- but until there is space for a request.
- (connq_connect_complete): New function.
- (connq_connect_cancel): New function.
- (connq_compress): Remove dead code.
- (connq_set_length): Rewrite.
-
- * socket.c (S_socket_connect): Create the server socket here...
- (S_socket_accept): ... not here.
-
-merged upstream
----
- pflocal/connq.c | 341
+++++++++++++++++++++++++++++--------------------------
- pflocal/connq.h | 39 +++---
- pflocal/io.c | 8 -
- pflocal/socket.c | 72 +++++------
- 4 files changed, 245 insertions(+), 215 deletions(-)
-
---- a/pflocal/connq.c
-+++ b/pflocal/connq.c
-@@ -26,31 +26,22 @@
- /* A queue for queueing incoming connections. */
- struct connq
- {
-- /* True if all connection requests should be treated as non-blocking. */
-- int noqueue;
--
- /* The connection request queue. */
-- struct connq_request **queue;
-- unsigned length;
-- /* Head is the position in QUEUE of the first request, and TAIL is the
-- first free position in the queue. If HEAD == TAIL, then the queue is
-- empty. Starting at HEAD, successive positions can be calculated by
-- using qnext(). */
-- unsigned head, tail;
-+ struct connq_request *head;
-+ struct connq_request **tail;
-+ unsigned count;
-+ unsigned max;
-
- /* Threads that have done an accept on this queue wait on this condition.
*/
- struct condition listeners;
- unsigned num_listeners;
-
-+ /* Threads that have done a connect on this queue wait on this condition.
*/
-+ struct condition connectors;
-+ unsigned num_connectors;
-+
- struct mutex lock;
- };
--
--/* Returns the position CQ's queue after POS. */
--static inline unsigned
--qnext (struct connq *cq, unsigned pos)
--{
-- return (pos + 1 == cq->length) ? 0 : pos + 1;
--}
-
- /* ---------------------------------------------------------------- */
-
-@@ -58,30 +49,50 @@ qnext (struct connq *cq, unsigned pos)
- get information from and to the thread. */
- struct connq_request
- {
-+ struct connq_request *next;
-+
- /* The socket that's waiting to connect. */
- struct sock *sock;
--
-- /* What the waiting thread blocks on. */
-- struct condition signal;
-- struct mutex lock;
--
-- /* Set to true when this request has been dealt with, to guard against
-- spurious conditions being signaled. */
-- int completed;
--
-- /* After the waiting thread is unblocked, this is the result, either 0 if
-- SOCK has been connected, or an error. */
-- error_t err;
- };
-
- static inline void
- connq_request_init (struct connq_request *req, struct sock *sock)
- {
-- req->err = 0;
- req->sock = sock;
-- req->completed = 0;
-- condition_init (&req->signal);
-- mutex_init (&req->lock);
-+}
-+
-+/* Enqueue connection request REQ onto CQ. CQ must be locked. */
-+static void
-+connq_request_enqueue (struct connq *cq, struct connq_request *req)
-+{
-+ assert (! mutex_try_lock (&cq->lock));
-+
-+ req->next = NULL;
-+ *cq->tail = req;
-+ cq->tail = &req->next;
-+
-+ cq->count ++;
-+}
-+
-+/* Dequeue a pending request from CQ. CQ must be locked and must not
-+ be empty. */
-+static struct connq_request *
-+connq_request_dequeue (struct connq *cq)
-+{
-+ struct connq_request *req;
-+
-+ assert (! mutex_try_lock (&cq->lock));
-+ assert (cq->head);
-+
-+ req = cq->head;
-+ cq->head = req->next;
-+ if (! cq->head)
-+ /* We just dequeued the last element. Fixup the tail pointer. */
-+ cq->tail = &cq->head;
-+
-+ cq->count --;
-+
-+ return req;
- }
-
- /* ---------------------------------------------------------------- */
-@@ -95,16 +106,20 @@ connq_create (struct connq **cq)
- struct connq *new = malloc (sizeof (struct connq));
-
- if (!new)
-- return ENOMEM;
-+ return ENOBUFS;
-+
-+ new->head = NULL;
-+ new->tail = &new->head;
-+ new->count = 0;
-+ /* By default, don't queue requests. */
-+ new->max = 0;
-
-- new->noqueue = 1; /* By default, don't queue requests. */
-- new->length = 0;
-- new->head = new->tail = 0;
-- new->queue = NULL;
- new->num_listeners = 0;
-+ new->num_connectors = 0;
-
- mutex_init (&new->lock);
- condition_init (&new->listeners);
-+ condition_init (&new->connectors);
-
- *cq = new;
- return 0;
-@@ -116,175 +131,189 @@ connq_destroy (struct connq *cq)
- {
- /* Everybody in the queue should hold a reference to the socket
- containing the queue. */
-- assert (cq->length == 0);
-- /* Nevertheless, malloc(0) or realloc(0) might allocate some small
-- space. */
-- if (cq->queue)
-- free (cq->queue);
-+ assert (! cq->head);
-+ assert (cq->count == 0);
-+
- free (cq);
- }
-
- /* ---------------------------------------------------------------- */
-
--/* Wait for a connection attempt to be made on CQ, and return the connecting
-- socket in SOCK, and a request tag in REQ. If REQ is NULL, the request is
-- left in the queue, otherwise connq_request_complete must be called on REQ
-- to allow the requesting thread to continue. If NOBLOCK is true,
-- EWOULDBLOCK is returned when there are no immediate connections
-- available. */
-+/* Return a connection request on CQ. If SOCK is NULL, the request is
-+ left in the queue. If NOBLOCK is true, EWOULDBLOCK is returned
-+ when there are no immediate connections available. */
- error_t
--connq_listen (struct connq *cq, int noblock,
-- struct connq_request **req, struct sock **sock)
-+connq_listen (struct connq *cq, int noblock, struct sock **sock)
- {
-+ error_t err = 0;
-+
- mutex_lock (&cq->lock);
-
-- if (noblock && cq->head == cq->tail)
-+ if (noblock && cq->count == 0 && cq->num_connectors == 0)
- {
- mutex_unlock (&cq->lock);
- return EWOULDBLOCK;
- }
-
-+ if (! sock && (cq->count > 0 || cq->num_connectors > 0))
-+ /* The caller just wants to know if a connection ready. */
-+ {
-+ mutex_unlock (&cq->lock);
-+ return 0;
-+ }
-+
- cq->num_listeners++;
-
-- while (cq->head == cq->tail)
-- if (hurd_condition_wait (&cq->listeners, &cq->lock))
-- {
-- cq->num_listeners--;
-- mutex_unlock (&cq->lock);
-- return EINTR;
-- }
-+ if (cq->count == 0)
-+ /* The request queue is empty. */
-+ {
-+ assert (! cq->head);
-+
-+ if (cq->num_connectors > 0)
-+ /* Someone is waiting for an acceptor. Signal that we can
-+ service their request. */
-+ condition_signal (&cq->connectors);
-+
-+ do
-+ if (hurd_condition_wait (&cq->listeners, &cq->lock))
-+ {
-+ cq->num_listeners--;
-+ err = EINTR;
-+ goto out;
-+ }
-+ while (cq->count == 0);
-+ }
-+
-+ assert (cq->head);
-
-- if (req != NULL)
-+ if (sock)
- /* Dequeue the next request, if desired. */
- {
-- *req = cq->queue[cq->head];
-- cq->head = qnext (cq, cq->head);
-- if (sock != NULL)
-- *sock = (*req)->sock;
-+ struct connq_request *req = connq_request_dequeue (cq);
-+ *sock = req->sock;
-+ free (req);
- }
-+ else if (cq->num_listeners > 0)
-+ /* The caller will not actually process this request but someone
-+ else could. (This case is rare but possible: it would require
-+ one thread to do a select on the socket and a second to do an
-+ accept.) */
-+ condition_signal (&cq->listeners);
-+ else
-+ /* There is no one else to process the request and the connection
-+ has now been initiated. This is not actually a problem as even
-+ if the current queue limit is 0, the connector will queue the
-+ request and another listener (should) eventually come along.
-+ (In fact it is very probably as the caller has likely done a
-+ select and will now follow up with an accept.) */
-+ ;
-
-- cq->num_listeners--;
--
-+ out:
- mutex_unlock (&cq->lock);
--
-- return 0;
--}
--
--/* Return the error code ERR to the thread that made the listen request REQ,
-- returned from a previous connq_listen. */
--void
--connq_request_complete (struct connq_request *req, error_t err)
--{
-- mutex_lock (&req->lock);
-- req->err = err;
-- req->completed = 1;
-- condition_signal (&req->signal);
-- mutex_unlock (&req->lock);
-+ return err;
- }
-
--/* Try to connect SOCK with the socket listening on CQ. If NOBLOCK is true,
-- then return EWOULDBLOCK immediately when there are no immediate
-- connections available. Neither SOCK nor CQ should be locked. */
-+/* Try to connect SOCK with the socket listening on CQ. If NOBLOCK is
-+ true, then return EWOULDBLOCK if there are no connections
-+ immediately available. On success, this call must be followed up
-+ either connq_connect_complete or connq_connect_cancel. */
- error_t
--connq_connect (struct connq *cq, int noblock, struct sock *sock)
-+connq_connect (struct connq *cq, int noblock)
- {
-- error_t err = 0;
-- unsigned next;
--
- mutex_lock (&cq->lock);
-
- /* Check for listeners after we've locked CQ for good. */
-- if ((noblock || cq->noqueue) && cq->num_listeners == 0)
-+
-+ if (noblock
-+ && cq->count + cq->num_connectors >= cq->max + cq->num_listeners)
-+ /* We are in non-blocking mode and would have to wait to secure an
-+ entry in the listen queue. */
- {
- mutex_unlock (&cq->lock);
- return EWOULDBLOCK;
- }
-
-- next = qnext (cq, cq->tail);
-- if (next == cq->tail)
-- /* The queue is full. */
-- err = ECONNREFUSED;
-- else
-- {
-- struct connq_request req;
-+ cq->num_connectors ++;
-+
-+ while (cq->count + cq->num_connectors > cq->max + cq->num_listeners)
-+ /* The queue is full and there is no immediate listener to service
-+ us. Block until we can get a slot. */
-+ if (hurd_condition_wait (&cq->connectors, &cq->lock))
-+ {
-+ cq->num_connectors --;
-+ mutex_unlock (&cq->lock);
-+ return EINTR;
-+ }
-
-- connq_request_init (&req, sock);
-+ mutex_unlock (&cq->lock);
-
-- cq->queue[cq->tail] = &req;
-- cq->tail = next;
-+ return 0;
-+}
-
-- /* Hold REQ.LOCK before we signal the condition so that we're sure
-- to be woken up. */
-- mutex_lock (&req.lock);
-- condition_signal (&cq->listeners);
-- mutex_unlock (&cq->lock);
-+/* Follow up to connq_connect. Completes the connect, SOCK is the new
-+ server socket. */
-+void
-+connq_connect_complete (struct connq *cq, struct sock *sock)
-+{
-+ struct connq_request *req;
-+
-+ req = malloc (sizeof (struct connq_request));
-+ if (! req)
-+ abort ();
-+
-+ connq_request_init (req, sock);
-
-- while (!req.completed)
-- condition_wait (&req.signal, &req.lock);
-- err = req.err;
-+ mutex_lock (&cq->lock);
-+
-+ assert (cq->num_connectors > 0);
-+ cq->num_connectors --;
-+
-+ connq_request_enqueue (cq, req);
-
-- mutex_unlock (&req.lock);
-+ if (cq->num_listeners > 0)
-+ /* Wake a listener up. We must consume the listener ref here as
-+ someone else might call this function before the listener
-+ thread dequeues this request. */
-+ {
-+ cq->num_listeners --;
-+ condition_signal (&cq->listeners);
- }
-
-- return err;
-+ mutex_unlock (&cq->lock);
- }
--
--#if 0
--/* `Compresses' CQ, by removing any NULL entries. CQ should be locked. */
--static void
--connq_compress (struct connq *cq)
-+
-+/* Follow up to connq_connect. Cancel the connect. */
-+void
-+connq_connect_cancel (struct connq *cq)
- {
-- unsigned pos;
-- unsigned comp_tail = cq->head;
-+ mutex_lock (&cq->lock);
-
-- /* Now compress the queue to remove any null entries we put in. */
-- for (pos = cq->head; pos != cq->tail; pos = qnext (cq, pos))
-- if (cq->queue[pos] != NULL)
-- /* This position has a non-NULL request, so move it to the end of the
-- compressed queue. */
-- {
-- cq->queue[comp_tail] = cq->queue[pos];
-- comp_tail = qnext (cq, comp_tail);
-- }
-+ assert (cq->num_connectors > 0);
-+ cq->num_connectors --;
-
-- /* Move back tail to only include what we kept in the queue. */
-- cq->tail = comp_tail;
-+ if (cq->count + cq->num_connectors >= cq->max + cq->num_listeners)
-+ /* A connector is blocked and could use the spot we reserved. */
-+ condition_signal (&cq->connectors);
-+
-+ mutex_unlock (&cq->lock);
- }
--#endif
-
--/* Set CQ's queue length to LENGTH. Any sockets already waiting for a
-- connections that are past the new length will fail with ECONNREFUSED. */
-+/* Set CQ's queue length to LENGTH. */
- error_t
--connq_set_length (struct connq *cq, int length)
-+connq_set_length (struct connq *cq, int max)
- {
-- mutex_lock (&cq->lock);
-+ int omax;
-
-- if (length > cq->length)
-- /* Growing the queue is simple... */
-- cq->queue = realloc (cq->queue, sizeof (struct connq_request *) * length);
-- else
-- /* Shrinking it less so. */
-- {
-- int i;
-- struct connq_request **new_queue =
-- malloc (sizeof (struct connq_request *) * length);
--
-- for (i = 0; i < cq->length && cq->head != cq->tail; i++)
-- {
-- if (i < length)
-- /* Keep this connect request in the queue. */
-- new_queue[length - i] = cq->queue[cq->head];
-- else
-- /* Punt this one. */
-- connq_request_complete (cq->queue[cq->head], ECONNREFUSED);
-- cq->head = qnext (cq, cq->head);
-- }
--
-- free (cq->queue);
-- cq->queue = new_queue;
-- }
-+ mutex_lock (&cq->lock);
-+ omax = cq->max;
-+ cq->max = max;
-
-- cq->noqueue = 0; /* Turn on queueing. */
-+ if (max > omax && cq->count >= omax && cq->count < max
-+ && cq->num_connectors >= cq->num_listeners)
-+ /* This is an increase in the number of connection slots which has
-+ made some slots available and there are waiting threads. Wake
-+ them up. */
-+ condition_broadcast (&cq->listeners);
-
- mutex_unlock (&cq->lock);
-
---- a/pflocal/connq.h
-+++ b/pflocal/connq.h
-@@ -23,9 +23,8 @@
-
- #include <errno.h>
-
--/* Unknown types */
-+/* Forward. */
- struct connq;
--struct connq_request;
- struct sock;
-
- /* Create a new listening queue, returning it in CQ. The resulting queue
-@@ -36,26 +35,26 @@ error_t connq_create (struct connq **cq)
- /* Destroy a queue. */
- void connq_destroy (struct connq *cq);
-
--/* Wait for a connection attempt to be made on CQ, and return the connecting
-- socket in SOCK, and a request tag in REQ. If REQ is NULL, the request is
-- left in the queue, otherwise connq_request_complete must be called on REQ
-- to allow the requesting thread to continue. If NOBLOCK is true,
-- EWOULDBLOCK is returned when there are no immediate connections
-- available. CQ should be unlocked. */
--error_t connq_listen (struct connq *cq, int noblock,
-- struct connq_request **req, struct sock **sock);
--
--/* Return the error code ERR to the thread that made the listen request REQ,
-- returned from a previous connq_listen. */
--void connq_request_complete (struct connq_request *req, error_t err);
-+/* Return a connection request on CQ. If SOCK is NULL, the request is
-+ left in the queue. If NOBLOCK is true, EWOULDBLOCK is returned
-+ when there are no immediate connections available. */
-+error_t connq_listen (struct connq *cq, int noblock, struct sock **sock);
-+
-+/* Try to connect SOCK with the socket listening on CQ. If NOBLOCK is
-+ true, then return EWOULDBLOCK if there are no connections
-+ immediately available. On success, this call must be followed up
-+ either connq_connect_complete or connq_connect_cancel. */
-+error_t connq_connect (struct connq *cq, int noblock);
-+
-+/* Follow up to connq_connect. Completes the connection, SOCK is the
-+ new server socket. */
-+void connq_connect_complete (struct connq *cq, struct sock *sock);
-+
-+/* Follow up to connq_connect. Cancel the connect. */
-+void connq_connect_cancel (struct connq *cq);
-
- /* Set CQ's queue length to LENGTH. Any sockets already waiting for a
-- connections that are past the new length will fail with ECONNREFUSED. */
-+ connections that are past the new length remain. */
- error_t connq_set_length (struct connq *cq, int length);
-
--/* Try to connect SOCK with the socket listening on CQ. If NOBLOCK is true,
-- then return EWOULDBLOCK immediately when there are no immediate
-- connections available. Neither SOCK nor CQ should be locked. */
--error_t connq_connect (struct connq *cq, int noblock, struct sock *sock);
--
- #endif /* __CONNQ_H__ */
---- a/pflocal/io.c
-+++ b/pflocal/io.c
-@@ -199,16 +199,16 @@ S_io_select (struct sock_user *user,
-
- if (*select_type & SELECT_READ)
- {
-- /* Wait for a connect. Passing in NULL for REQ means that the
-- request won't be dequeued. */
-- if (connq_listen (sock->listen_queue, 1, NULL, NULL) == 0)
-+ /* Wait for a connect. Passing in NULL for SOCK means that
-+ the request won't be dequeued. */
-+ if (connq_listen (sock->listen_queue, 1, NULL) == 0)
- /* We can satisfy this request immediately. */
- return 0;
- else
- /* Gotta wait... */
- {
- ports_interrupt_self_on_port_death (user, reply);
-- return connq_listen (sock->listen_queue, 0, NULL, NULL);
-+ return connq_listen (sock->listen_queue, 0, NULL);
- }
- }
- }
---- a/pflocal/socket.c
-+++ b/pflocal/socket.c
-@@ -110,7 +110,7 @@ S_socket_connect (struct sock_user *user
- else if (sock->flags & SOCK_CONNECTED)
- /* SOCK_CONNECTED is only set for connection-oriented sockets,
- which can only ever connect once. [If we didn't do this test
-- here, it would eventually fail when it the listening socket
-+ here, it would eventually fail when the listening socket
- tried to accept our connection request.] */
- err = EISCONN;
- else
-@@ -118,16 +118,35 @@ S_socket_connect (struct sock_user *user
- /* Assert that we're trying to connect, so anyone else trying
- to do so will fail with EALREADY. */
- sock->connect_queue = cq;
-- mutex_unlock (&sock->lock); /* Unlock SOCK while waiting. */
-+ /* Unlock SOCK while waiting. */
-+ mutex_unlock (&sock->lock);
-
-- /* Try to connect. */
-- err = connq_connect (cq, sock->flags & SOCK_NONBLOCK, sock);
-+ err = connq_connect (peer->listen_queue,
-+ sock->flags & SOCK_NONBLOCK);
-+ if (!err)
-+ {
-+ struct sock *server;
-+
-+ err = sock_clone (peer, &server);
-+ if (!err)
-+ {
-+ err = sock_connect (sock, server);
-+ if (!err)
-+ connq_connect_complete (peer->listen_queue, server);
-+ else
-+ sock_free (server);
-+ }
-
-- /* We can safely set CONNECT_QUEUE to NULL, as no one else can
-+ mutex_lock (&sock->lock);
-+ if (err)
-+ connq_connect_cancel (peer->listen_queue);
-+ }
-+
-+ /* We must set CONNECT_QUEUE to NULL, as no one else can
- set it until we've done so. */
-- mutex_lock (&sock->lock);
- sock->connect_queue = NULL;
- }
-+
- mutex_unlock (&sock->lock);
- }
- else
-@@ -157,42 +176,25 @@ S_socket_accept (struct sock_user *user,
- err = ensure_connq (sock);
- if (!err)
- {
-- struct connq_request *req;
- struct sock *peer_sock;
-
-- err =
-- connq_listen (sock->listen_queue, sock->flags & SOCK_NONBLOCK,
-- &req, &peer_sock);
-+ err = connq_listen (sock->listen_queue, sock->flags & SOCK_NONBLOCK,
-+ &peer_sock);
- if (!err)
- {
-- struct sock *conn_sock;
--
-- err = sock_clone (sock, &conn_sock);
-+ struct addr *peer_addr;
-+ *port_type = MACH_MSG_TYPE_MAKE_SEND;
-+ err = sock_create_port (peer_sock, port);
-+ if (!err)
-+ err = sock_get_addr (peer_sock, &peer_addr);
- if (!err)
- {
-- err = sock_connect (conn_sock, peer_sock);
-- if (!err)
-- {
-- struct addr *peer_addr;
-- *port_type = MACH_MSG_TYPE_MAKE_SEND;
-- err = sock_create_port (conn_sock, port);
-- if (!err)
-- err = sock_get_addr (peer_sock, &peer_addr);
-- if (!err)
-- {
-- *peer_addr_port = ports_get_right (peer_addr);
-- *peer_addr_port_type = MACH_MSG_TYPE_MAKE_SEND;
-- ports_port_deref (peer_addr);
-- }
-- else
-- /* TEAR DOWN THE CONNECTION XXX */;
-- }
-- if (err)
-- sock_free (conn_sock);
-+ *peer_addr_port = ports_get_right (peer_addr);
-+ *peer_addr_port_type = MACH_MSG_TYPE_MAKE_SEND;
-+ ports_port_deref (peer_addr);
- }
--
-- /* Communicate any error (or success) to the connecting thread. */
-- connq_request_complete (req, err);
-+ else
-+ /* TEAR DOWN THE CONNECTION XXX */;
- }
- }
-
diff --git a/debian/patches/runsystem_setup_pflocal.patch
b/debian/patches/runsystem_setup_pflocal.patch
deleted file mode 100644
index 3f8a96e..0000000
--- a/debian/patches/runsystem_setup_pflocal.patch
+++ /dev/null
@@ -1,22 +0,0 @@
-Automatically setup local sockets if not already done so. Needed early for
shell
-pipelines.
----
- daemons/runsystem.sh | 6 ++++++
- 1 file changed, 6 insertions(+)
-
-merged upstream
---- a/daemons/runsystem.sh
-+++ b/daemons/runsystem.sh
-@@ -53,6 +53,12 @@ function singleuser ()
- }
-
-
-+# See whether pflocal is setup already, and do so if not (install case)
-+
-+if ! test -e /servers/socket/1 && which settrans >/dev/null ; then
-+ settrans -c /servers/socket/1 /hurd/pflocal
-+fi
-+
- # We expect to be started by console-run, which gives us no arguments and
- # puts FALLBACK_CONSOLE=file-name in the environment if our console is
- # other than a normal /dev/console.
diff --git a/debian/patches/series b/debian/patches/series
index 9e65311..299b945 100644
--- a/debian/patches/series
+++ b/debian/patches/series
@@ -3,12 +3,9 @@ ext2fs_large_stores.patch
ext2fs_large_stores_pthread.patch
hurd_console_startup.patch
init_try_runsystem.gnu.patch
-#install-msgids.diff
makedev.diff
pfinet_dhcp.patch
-#pflocal.patch
rc.patch
-#runsystem_setup_pflocal.patch
startup-usr-support.patch
tmp_exec_startup.patch
uptime_w_path_fix.patch
@@ -16,13 +13,7 @@ stat_round.patch
libports_stability.patch
proxy-defpager.diff
external.patch
-#makedev_keep_options.patch
-#makedev_parted_store.patch
console_ignore_bdf_err.patch
-#ext2fs_nowait.patch
-#libpager_deadlock.patch
-#libdiskfs_self-reauth.patch
-#libdiskfs_sync.patch
posix-sigcodes.patch
random-default-fast.patch
libexec.patch
--
Debian GNU Hurd packaging
- [SCM] Debian GNU Hurd packaging branch, master, updated. debian/20120710-3-79-g768f71d, Samuel Thibault, 2012/11/27
- [SCM] Debian GNU Hurd packaging branch, master, updated. debian/20120710-3-79-g768f71d, Samuel Thibault, 2012/11/27
- [SCM] Debian GNU Hurd packaging branch, master, updated. debian/20120710-3-79-g768f71d, Samuel Thibault, 2012/11/27
- [SCM] Debian GNU Hurd packaging branch, master, updated. debian/20120710-3-79-g768f71d, Samuel Thibault, 2012/11/27
- [SCM] Debian GNU Hurd packaging branch, master, updated. debian/20120710-3-79-g768f71d,
Samuel Thibault <=