Whamcloud - gitweb
fs/lustre-release.git
13 years agocollect more info in checkstack
nikita [Thu, 28 Sep 2006 18:12:39 +0000 (18:12 +0000)]
collect more info in checkstack

13 years agosrc needs to be in the options list if it is going to be available for use.
brian [Wed, 20 Sep 2006 22:01:45 +0000 (22:01 +0000)]
src needs to be in the options list if it is going to be available for use.

13 years agofile quota-umount-race-fix.patch was initially added on branch b1_4.
niu [Fri, 15 Sep 2006 03:55:53 +0000 (03:55 +0000)]
file quota-umount-race-fix.patch was initially added on branch b1_4.

13 years agofile quota-deadlock-on-pagelock-ext3.patch was initially added on branch b1_4.
niu [Fri, 15 Sep 2006 03:55:52 +0000 (03:55 +0000)]
file quota-deadlock-on-pagelock-ext3.patch was initially added on branch b1_4.

13 years agofile quota-deadlock-on-pagelock-core.patch was initially added on branch b1_4.
niu [Fri, 15 Sep 2006 03:55:51 +0000 (03:55 +0000)]
file quota-deadlock-on-pagelock-core.patch was initially added on branch b1_4.

13 years agob=22481
nathan [Tue, 12 Sep 2006 17:48:18 +0000 (17:48 +0000)]
b=22481
r=nathan
Perl script to parse and graph IOR results data
(linsyssoft)

13 years agofile flocks_test.c was initially added on branch b1_4.
deen [Mon, 11 Sep 2006 16:31:21 +0000 (16:31 +0000)]
file flocks_test.c was initially added on branch b1_4.

13 years agoBranch b1_4
adilger [Fri, 8 Sep 2006 23:45:13 +0000 (23:45 +0000)]
Branch b1_4
Add x86_64 support for checkstack.pl

13 years agorevert previous changes as per the discussion between adilger and brian.
wangyb [Fri, 8 Sep 2006 04:20:16 +0000 (04:20 +0000)]
revert previous changes as per the discussion between adilger and brian.

13 years agofile bitops_ext2_find_next_le_bit-2.6.patch was initially added on branch b1_4.
bobijam [Wed, 6 Sep 2006 03:58:05 +0000 (03:58 +0000)]
file bitops_ext2_find_next_le_bit-2.6.patch was initially added on branch b1_4.

13 years agofile lustre_rmmod.sh was initially added on branch b1_5.
nathan [Sat, 2 Sep 2006 00:08:10 +0000 (00:08 +0000)]
file lustre_rmmod.sh was initially added on branch b1_5.

13 years agocvsignore quilt's series
qiyong [Fri, 1 Sep 2006 05:54:36 +0000 (05:54 +0000)]
cvsignore quilt's series

13 years agofile export-do_kern_mount.patch was initially added on branch b1_5.
qiyong [Thu, 31 Aug 2006 06:22:24 +0000 (06:22 +0000)]
file export-do_kern_mount.patch was initially added on branch b1_5.

13 years agofile lookup_bdev_init_intent-2.6.18-vanilla.patch was initially added on branch b1_5.
qiyong [Thu, 31 Aug 2006 05:18:04 +0000 (05:18 +0000)]
file lookup_bdev_init_intent-2.6.18-vanilla.patch was initially added on branch b1_5.

13 years agor=adilger
wangyb [Wed, 30 Aug 2006 08:31:06 +0000 (08:31 +0000)]
r=adilger
Ltest uses 'lbuild' to make rpms. When there're multiple ltestds do 'lbuild' at the same time, there's a good chance in a kernel or lustre building race if they build the rpms in a same directory. So we should use separate buildroot to avoid such issue. Also make changes to get 'make rpms' and 'make srpm' work smoothly with this change to lustre.spec.in.

13 years agolustre-kernel-2.4.spec uses '_tmppath' instead of '_tmpdir' in its Builtroot variable.
wangyb [Wed, 30 Aug 2006 06:41:59 +0000 (06:41 +0000)]
lustre-kernel-2.4.spec uses '_tmppath' instead of '_tmpdir' in its Builtroot variable.
so let's pass the right macro definition to it.

13 years agoTry to avoid a race where two rpmbuilds on the same node are trying to stage
brian [Sun, 27 Aug 2006 14:12:24 +0000 (14:12 +0000)]
Try to avoid a race where two rpmbuilds on the same node are trying to stage
the same rpm -- which ends up happening in the same directory. In order for
this race avoidance to work, the caller has to supply a unique TMPDIR which
rpmbuild will use at it's _tmpdir. Default to /var/tmp if not supplied.

13 years agoadd libiam.a if its source is in lustre tree
tappro [Sun, 27 Aug 2006 12:42:56 +0000 (12:42 +0000)]
add libiam.a if its source is in lustre tree

13 years agoWhile this looks like it should work, and Yury says he tested it working
brian [Sun, 27 Aug 2006 01:13:36 +0000 (01:13 +0000)]
While this looks like it should work, and Yury says he tested it working
on both b_new_cmd and b1_5, it's obviously not working for some cases:
http://buffalo.lustre.org:8080/display_report.pl?report_id=278455
I am not sure why though.  It would seem like lustre.spec.in is not getting
processed by configure, but all indications are that it is.  Why is the
@RPM_LIBIAM_PATH@ substitution not being done then?

13 years agob=10851
cliffw [Thu, 24 Aug 2006 20:34:46 +0000 (20:34 +0000)]
b=10851

Includes new ost-survey script, fixes to obdfiler-survey from scjody, new echo.sh for obdfilter setup from eeb

13 years ago- add libiam.a (if exists) to rpm, this is neededfor testing our branch;
yury [Thu, 24 Aug 2006 15:27:24 +0000 (15:27 +0000)]
- add libiam.a (if exists) to rpm, this is neededfor testing our branch;
- replace all memcmp of fids by lu_fid_eq;
- allocate new fid for create case in ll_revalidate_it() in case of doing "goto do_lookup";
- removed needless lmv_fld_lookup() in lmv;
- fixed lmv_obj refcount management in error cases;
- comments are adjusted to be coherent with 80 columns style.

13 years ago- switch to b_hd_newconfig
yury [Sun, 20 Aug 2006 13:59:30 +0000 (13:59 +0000)]
- switch to b_hd_newconfig

13 years agofile jbd-stats-2.6.5.patch was initially added on branch b1_4.
alex [Wed, 16 Aug 2006 20:26:09 +0000 (20:26 +0000)]
file jbd-stats-2.6.5.patch was initially added on branch b1_4.

13 years agofile jbd-stats-2.6.9.patch was initially added on branch b1_4.
alex [Wed, 16 Aug 2006 20:26:08 +0000 (20:26 +0000)]
file jbd-stats-2.6.9.patch was initially added on branch b1_4.

13 years agoAnswer for a [Y/?] question.
brian [Wed, 16 Aug 2006 18:44:42 +0000 (18:44 +0000)]
Answer for a [Y/?] question.

13 years agoGet the RC from update_oldconfig properly.
brian [Wed, 16 Aug 2006 18:37:09 +0000 (18:37 +0000)]
Get the RC from update_oldconfig properly.

13 years agoTeach lustrecvs and buildcvs about b_release_1_4_7-test branch. sigh.
green [Wed, 16 Aug 2006 15:05:03 +0000 (15:05 +0000)]
Teach lustrecvs and buildcvs about b_release_1_4_7-test branch. sigh.

13 years agofile 2.6-sles10.series was initially added on branch b1_5.
qiyong [Fri, 4 Aug 2006 06:44:10 +0000 (06:44 +0000)]
file 2.6-sles10.series was initially added on branch b1_5.

13 years agofile tcp-zero-copy-2.6-sles10.patch was initially added on branch b1_5.
qiyong [Fri, 4 Aug 2006 06:44:09 +0000 (06:44 +0000)]
file tcp-zero-copy-2.6-sles10.patch was initially added on branch b1_5.

13 years agofile vfs_nointent-2.6-sles10.patch was initially added on branch b1_5.
qiyong [Fri, 4 Aug 2006 06:44:08 +0000 (06:44 +0000)]
file vfs_nointent-2.6-sles10.patch was initially added on branch b1_5.

13 years agofile vfs_intent-2.6-sles10.patch was initially added on branch b1_5.
qiyong [Fri, 4 Aug 2006 06:44:07 +0000 (06:44 +0000)]
file vfs_intent-2.6-sles10.patch was initially added on branch b1_5.

13 years agofile ext3-multi-mount-protection-2.6-fc5.patch was initially added on branch b1_5.
qiyong [Mon, 24 Jul 2006 03:03:33 +0000 (03:03 +0000)]
file ext3-multi-mount-protection-2.6-fc5.patch was initially added on branch b1_5.

13 years agoadd an entry for b1_8.
ericm [Fri, 21 Jul 2006 19:27:34 +0000 (19:27 +0000)]
add an entry for b1_8.

14 years agoUTS_RELEASE moves from include/linux/version.h to utsrelease.h since v2.6.18.
qiyong [Mon, 17 Jul 2006 07:02:50 +0000 (07:02 +0000)]
UTS_RELEASE moves from include/linux/version.h to utsrelease.h since v2.6.18.

14 years agoBranch b1_4
scjody [Wed, 12 Jul 2006 22:33:58 +0000 (22:33 +0000)]
Branch b1_4

Add directories used by quilt to .cvsignore

14 years ago* landing build diffs from b_release_1_4_7_1
eeb [Mon, 10 Jul 2006 20:17:23 +0000 (20:17 +0000)]
*   landing build diffs from b_release_1_4_7_1

14 years agoBe more specific so that other lustre* cruft is not picked up.
brian [Mon, 3 Jul 2006 16:40:18 +0000 (16:40 +0000)]
Be more specific so that other lustre* cruft is not picked up.

14 years agoIf the kernel doesn't have a patch series, it's not lustre specific.
brian [Mon, 3 Jul 2006 01:12:40 +0000 (01:12 +0000)]
If the kernel doesn't have a patch series, it's not lustre specific.
Take two.

14 years agoLet's even be specific about the fact that it's the script library that we
brian [Sat, 1 Jul 2006 18:57:27 +0000 (18:57 +0000)]
Let's even be specific about the fact that it's the script library that we
want packaged.

14 years agoNeed to package the scripts in $libdir/lustre/.
brian [Sat, 1 Jul 2006 18:54:59 +0000 (18:54 +0000)]
Need to package the scripts in $libdir/lustre/.

14 years agoTrying to strace this sucker to see where it's going wrong on the occasions
brian [Fri, 30 Jun 2006 17:05:12 +0000 (17:05 +0000)]
Trying to strace this sucker to see where it's going wrong on the occasions
that it does has it's own set of problems.  ~sigh~

14 years agoLustre support needed to be able to build a 2.6 patchless kernel. This
brian [Tue, 27 Jun 2006 02:29:32 +0000 (02:29 +0000)]
Lustre support needed to be able to build a 2.6 patchless kernel.  This
simply builds a vanilla 2.6.16.22 kernel for the client nodes.

14 years agoAdded scripts to parse output per Peter Braam
cliffw [Thu, 22 Jun 2006 23:31:52 +0000 (23:31 +0000)]
Added scripts to parse output per Peter Braam

14 years agoBranch HEAD
scjody [Fri, 16 Jun 2006 03:27:40 +0000 (03:27 +0000)]
Branch HEAD
r=brian

Merge from b_release_1_4_6: Run update_oldconfig only if oldconfig fails;
always strace update_oldconfig so we can debug it.

14 years agob=8007
nathan [Wed, 7 Jun 2006 18:33:45 +0000 (18:33 +0000)]
b=8007
use default mode for modified modprobe.conf instead of mktemp mode

14 years agob=8007
nathan [Wed, 7 Jun 2006 16:52:52 +0000 (16:52 +0000)]
b=8007
remove lustre-ldap

14 years agob=8007
nathan [Wed, 7 Jun 2006 16:42:50 +0000 (16:42 +0000)]
b=8007
  - change module name from "llite.ko" to "lustre.ko"
  - drop ldap and python from rpms
  - remove initd scripts from rpms

14 years agoStarting with 1.6.0 our kernel package will now actually be
brian [Wed, 7 Jun 2006 03:39:10 +0000 (03:39 +0000)]
Starting with 1.6.0 our kernel package will now actually be
"kernel-lustre"[-smp] so that package management (i.e. yum) does not treat
our kernel equal to the vendor's kernel and allow inadvertently upgrading
from one to the other.
We still "provide" a "kernel" functionality, so in fact our
kernel-lustre[-smp] can be installed on say RHEL4 and then the vendor kernel
package can be completely removed because our kernel-lustre[-smp] will
satisfy all required dependecies with regard to "kernel".

14 years agor=nathan
brian [Wed, 7 Jun 2006 03:22:07 +0000 (03:22 +0000)]
r=nathan
build-core's HEAD is going to diverge for 1.6.0, so "capture" build-core's
current state in a branch mainly for b1_4.
Note that b_release_1_4_6-patchless is not really a release branch and thus
tracks b1_4.
Also, v1_2_8 and v1_4_0 pre-existed this build system so which branch to
check out for them is actually quite moot.  Let's just keep them tracking
b1_4.

14 years agob=8007
nathan [Tue, 6 Jun 2006 23:36:24 +0000 (23:36 +0000)]
b=8007
aagh, broke old b1_4 stuff - wait until we branch lustre-build

14 years agob=8007
nathan [Tue, 6 Jun 2006 22:35:09 +0000 (22:35 +0000)]
b=8007
- change module name from "llite.ko" to "lustre.ko" (r=adilger)
- drop ldap and python from rpms
- remove initd scripts from rpms
- remove old cluster_scripts dir

14 years agoTemp fix: gcc with 64-bit support (like in debian) doesn't mean the target is 64...
qiyong [Tue, 6 Jun 2006 06:05:01 +0000 (06:05 +0000)]
Temp fix: gcc with 64-bit support (like in debian) doesn't mean the target is 64-bit.

14 years agoNeed to use the flavtgt instead of flavor. flavtgt == "" when flavor == up.
brian [Wed, 31 May 2006 20:44:21 +0000 (20:44 +0000)]
Need to use the flavtgt instead of flavor.  flavtgt == "" when flavor == up.

14 years agoBranch b1_4
adilger [Wed, 31 May 2006 19:37:15 +0000 (19:37 +0000)]
Branch b1_4
Move checkstack target up to top-level autoMakefile so "make checkstack"
works more easily.  This will print out the top stack users (though not
total callchain stack usage).

14 years agoreally hook gss configuration into build system.
ericm [Wed, 31 May 2006 18:36:09 +0000 (18:36 +0000)]
really hook gss configuration into build system.

14 years ago* Applied obdfilter-survey fix from 10490 (all the threads were working on
eeb [Thu, 18 May 2006 17:28:35 +0000 (17:28 +0000)]
*   Applied obdfilter-survey fix from 10490 (all the threads were working on
    the first object).

14 years agoJust when you thought you had all the possibilities covered...
brian [Thu, 18 May 2006 01:19:44 +0000 (01:19 +0000)]
Just when you thought you had all the possibilities covered...

14 years agoDisable the stracing of update_oldconfig.
brian [Wed, 17 May 2006 19:04:12 +0000 (19:04 +0000)]
Disable the stracing of update_oldconfig.

14 years agoBranch: HEAD
wangdi [Thu, 11 May 2006 03:38:32 +0000 (03:38 +0000)]
Branch: HEAD
create new branch b_new_portals for colibri team

14 years agoBranch: HEAD
wangdi [Tue, 9 May 2006 04:21:59 +0000 (04:21 +0000)]
Branch: HEAD
create new branch b_new_lnet for new cmd

14 years agoBranch b1_4
adilger [Thu, 4 May 2006 22:08:42 +0000 (22:08 +0000)]
Branch b1_4
Add portals back as an "optional" directory, so build doesn't fail when
it is already checked out.

14 years agoBranch b1_4
adilger [Thu, 4 May 2006 20:02:44 +0000 (20:02 +0000)]
Branch b1_4
Don't try to check out portals anymore, we don't need it.
r=eeb

14 years agoBranch b1_4
adilger [Thu, 4 May 2006 08:04:02 +0000 (08:04 +0000)]
Branch b1_4
Per Eric's request - remove portals from the build tree, as he is testing
with Cray portals directly, when testing the ptllnd.
r=eeb

14 years agob1_4-patchless lnet version requirements are same as for b1_4
green [Mon, 1 May 2006 14:06:12 +0000 (14:06 +0000)]
b1_4-patchless lnet version requirements are same as for b1_4

14 years agob=10466
zhanghc [Fri, 28 Apr 2006 04:39:06 +0000 (04:39 +0000)]
b=10466
r=adilger@clusterfs.com
description: After calling 'symlink(path, newpath)', of which path(link target)
doesn't exist,  the second symlink call still success, which should return
-EEXIST.

14 years agoremoving bonsai test file; please ignore
scjody [Fri, 14 Apr 2006 17:36:59 +0000 (17:36 +0000)]
removing bonsai test file; please ignore

14 years agobonsai test; please ignore
scjody [Fri, 14 Apr 2006 17:36:20 +0000 (17:36 +0000)]
bonsai test; please ignore

14 years agobonsai test; please ignore
scjody [Wed, 12 Apr 2006 12:57:40 +0000 (12:57 +0000)]
bonsai test; please ignore

14 years agobonsai test; please ignore
scjody [Tue, 11 Apr 2006 20:38:51 +0000 (20:38 +0000)]
bonsai test; please ignore

14 years agobonsai test; please ignore
scjody [Tue, 11 Apr 2006 20:19:51 +0000 (20:19 +0000)]
bonsai test; please ignore

14 years ago* turns out we actually need to build against real sources for patchless
mjmac [Fri, 7 Apr 2006 17:47:04 +0000 (17:47 +0000)]
* turns out we actually need to build against real sources for patchless
* massage the kernel Makefile's EXTRAVERSION to match the running kernel

14 years agoThe variable, not the literal. Thanks mjmac.
brian [Thu, 6 Apr 2006 14:15:05 +0000 (14:15 +0000)]
The variable, not the literal.  Thanks mjmac.

14 years agoFor the patchless client, LNET needs to come from it's own branch too.
brian [Thu, 6 Apr 2006 14:14:14 +0000 (14:14 +0000)]
For the patchless client, LNET needs to come from it's own branch too.

14 years agoRevert my use of the configure_flags macro in the lustre.spec file and just
brian [Thu, 6 Apr 2006 14:03:12 +0000 (14:03 +0000)]
Revert my use of the configure_flags macro in the lustre.spec file and just
pass the --disable-server argument to configure with the CONFIGURE_FLAGS
variable.
Prefer (by boolean switch) mjmac's use of kernel-devel for building the
patchless client.

14 years ago~sigh~ rpm macros are not empty if undefined.
brian [Wed, 5 Apr 2006 17:35:48 +0000 (17:35 +0000)]
~sigh~  rpm macros are not empty if undefined.

14 years agoNo RPM is built when NO_SRC=0, but we should have the rpms being built.
wangyb [Wed, 5 Apr 2006 07:46:05 +0000 (07:46 +0000)]
No RPM is built when NO_SRC=0, but we should have the rpms being built.

14 years ago* fix a tyop
mjmac [Wed, 5 Apr 2006 02:02:36 +0000 (02:02 +0000)]
* fix a tyop
* add some sanity-checking

14 years agoNeed a bit more kernel prep fu. Thanx Cliff and bug 7108.
brian [Wed, 5 Apr 2006 00:30:27 +0000 (00:30 +0000)]
Need a bit more kernel prep fu.  Thanx Cliff and bug 7108.
Need to be able to tell the lustre.spec which kernel sources to build against.
lmake "cheats" and re-substs the lustre.spec.in template file.

14 years agoOf course, we need to work outside of $DO_SRC == 1.
brian [Tue, 4 Apr 2006 22:33:32 +0000 (22:33 +0000)]
Of course, we need to work outside of $DO_SRC == 1.

14 years agoAhhh. For patchless client, we still need the kernel source unpacked and
brian [Tue, 4 Apr 2006 22:02:15 +0000 (22:02 +0000)]
Ahhh.  For patchless client, we still need the kernel source unpacked and
configure needs to know where it is.

14 years agoA lack of a SERIES file specified in a target file means that we don't have
brian [Tue, 4 Apr 2006 21:04:12 +0000 (21:04 +0000)]
A lack of a SERIES file specified in a target file means that we don't have
patches for a kernel, so why build one.  Specifically this is for building a
patchless client.
This should all work if the concept of:
if [ -z "$LINUX" ] ; then
...
else
    build_lustre
fi
works.  I don't know that we test the build_lustre() function of lbuild with
any regularity, if at all though.

14 years agob_release_1_4_6-patchless is now using lustre-build from HEAD. Happy hacking.
brian [Tue, 4 Apr 2006 20:47:45 +0000 (20:47 +0000)]
b_release_1_4_6-patchless is now using lustre-build from HEAD.  Happy hacking.

14 years agor=green
brian [Tue, 4 Apr 2006 20:45:54 +0000 (20:45 +0000)]
r=green
b_release_1_4_6-patchless is now using lustre-build from HEAD.  Happy hacking.

14 years agoRemoved per Peter Bojanic. Will return if the code can be fixed and tested.
cliffw [Mon, 3 Apr 2006 15:55:06 +0000 (15:55 +0000)]
Removed per Peter Bojanic. Will return if the code can be fixed and tested.

14 years agob=10371
cliffw [Fri, 31 Mar 2006 16:53:37 +0000 (16:53 +0000)]
b=10371
r=scjody@clusterfs.com

Fixed error with multiple scsidevs

14 years agob=10329
zhanghc [Wed, 29 Mar 2006 10:56:02 +0000 (10:56 +0000)]
b=10329
DESCRIPTION:
    When doing fcntl() file locking calls in a liblustre client via libsysio,
successful lock calls return random non-zero values due to use uninitialized
value as return value.

14 years agob=10163
zhanghc [Wed, 29 Mar 2006 06:24:50 +0000 (06:24 +0000)]
b=10163
DESCRIPTION:
    keep CFS up to date with latest libsysio version

14 years agoBranch: HEAD
wangdi [Tue, 28 Mar 2006 02:57:14 +0000 (02:57 +0000)]
Branch: HEAD
add b_new_cmd in buildcvs

14 years agoTrickery to get the exit status of strace's subject.
brian [Fri, 24 Mar 2006 21:05:36 +0000 (21:05 +0000)]
Trickery to get the exit status of strace's subject.

14 years agostrace that update_oldconfig to see why make oldconfig is timing out the expect script.
brian [Fri, 24 Mar 2006 18:06:20 +0000 (18:06 +0000)]
strace that update_oldconfig to see why make oldconfig is timing out the expect script.

14 years agofix shebang -d problem
wangyb [Fri, 24 Mar 2006 04:44:37 +0000 (04:44 +0000)]
fix shebang -d problem

14 years agoAdded support for raw devices
cliffw [Thu, 23 Mar 2006 22:21:33 +0000 (22:21 +0000)]
Added support for raw devices

14 years agoGet some diagnostic output from expect to see if I can see why this fails on
brian [Thu, 23 Mar 2006 22:10:29 +0000 (22:10 +0000)]
Get some diagnostic output from expect to see if I can see why this fails on
RHEL3 on only boulder.  Attempts to reproduce in a debugging environment are
fruitless.  ~sigh~

14 years agoMerge with old CVS version; this should now contain the best of old CVS and 20051107...
scjody [Thu, 23 Mar 2006 20:58:04 +0000 (20:58 +0000)]
Merge ... old CVS version; this should now contain the best of old CVS and 20051107 from the website.

14 years agor=adilger
brian [Tue, 21 Mar 2006 20:20:29 +0000 (20:20 +0000)]
r=adilger
Try to find the cvs-modified-files.pl in the same place as the cvsdiffclient
if it's not in $PATH.

14 years agoBe a bit more verbose than just "timeout".
brian [Tue, 21 Mar 2006 16:19:32 +0000 (16:19 +0000)]
Be a bit more verbose than just "timeout".

14 years agoYet another expression found in the RHEL3/x86_64 make oldconfig.
brian [Tue, 21 Mar 2006 15:55:51 +0000 (15:55 +0000)]
Yet another expression found in the RHEL3/x86_64 make oldconfig.

14 years agoAnother different make config pattern. Where are these all coming from all
brian [Tue, 21 Mar 2006 06:56:44 +0000 (06:56 +0000)]
Another different make config pattern.  Where are these all coming from all
of a sudden?

14 years agoPut the return status in a variable, and use the PIPESTATUS array instead of
brian [Mon, 20 Mar 2006 21:11:25 +0000 (21:11 +0000)]
Put the return status in a variable, and use the PIPESTATUS array instead of
the easily broken $?.

14 years agoRecognize and answer an additional expression.
brian [Mon, 20 Mar 2006 18:48:33 +0000 (18:48 +0000)]
Recognize and answer an additional expression.
Be a bit more liberal with the format of the questions.