Hi Larry,
Is it possible to reset the (BK-)timestamps on the following files in the
http://linus.bkbits.net:8080/linux-2.5 repository? Somehow, they've gotten a
timestamp which lies in the future, causing lots of warnings when I use a bk
exported tree.
./drivers/base/class.c
./drivers/pci/hotplug/rpaphp_pci.c
./drivers/pci/hotplug/rpaphp_slot.c
./drivers/pci/hotplug/rpaphp_vio.c
./include/linux/kobject.h
./lib/kobject.c
Regards,
--
Jasper Spaans http://jsp.vs19.net/
10:06:49 up 9970 days, 53 min, 0 users, load average: 6.76 6.14 5.69
-... .- -.. --. . .-. -... .- -.. --. . .-. -... .- -.. --. . .-.
Jasper Spaans wrote:
> Hi Larry,
>
> Is it possible to reset the (BK-)timestamps on the following files in the
> http://linus.bkbits.net:8080/linux-2.5 repository? Somehow, they've gotten a
> timestamp which lies in the future, causing lots of warnings when I use a bk
> exported tree.
>
> ./drivers/base/class.c
> ./drivers/pci/hotplug/rpaphp_pci.c
> ./drivers/pci/hotplug/rpaphp_slot.c
> ./drivers/pci/hotplug/rpaphp_vio.c
> ./include/linux/kobject.h
> ./lib/kobject.c
Sounds like a local problem. I bet this will fix it:
ntpdate time.nist.gov
bk -r clean
bk -r co -q
The timestamps on checked-out files are not generated from any BitKeeper
metadata, but instead reflect the machine's system time when the file
was checked out.
Jeff
On Fri, Jun 04, 2004 at 10:19:26AM +0200, Jasper Spaans wrote:
> Is it possible to reset the (BK-)timestamps on the following files in the
> http://linus.bkbits.net:8080/linux-2.5 repository? Somehow, they've gotten a
> timestamp which lies in the future, causing lots of warnings when I use a bk
> exported tree.
Sorry, the timestamps are part of the BK metadata and there isn't any way
to alter them after the fact.
I'd suggest you not use the -T option to export or write a script that
looks for timestamps in the future and fix them.
For everyone else, please keep your clocks in sync. Whoever generated
these deltas must be running a very old version of BK (which is a license
violation by the way [whoohoo, we can now have 3 weeks of flames about
that horrible BK license that if people were obeying this wouldn't
have happened]). Current versions of BK insist that your clock isn't
off by more than 24 hours (I suspect that this works only if you have
a net connection though, I don't remember).
--
---
Larry McVoy lm at bitmover.com http://www.bitkeeper.com
On Fri, Jun 04, 2004 at 08:56:13AM -0700, Larry McVoy wrote:
> > Is it possible to reset the (BK-)timestamps on the following files in the
> > http://linus.bkbits.net:8080/linux-2.5 repository? Somehow, they've gotten a
> > timestamp which lies in the future, causing lots of warnings when I use a bk
> > exported tree.
>
> Sorry, the timestamps are part of the BK metadata and there isn't any way
> to alter them after the fact.
Ack. I just tried something with the following result:
---8<---
# This is a BitKeeper generated diff -Nru style patch.
#
# drivers/base/class.c
# 2004/06/04 19:55:06+02:00 [email protected] +0 -1
# Fix timestamp
#
# drivers/base/class.c
# 2004/06/04 19:54:56+02:00 [email protected] +1 -0
#
# ChangeSet
# 2004/06/04 19:55:32+02:00 [email protected]
# class.c:
# Fix timestamp
#
---8<---
Can anyone confirm that importing this does the trick? If so, I'll whip up
the 'patch' for all six files.
Regards,
--
Jasper Spaans http://jsp.vs19.net/
19:56:31 up 9970 days, 10:43, 0 users, load average: 6.19 6.08 5.67
-... .- -.. --. . .-. -... .- -.. --. . .-. -... .- -.. --. . .-.