headers
Kenneth Crudup | 17 Jul 19:06

So the ToI GIT isn't guaranteed to build, then?


I just pulled the latest GIT last night (about 18:30 local Pacific USA
time) and this morning (Thu Jul 17 07:04:13 Nigel's time) and it doesn't
compile:

----
  CC      kernel/power/tuxonice_prepare_image.o
kernel/power/tuxonice_prepare_image.c: In function ‘get_extra_pd1_allowance’:
kernel/power/tuxonice_prepare_image.c:376: error: too few arguments to function ‘device_power_up’
kernel/power/tuxonice_prepare_image.c:378: error: too few arguments to function ‘device_resume’
make[2]: *** [kernel/power/tuxonice_prepare_image.o] Error 1
make[1]: *** [kernel/power] Error 2
make: *** [kernel] Error 2
---

I'll fix it, but I was just surprised this happened.

	-Kenny

--

-- 
Kenneth R. Crudup  Sr. SW Engineer, Scott County Consulting, Los Angeles
O: 3630 S. Sepulveda Blvd. #138, L.A., CA 90034-6809      (888) 454-8181
_______________________________________________
TuxOnIce-devel mailing list
TuxOnIce-devel <at> lists.tuxonice.net
http://lists.tuxonice.net/mailman/listinfo/tuxonice-devel
Permalink | Reply |
headers
Kenneth Crudup | 17 Jul 19:27

Re: So the ToI GIT isn't guaranteed to build, then?


> I just pulled the latest GIT last night (about 18:30 local Pacific USA
> time) and this morning (Thu Jul 17 07:04:13 Nigel's time) and it doesn't
> compile:

Some info:

$ git status
# On branch tuxonice
nothing to commit (working directory clean)
$

	-Kenny

--

-- 
Kenneth R. Crudup  Sr. SW Engineer, Scott County Consulting, Los Angeles
O: 3630 S. Sepulveda Blvd. #138, L.A., CA 90034-6809      (888) 454-8181
Permalink | Reply |
headers
Nigel Cunningham | 17 Jul 23:23
Favicon

Re: So the ToI GIT isn't guaranteed to build, then?

Hi.

On Thu, 2008-07-17 at 10:07 -0700, Kenneth Crudup wrote:
> I just pulled the latest GIT last night (about 18:30 local Pacific USA
> time) and this morning (Thu Jul 17 07:04:13 Nigel's time) and it doesn't
> compile:
> 
> ----
>   CC      kernel/power/tuxonice_prepare_image.o
> kernel/power/tuxonice_prepare_image.c: In function ‘get_extra_pd1_allowance’:
> kernel/power/tuxonice_prepare_image.c:376: error: too few arguments to function ‘device_power_up’
> kernel/power/tuxonice_prepare_image.c:378: error: too few arguments to function ‘device_resume’
> make[2]: *** [kernel/power/tuxonice_prepare_image.o] Error 1
> make[1]: *** [kernel/power] Error 2
> make: *** [kernel] Error 2
> ---
> 
> I'll fix it, but I was just surprised this happened.

Please remember that I'm doing this in whatever 'spare' time I can
make.

It almost always will, but I didn't get enough time to test these
changes yesterday, and my script pushed the merge result automatically
because it's normally okay to do so, so I made that the default.

Nigel

_______________________________________________
TuxOnIce-devel mailing list
(Continue reading)

Permalink | Reply |
headers
Kenneth Crudup | 17 Jul 23:31

Re: So the ToI GIT isn't guaranteed to build, then?


On Fri, 18 Jul 2008, Nigel Cunningham wrote:

> Please remember that I'm doing this in whatever 'spare' time I can make.

Sure, no worries (well, one- see below). I didn't know this part:

> and my script pushed the merge result automatically

Ah. I'd wanted to ask you about that- 2.6.26 was available for GITting
nearly as soon as Linus et al. had pushed it- do you have scripts that
also automatically do <whatever does merges in git>, as well?

Oh, and my big concern was while I've figured out what to do in cases
like this ("git checkout", edit, then- and this was the scary part-
"git commit", 'cause I didn't want my private changes polluting the
upstream trees by mistake!) I don't know how to do the equivalent of
"revert" or "unedit" yet in GIT, so when you did finally make the fix
I wanted to pull out my version.

(BTW, I ended up having to wait for your fix, anyway- I couldn't decide
from a precursory inspection which pm_message_t enum was appropriate for
that situation :)

	-Kenny

--

-- 
Kenneth R. Crudup  Sr. SW Engineer, Scott County Consulting, Los Angeles
O: 3630 S. Sepulveda Blvd. #138, L.A., CA 90034-6809      (888) 454-8181
(Continue reading)

Permalink | Reply |
headers
Nigel Cunningham | 18 Jul 00:38
Favicon

Re: So the ToI GIT isn't guaranteed to build, then?

Hi.

On Thu, 2008-07-17 at 14:31 -0700, Kenneth Crudup wrote:
> On Fri, 18 Jul 2008, Nigel Cunningham wrote:
> 
> > Please remember that I'm doing this in whatever 'spare' time I can make.
> 
> Sure, no worries (well, one- see below). I didn't know this part:
> 
> > and my script pushed the merge result automatically
> 
> Ah. I'd wanted to ask you about that- 2.6.26 was available for GITting
> nearly as soon as Linus et al. had pushed it- do you have scripts that
> also automatically do <whatever does merges in git>, as well?

I usually only run the script once a day at most, but was running it
more often yesterday because I wanted to make sure I wasn't going to
cause problems for Stephen with linux-next. Unfortunately that meant I
did merge stuff I didn't want and did end up saying "Let's skip it for
today" to him.

The script does (in essence - we bail if any step fails)

git checkout origin (upstream branch copy)
git pull (get upstream changes - sometimes I have to control-c here
because Git thinks there are no commits in common when there are. It
gets it right the second time. I guess this is a bug in Ubuntu's version
that's been fixed in upstream)
git checkout tuxonice (switch to tuxonice branch)
git merge origin (merge changes to origin)
(Continue reading)

Permalink | Reply |
headers
Kenneth Crudup | 18 Jul 00:45

Re: So the ToI GIT isn't guaranteed to build, then?

On Fri, 18 Jul 2008, Nigel Cunningham wrote:

> To avoid 'polluting' your copy of the upstream tree, I'd do:

I'm *really* going to have to find a GIT primer somewhere! But, as you said:

> I'm not sure how long I'll be - I have other work to do that's higher priority.

Yeah, M3 T00!! (Besides, I'm quite happy with the build as of yesterday
about noon USA PDT- everything works well- I just wish I'd snapshotted it!)

	-Kenny

--

-- 
Kenneth R. Crudup  Sr. SW Engineer, Scott County Consulting, Los Angeles
O: 3630 S. Sepulveda Blvd. #138, L.A., CA 90034-6809      (888) 454-8181
Permalink | Reply |
headers
Nigel Cunningham | 18 Jul 01:54
Favicon

Re: So the ToI GIT isn't guaranteed to build, then?

Hi.

On Thu, 2008-07-17 at 15:45 -0700, Kenneth Crudup wrote:
> On Fri, 18 Jul 2008, Nigel Cunningham wrote:
> 
> > To avoid 'polluting' your copy of the upstream tree, I'd do:
> 
> I'm *really* going to have to find a GIT primer somewhere! But, as you said:
> 
> > I'm not sure how long I'll be - I have other work to do that's higher priority.
> 
> Yeah, M3 T00!! (Besides, I'm quite happy with the build as of yesterday
> about noon USA PDT- everything works well- I just wish I'd snapshotted it!)

Run gitk to see the history. That tells me that the last merge which
should work fine is 4e6d158c7ae90a6e01538704565469a8a2e96aed. So if you
do 

git branch good 4e6d158c7ae90a6e01538704565469a8a2e96aed
git checkout good

You should be on the commit prior to yesterday's merge of Rafael's
conflicting work.

Regards,

Nigel
Permalink | Reply |
headers
Kenneth Crudup | 18 Jul 02:18

Re: So the ToI GIT isn't guaranteed to build, then?


On Fri, 18 Jul 2008, Nigel Cunningham wrote:

> Run gitk to see the history. That tells me that the last merge which
> should work fine is 4e6d158c7ae90a6e01538704565469a8a2e96aed.

It appears to be the one before it; that one produces:

----
  CC      kernel/power/process.o
kernel/power/process.c: In function ‘thaw_kernel_threads’:
kernel/power/process.c:302: error: ‘FREEZER_KERNEL_THREADS’ undeclared (first use in this function)
kernel/power/process.c:302: error: (Each undeclared identifier is reported only once
kernel/power/process.c:302: error: for each function it appears in.)
make[2]: *** [kernel/power/process.o] Error 1
make[1]: *** [kernel/power] Error 2
make: *** [kernel] Error 2
----

... but now that I know how to find the history, I'll work back one and
get that one. Thanks!

	-Kenny, learning GIT one painful command at a time

--

-- 
Kenneth R. Crudup  Sr. SW Engineer, Scott County Consulting, Los Angeles
O: 3630 S. Sepulveda Blvd. #138, L.A., CA 90034-6809      (888) 454-8181
_______________________________________________
TuxOnIce-devel mailing list
TuxOnIce-devel <at> lists.tuxonice.net
(Continue reading)

Permalink | Reply |
headers
Kenneth Crudup | 18 Jul 02:26

Re: So the ToI GIT isn't guaranteed to build, then?


... but now that I know how to find the history, I'll work back one and
get that one. Thanks!

D-mn, that failed *spectacularly*! I ended up with a ToI-free kernel
somehow, and I can't interpret how/where to get the previous ToI-enabled
kernel from the graph in gitk ...

I'll wait for your release.

	-Kenny

--

-- 
Kenneth R. Crudup  Sr. SW Engineer, Scott County Consulting, Los Angeles
O: 3630 S. Sepulveda Blvd. #138, L.A., CA 90034-6809      (888) 454-8181
Permalink | Reply |
headers
Nigel Cunningham | 18 Jul 02:58
Favicon

Re: So the ToI GIT isn't guaranteed to build, then?

Hi.

On Thu, 2008-07-17 at 17:26 -0700, Kenneth Crudup wrote:
> ... but now that I know how to find the history, I'll work back one and
> get that one. Thanks!
> 
> D-mn, that failed *spectacularly*! I ended up with a ToI-free kernel
> somehow, and I can't interpret how/where to get the previous ToI-enabled
> kernel from the graph in gitk ...

When in gitk, you'll see the diff for the current commit in the bottom
left section. Near the top of that, you'll see clickable links to the
parent or parents of this node. If you use them, you should be right. To
know which parent (when there are more than one at merges), look at the
colours in the tree at the top. For example, on the last merge I did, I
see a red line on the far left that goes down to commit 4e6d158c. This
is why "Parent" beside that commit is red. If you click on 4e6d158c, it
will take you down to that previous commit. Go one more (to f6fc9dbd)
and I think you'll have the one you want. Hope that helps.

Nigel
Permalink | Reply |
headers
Kenneth Crudup | 19 Jul 03:37

Re: So the ToI GIT isn't guaranteed to build, then?


GIT's "Steep learning curve" is right- but- you're right about once you
have it, you appreciate it. Until you have time, I think I've figured
out how to get the "official" 2.6.26 tree, add in a few commits that
came after that.

	-Kenny

--

-- 
Kenneth R. Crudup  Sr. SW Engineer, Scott County Consulting, Los Angeles
O: 3630 S. Sepulveda Blvd. #138, L.A., CA 90034-6809      (888) 454-8181
Permalink | Reply |

Gmane