factory-utils: add factory scripts from crosutils (copy only)

factory-utils: add factory scripts from crosutils (copy only)

BUG=n0ne
TEST=

Committed: http://chrome-svn/viewvc/chromeos?view=rev&revision=b67893d

[Nick Sanders, 2011-02-04 19:55:13]

http://codereview.chromium.org/6368069/diff/1/mk_memento_images.sh
File mk_memento_images.sh (right):

http://codereview.chromium.org/6368069/diff/1/mk_memento_images.sh#newcode34
mk_memento_images.sh:34: . "${SCRIPT_ROOT}/common.sh" || (echo "Unable to load
common.sh" && exit 1)
Why is it here twice?

[zbehan, 2011-02-04 20:09:37]

It's twice in the original, I suspect after greg's huge CL. I merely made a
copy. I don't think refactoring is in place at this point, but I can fix it
if you want.

On Fri, Feb 4, 2011 at 8:55 PM, <nsanders@chromium.org> wrote:

>
> http://codereview.chromium.org/6368069/diff/1/mk_memento_images.sh
> File mk_memento_images.sh (right):
>
>
> http://codereview.chromium.org/6368069/diff/1/mk_memento_images.sh#newcode34
> mk_memento_images.sh:34: . "${SCRIPT_ROOT}/common.sh" || (echo "Unable
> to load common.sh" && exit 1)
> Why is it here twice?
>
>
> http://codereview.chromium.org/6368069/
>

[Nick Sanders, 2011-02-04 21:29:27]

lgtm if you remove the dup line

[Nick Sanders, 2011-02-05 22:08:40]

lgtm to move as is

[zbehan, 2011-02-05 23:15:51]

OK, I'm actually considering a little modification. I'd rename your copy of
the script to something like mk_memento_images_factory.sh (and update the
appropriate calls).

Reasoning behind this is, many people use mk_memento_images.sh, but factory,
as usually, calls it in a different way. Namely, it's supposed to be
exclusively called via cros_create_update_payload (may not be 100% precise
name), and while everyone else does this, factory doesn't. This calls for a
fork of the script, so that factory remains entirely independent on
everything else (it already is), and let me worry about the autoupdate stuff
separately, in a different CL.

See http://codereview.chromium.org/6349060/ for issues with moving this
particular file.

The whole thing is going to be needing a refactoring. It may be that we'll
manage to switch over everything else into the new au update format and get
rid of the script altogether. But I'd like to be able to treat those as
separate issues.

Thoughts?

On Sat, Feb 5, 2011 at 11:08 PM, <nsanders@chromium.org> wrote:

> lgtm to move as is
>
>
> http://codereview.chromium.org/6368069/
>

[anush, 2011-02-05 23:20:38]

My suggestion would be to move as is. And file bugs for refactoring
and come back to it. Trying to refactor/rewrite everything this
quarter may drag the migration of src/scripts more than necessary.
But if you think it is easy to do without being a time sink go for it.

On Sat, Feb 5, 2011 at 3:15 PM, Zdenek Behan <zbehan@chromium.org> wrote:
> OK, I'm actually considering a little modification. I'd rename your copy of
> the script to something like mk_memento_images_factory.sh (and update the
> appropriate calls).
> Reasoning behind this is, many people use mk_memento_images.sh, but factory,
> as usually, calls it in a different way. Namely, it's supposed to be
> exclusively called via cros_create_update_payload (may not be 100% precise
> name), and while everyone else does this, factory doesn't. This calls for a
> fork of the script, so that factory remains entirely independent on
> everything else (it already is), and let me worry about the autoupdate stuff
> separately, in a different CL.
> See http://codereview.chromium.org/6349060/ for issues with moving this
> particular file.
> The whole thing is going to be needing a refactoring. It may be that we'll
> manage to switch over everything else into the new au update format and get
> rid of the script altogether. But I'd like to be able to treat those as
> separate issues.
> Thoughts?
>
> On Sat, Feb 5, 2011 at 11:08 PM, <nsanders@chromium.org> wrote:
>>
>> lgtm to move as is
>>
>> http://codereview.chromium.org/6368069/
>
>

[zbehan, 2011-02-05 23:23:07]

At this point, I'm only wanting to do the fork, to free my hands from the
various consequences of moving the script. Refactoring later.

2011/2/6 Anush Elangovan(அனுஷ்) <anush@chromium.org>

> My suggestion would be to move as is. And file bugs for refactoring
> and come back to it. Trying to refactor/rewrite everything this
> quarter may drag the migration of src/scripts more than necessary.
> But if you think it is easy to do without being a time sink go for it.
>
> On Sat, Feb 5, 2011 at 3:15 PM, Zdenek Behan <zbehan@chromium.org> wrote:
> > OK, I'm actually considering a little modification. I'd rename your copy
> of
> > the script to something like mk_memento_images_factory.sh (and update the
> > appropriate calls).
> > Reasoning behind this is, many people use mk_memento_images.sh, but
> factory,
> > as usually, calls it in a different way. Namely, it's supposed to be
> > exclusively called via cros_create_update_payload (may not be 100%
> precise
> > name), and while everyone else does this, factory doesn't. This calls for
> a
> > fork of the script, so that factory remains entirely independent on
> > everything else (it already is), and let me worry about the autoupdate
> stuff
> > separately, in a different CL.
> > See http://codereview.chromium.org/6349060/ for issues with moving this
> > particular file.
> > The whole thing is going to be needing a refactoring. It may be that
> we'll
> > manage to switch over everything else into the new au update format and
> get
> > rid of the script altogether. But I'd like to be able to treat those as
> > separate issues.
> > Thoughts?
> >
> > On Sat, Feb 5, 2011 at 11:08 PM, <nsanders@chromium.org> wrote:
> >>
> >> lgtm to move as is
> >>
> >> http://codereview.chromium.org/6368069/
> >
> >
>

[anush, 2011-02-05 23:24:30]

thats fine with me if it is ok with nsanders.

On Sat, Feb 5, 2011 at 3:23 PM, Zdenek Behan <zbehan@chromium.org> wrote:
> At this point, I'm only wanting to do the fork, to free my hands from the
> various consequences of moving the script. Refactoring later.
>
> 2011/2/6 Anush Elangovan(அனுஷ்) <anush@chromium.org>
>>
>> My suggestion would be to move as is. And file bugs for refactoring
>> and come back to it. Trying to refactor/rewrite everything this
>> quarter may drag the migration of src/scripts more than necessary.
>> But if you think it is easy to do without being a time sink go for it.
>>
>> On Sat, Feb 5, 2011 at 3:15 PM, Zdenek Behan <zbehan@chromium.org> wrote:
>> > OK, I'm actually considering a little modification. I'd rename your copy
>> > of
>> > the script to something like mk_memento_images_factory.sh (and update
>> > the
>> > appropriate calls).
>> > Reasoning behind this is, many people use mk_memento_images.sh, but
>> > factory,
>> > as usually, calls it in a different way. Namely, it's supposed to be
>> > exclusively called via cros_create_update_payload (may not be 100%
>> > precise
>> > name), and while everyone else does this, factory doesn't. This calls
>> > for a
>> > fork of the script, so that factory remains entirely independent on
>> > everything else (it already is), and let me worry about the autoupdate
>> > stuff
>> > separately, in a different CL.
>> > See http://codereview.chromium.org/6349060/ for issues with moving this
>> > particular file.
>> > The whole thing is going to be needing a refactoring. It may be that
>> > we'll
>> > manage to switch over everything else into the new au update format and
>> > get
>> > rid of the script altogether. But I'd like to be able to treat those as
>> > separate issues.
>> > Thoughts?
>> >
>> > On Sat, Feb 5, 2011 at 11:08 PM, <nsanders@chromium.org> wrote:
>> >>
>> >> lgtm to move as is
>> >>
>> >> http://codereview.chromium.org/6368069/
>> >
>> >
>
>

[Nick Sanders, 2011-02-05 23:26:34]

Agree, refactors and moves should be separate.

On Sat, Feb 5, 2011 at 3:15 PM, Zdenek Behan <zbehan@chromium.org> wrote:

>  > OK, I'm actually considering a little modification. I'd rename your copy
> of
> > the script to something like mk_memento_images_factory.sh (and update the
> > appropriate calls).
> > Reasoning behind this is, many people use mk_memento_images.sh, but
> factory,
> > as usually, calls it in a different way. Namely, it's supposed to be
>
> exclusively called via cros_create_update_payload
>

Nobody should be calling make_memento_images other than factory, calling it
from
cros_create_update_payload is not supported. The comment in the code
is incorrect.


> (may not be 100% precise
> > name), and while everyone else does this, factory doesn't. This calls for
> a
> > fork of the script, so that factory remains entirely independent on
> > everything else (it already is), and let me worry about the autoupdate
> stuff
> > separately, in a different CL.
> > See http://codereview.chromium.org/6349060/ for issues with moving this
> > particular file.
> > The whole thing is going to be needing a refactoring. It may be that
> we'll
> > manage to switch over everything else into the new au update format and
> get
> > rid of the script altogether. But I'd like to be able to treat those as
> > separate issues.
> > Thoughts?
> >
> > On Sat, Feb 5, 2011 at 11:08 PM, <nsanders@chromium.org> wrote:
> >>
> >> lgtm to move as is
> >>
> >> http://codereview.chromium.org/6368069/
> >
> >
>

[Nick Sanders, 2011-02-05 23:27:18]

sgtm.

On Sat, Feb 5, 2011 at 3:24 PM, Anush Elangovan(அனுஷ்)
<anush@chromium.org>wrote:

> thats fine with me if it is ok with nsanders.
>
> On Sat, Feb 5, 2011 at 3:23 PM, Zdenek Behan <zbehan@chromium.org> wrote:
> > At this point, I'm only wanting to do the fork, to free my hands from the
> > various consequences of moving the script. Refactoring later.
> >
> > 2011/2/6 Anush Elangovan(அனுஷ்) <anush@chromium.org>
> >>
> >> My suggestion would be to move as is. And file bugs for refactoring
> >> and come back to it. Trying to refactor/rewrite everything this
> >> quarter may drag the migration of src/scripts more than necessary.
> >> But if you think it is easy to do without being a time sink go for it.
> >>
> >> On Sat, Feb 5, 2011 at 3:15 PM, Zdenek Behan <zbehan@chromium.org>
> wrote:
> >> > OK, I'm actually considering a little modification. I'd rename your
> copy
> >> > of
> >> > the script to something like mk_memento_images_factory.sh (and update
> >> > the
> >> > appropriate calls).
> >> > Reasoning behind this is, many people use mk_memento_images.sh, but
> >> > factory,
> >> > as usually, calls it in a different way. Namely, it's supposed to be
> >> > exclusively called via cros_create_update_payload (may not be 100%
> >> > precise
> >> > name), and while everyone else does this, factory doesn't. This calls
> >> > for a
> >> > fork of the script, so that factory remains entirely independent on
> >> > everything else (it already is), and let me worry about the autoupdate
> >> > stuff
> >> > separately, in a different CL.
> >> > See http://codereview.chromium.org/6349060/ for issues with moving
> this
> >> > particular file.
> >> > The whole thing is going to be needing a refactoring. It may be that
> >> > we'll
> >> > manage to switch over everything else into the new au update format
> and
> >> > get
> >> > rid of the script altogether. But I'd like to be able to treat those
> as
> >> > separate issues.
> >> > Thoughts?
> >> >
> >> > On Sat, Feb 5, 2011 at 11:08 PM, <nsanders@chromium.org> wrote:
> >> >>
> >> >> lgtm to move as is
> >> >>
> >> >> http://codereview.chromium.org/6368069/
> >> >
> >> >
> >
> >
>

[zbehan, 2011-02-05 23:29:53]

2011/2/6 Nick Sanders <nsanders@chromium.org>

> Agree, refactors and moves should be separate.
>
> On Sat, Feb 5, 2011 at 3:15 PM, Zdenek Behan <zbehan@chromium.org> wrote:
>
>>  > OK, I'm actually considering a little modification. I'd rename your
>> copy of
>> > the script to something like mk_memento_images_factory.sh (and update
>> the
>> > appropriate calls).
>> > Reasoning behind this is, many people use mk_memento_images.sh, but
>> factory,
>> > as usually, calls it in a different way. Namely, it's supposed to be
>>
> > exclusively called via cros_create_update_payload
>>
>
> Nobody should be calling make_memento_images other than factory, calling it
> from
> cros_create_update_payload is not supported. The comment in the code
> is incorrect.
>

Unfortunately, they are using it. My CL broke workflow for some people, in
particular the test people, and who knows who else.


> (may not be 100% precise
>> > name), and while everyone else does this, factory doesn't. This calls
>> for a
>> > fork of the script, so that factory remains entirely independent on
>> > everything else (it already is), and let me worry about the autoupdate
>> stuff
>> > separately, in a different CL.
>> > See http://codereview.chromium.org/6349060/ for issues with moving this
>> > particular file.
>> > The whole thing is going to be needing a refactoring. It may be that
>> we'll
>> > manage to switch over everything else into the new au update format and
>> get
>> > rid of the script altogether. But I'd like to be able to treat those as
>> > separate issues.
>> > Thoughts?
>> >
>> > On Sat, Feb 5, 2011 at 11:08 PM, <nsanders@chromium.org> wrote:
>> >>
>> >> lgtm to move as is
>> >>
>> >> http://codereview.chromium.org/6368069/
>> >
>> >
>>
>
>

[zbehan, 2011-02-06 03:25:11]

mk_memento_images renamed, and reference changed. I also removed the
duplicate line.

Also, this CL intends to "move" the scripts, with as few modifications as
possible. After I introduce the ebuild to install these, I'll submit another
subsequent CL to fix all necessary paths and test things if they work. Then
I'll remove these from crosutils.

On Sun, Feb 6, 2011 at 4:22 AM, <zbehan@chromium.org> wrote:

> http://codereview.chromium.org/6368069/
>