proper wic integration

proper wic integration

[Henning Schild]

Hi,

i started looking into intetgrating wic into the build process. And i
have a first working version of that code.
On the way i found that Isar is using a massively messed-with version
of wic. In my point of view that is an absolute show-blocker for any
code that builds on top of that.
Before anything gets done with wic, we need to revert all patches that
touch wic. We should not have a fork of wic but use an unmodified
version of it.

So this mail is just to let you know that i am looking into that.
A first working commit can be found here:
https://github.com/henning-schild-work/isar/commit/eeba383177a95bf8b03cba70e1525f65813b4df5

To get that merged i will first move to an untouched version of wic and
rebase it on top of that.

Henning

Re: proper wic integration

[Baurzhan Ismagulov]

On Fri, Jan 19, 2018 at 01:18:51PM +0100, Henning Schild wrote:
> i started looking into intetgrating wic into the build process. And i
> have a first working version of that code.
> On the way i found that Isar is using a massively messed-with version
> of wic. In my point of view that is an absolute show-blocker for any
> code that builds on top of that.

It depends on your priorities. Our goal at ilbers is to give people working
code first, merge with upstream second.


> Before anything gets done with wic, we need to revert all patches that
> touch wic. We should not have a fork of wic but use an unmodified
> version of it.

I agree with the latter. If you come up with an implementation that does it
without modifying wic, we'd gladly review it. Before that, nothing will be
reverted in master.


> So this mail is just to let you know that i am looking into that.

Thanks, this is important to know.


> A first working commit can be found here:
> https://github.com/henning-schild-work/isar/commit/eeba383177a95bf8b03cba70e1525f65813b4df5

The commit comment contains dirty words. Please use respectful expressions and
tone.


> To get that merged i will first move to an untouched version of wic and
> rebase it on top of that.

That is fine with me.


With kind regards,
-- 
Baurzhan Ismagulov
ilbers GmbH
Baierbrunner Str. 28c
D-81379 M�nchen
+49 (89) 122 67 24-0
ibr@ilbers.de
http://ilbers.de/
Amtsgericht M�nchen, HRB 214197
Gesch�ftsf�hrer: Baurzhan Ismagulov

Re: proper wic integration

[Henning Schild]

Am Fri, 19 Jan 2018 15:02:10 +0100
schrieb Baurzhan Ismagulov <ibr@radix50.net>:

> On Fri, Jan 19, 2018 at 01:18:51PM +0100, Henning Schild wrote:
> > i started looking into intetgrating wic into the build process. And
> > i have a first working version of that code.
> > On the way i found that Isar is using a massively messed-with
> > version of wic. In my point of view that is an absolute
> > show-blocker for any code that builds on top of that.  
> 
> It depends on your priorities. Our goal at ilbers is to give people
> working code first, merge with upstream second.
> 
> 
> > Before anything gets done with wic, we need to revert all patches
> > that touch wic. We should not have a fork of wic but use an
> > unmodified version of it.  
> 
> I agree with the latter. If you come up with an implementation that
> does it without modifying wic, we'd gladly review it. Before that,
> nothing will be reverted in master.

Of cause not, i will send the reverts with the working code.

> > So this mail is just to let you know that i am looking into that.  
> 
> Thanks, this is important to know.

That is why i shared an early version of my work.
 
> > A first working commit can be found here:
> > https://github.com/henning-schild-work/isar/commit/eeba383177a95bf8b03cba70e1525f65813b4df5  
> 
> The commit comment contains dirty words. Please use respectful
> expressions and tone.

Yes sorry about that, that where comments that where not supposed to
leave my local branch.

> > To get that merged i will first move to an untouched version of wic
> > and rebase it on top of that.  
> 
> That is fine with me.

I came up with something that comes very close, only one tiny bit
missing with fsck failing for some reason.

https://github.com/henning-schild-work/isar/commits/henning/staging

Henning
 
> 
> With kind regards,

Re: proper wic integration

[Baurzhan Ismagulov]

On Fri, Jan 19, 2018 at 01:18:51PM +0100, Henning Schild wrote:
> To get that merged i will first move to an untouched version of wic and
> rebase it on top of that.

I'd strongly suggest that you clean up this version so that it is merged and
Christian can use it, and refactor wic in a separate series.

With kind regards,
Baurzhan.

Re: proper wic integration

[Jan Kiszka]

On 2018-01-23 12:54, Baurzhan Ismagulov wrote:
> On Fri, Jan 19, 2018 at 01:18:51PM +0100, Henning Schild wrote:
>> To get that merged i will first move to an untouched version of wic and
>> rebase it on top of that.
> 
> I'd strongly suggest that you clean up this version so that it is merged and
> Christian can use it, and refactor wic in a separate series.

Patches are there, they surely need some cleanups, but I don't see your
point in reordering.

In the end, the effort to remove the upstream wic modification wasn't
that excessive as we feared. In fact, it would have been probably
simpler to go that way in the beginning.

But Henning is on vacation this week, so don't expect an update before
mid of next week.

Jan

Re: proper wic integration

[Baurzhan Ismagulov]

On Tue, Jan 23, 2018 at 01:46:25PM +0100, Jan Kiszka wrote:
> >> To get that merged i will first move to an untouched version of wic and
> >> rebase it on top of that.
> > 
> > I'd strongly suggest that you clean up this version so that it is merged and
> > Christian can use it, and refactor wic in a separate series.
> 
> Patches are there, they surely need some cleanups, but I don't see your
> point in reordering.

As discussed offline, looking from outside: 1. Wic build integration is
finished, 2. wic refactoring is going to start. So, looking from outside,
providing 2, then 1 is reordering, not the other way around :) . The point is,
as usual, moving in small steps. But let's see Henning's motivation for that.

With kind regards,
Baurzhan.