Could you please clarify what is the intended purpose of each of these changes?

No. I can guess what the code does by looking at it. I can not guess the intended purpose. In case when author does a mistake and writes something else than he wants to, I have only a commit message as a clue. In all three cases here, I guess the patch does what commit message describes.

(In reply to Frantisek Kluknavsky from comment #3)
> No. I can guess what the code does by looking at it. I can not guess the
> intended purpose. In case when author does a mistake and writes something
> else than he wants to, I have only a commit message as a clue. In all three
> cases here, I guess the patch does what commit message describes.

Well, indeed the commit messages in the sandbos repo are not useful: they describe the implementation (not sure if correctly, but that's not really my interest) but not the intent.  (Frankly I was expecting that owner of the bug would know best what is the intent, but you know, a man can dream...)

At this point, there is almost nothing QA can do.  From my POV this looks like the changes have no documented justification or purpose (ok maybe for one of them we could guess...).

(Peeking at the code, I don't even have any reason to believe the changes are really "minor"---not that I think such way of quantification would be safe.)

So if this is ever going to be "order for QA service", I can already see that the "order" is incomplete.  I hope you understand that I can't ack an incomplete order.

TL;DR: So please (somebody) clarify intent of the changes.  (If the answer is really "I/we don't know", then maybe the commits shouldn't have existed in the first place?)

Hi Alois, about my third patch - sed hasn't parameter "-E" for extended regexps as grep, but it has "-r" for this purpose.

First commit contains a default set of users provided by setup component. We need them for clarification what users are created by default on system. Users mentioned in that file are not included in kickstart, though.

Second is a bug which insert the comment to available-repos file (which is not needeed) and kickstart has a problem with parsing it.
Therefore I have remove it.

(In reply to pstodulk from comment #5)
> Hi Alois, about my third patch - sed hasn't parameter "-E" for extended
> regexps as grep, but it has "-r" for this purpose.

Yep, that's a justification why that is a bug, but does not answer my question:

How this change is supposed to affect user?

sed -E is not indeed an expected way how to use sed.  According to quick test it generates help text on stderr.  So two effects:

 *  help text that has been generated somewhere will disappear,

 *  and more importantly, list that has been empty until now, will contain
    items now.

What I'm asking is what kind of list is that, because I need to reproduce the case.  How can I see the effect of the test not being empty now?  (And possibly the effect of the error message not being seen?)

(In reply to Petr Hracek from comment #6)
> First commit contains a default set of users provided by setup component. We
> need them for clarification what users are created by default on system.
> Users mentioned in that file are not included in kickstart, though.

What do you mean by the clarification?  Is this list supposed to be displayed to user or used somehow else?

How can I reproduce and see the effect?

(In reply to Alois Mahdal from comment #7)
> (In reply to pstodulk from comment #5)
> > Hi Alois, about my third patch - sed hasn't parameter "-E" for extended
> > regexps as grep, but it has "-r" for this purpose.
> 
> Yep, that's a justification why that is a bug, but does not answer my
> question:
> 
> How this change is supposed to affect user?
> 
> sed -E is not indeed an expected way how to use sed.  According to quick
> test it generates help text on stderr.  So two effects:
> 
>  *  help text that has been generated somewhere will disappear,
> 
>  *  and more importantly, list that has been empty until now, will contain
>     items now.
> 
> What I'm asking is what kind of list is that, because I need to reproduce
> the case.  How can I see the effect of the test not being empty now?  (And
> possibly the effect of the error message not being seen?)

OK. I hope that I understand well. So exactly part of this change generates list of all installed packages, which is used to get all packages, which contains some utilities, which are mentioned inside "/root/preupgrade/RHEL6_7/common/$(arch)/Common_utilities". These utilities are printed in results, peculiarly with additional informations.

Petr, the first commit remains unclear--the first part of above comment (before tilde).

Clarified the first change with phracek:

The intended change is to deliver

 *  result from ReservedIDs,

 *  "old" intact /etc/passwd (as provided by 'setup' package)

into kickstart data.

Now if you call preupg with --kickstart parameter, the tarball (or something like that) must contain this data.

Verified with:
   preupgrade-assistant-2.1.1-5.el6.x86_64
   preupgrade-assistant-contents-0.6.39-1.el6.noarch
   redhat-upgrade-tool-0.7.41-1.el6.noarch

All fixes was verified manually. For the first one was created test case.

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://rhn.redhat.com/errata/RHBA-2015-2498.html