Bug 104868

Please attach the log files.

Presence of windows partition is checked with this call:

if (size (Storage::GetWinPrimPartitions(Storage::GetTargetMap()))>0)
...

Thomas?

Not possible to say anything useful without log files.

BTW: I had a look into a log file of my installation (where I created a
windows primary partition to verify) and I could not see any call to
Storage::GetWinPrimPartitions at all. 
Are you sure this is called. If yes, when should it get called. I assume
necessarily before the timezone proposal but I did not see a log entry for 
such a call.

It's in "MakeProposal" call of timezone_proposal.ycp

Can you see a line with GetWinPrimPartitions in your logfile?
I cannot find one here even after timezone proposal is finished.

In my logs from beta1 installation, I can see it:

2005-08-08 15:47:35 <1> linux(2992) [Interpreter] clients/inst_proposal.ycp:67
Calling YaST client timezone_proposal (arguments: ["MakeProposal",
$["force_reset":false, "language_changed":false]])
2005-08-08 15:47:35 <1> linux(2992) [YCP] clients/timezone_proposal.ycp:28 date
call: $["exit":0, "stderr":"", "stdout":"2005\n"]

And now some calls from Storage begin:

2005-08-08 15:47:35,410 INFO  libstorage - Storage.cc(getFreeInfo):3531
device:/dev/hda1


But we're looking for the Ulrich's logs.

Created attachment 46261 [details]
Archive with YaST's logs

The log contains entries like:

Storage.ycp:3163 GetWinPrimPartitions num_linux 0 num_win 1 num_dos 0 num_os2 0
Storage.ycp:3217 GetWinPrimPartitions ret [$["device":"/dev/hda1",
"string":"windows"]]
clients/timezone_proposal.ycp:46 windows partition found: assuming local time

So the detection of the windows partition in yast2-storage seems to work fine.

And this log says what is shown in proposal screen:

"Europa / Deutschland - Hardware Clock Set To Local Time 15:00:10 - 14-08-2005"

Ulrich, you either provided wrong logfiles or the behaviour is correct.

AFAIR I manually changed the timezone for the RTC. Can you see that from the
logs?  I've been updating installing so many times recently that I could have
mixed things, but I don't think so.

So what are you reporting?

You said "installation suggests the RTC runs UTC instead of local time"; this is
not true, according to the logs.

Created attachment 46803 [details]
Logs taken from the installation system

Even on a second installationm attempt (which had been aborted), the suggestion
was to use UTC for the RTC.

now I see it: inst_timezone is started before timezone_proposal.ycp

Thomas, is it possible to call GetWinPrimPartitions before the proposal? Do I
get expected result?

(In current installation workflow, timezone setting is before creating the
proposal.)

Yes, it should not matter when you call GetWinPrimPartitions(), to be on the
save side you can just check the debug output. It is also printed if no
Windows partitions are detected.

OK, fixed.

Ulrich, please test it as soon as you'll get beta3.

The proposal is correct in beta3, however the time displayed is wrong (timezone
correction applied to localtime). However after installation the clock seems
correct. Now if the user would "correct" the time during installation (because
of the wrong display), the time would be wrong at the end.
I know that this bug as a very long tradition is SuSE Linux, but that's no
reason not to fix it some day.

What is displayd wrongly when proposal is correct?
Does it mean that only proposal for using local time is correct?

I think it should be meanwhile fixed with the fix of bug #112900

(In reply to comment #19)
> What is displayd wrongly when proposal is correct?

It is correctly proposed that the RTC runs in local time, but the time displayed
is the local time plus the timezone offset (2 hours for MESZ). So the time
displayed is two hours ahead. (It would be correct if RTC were running UTC).