torios-dev team mailing list archive

[Nio Wiklund <nio.wiklund@xxxxxxxxx>]

Den 2016-02-06 kl. 10:39, skrev Nio Wiklund:
> [bottom posting]
> 
> Den 2016-02-06 kl. 09:30, skrev Nio Wiklund:
>> Hi Israel,
>> [inline and bottom posting]
>> Best regards
>> Nio
>>
>> Den 2016-02-05 kl. 21:09, skrev Israel:
>>> Hi Jack,
>>> (inlines)
>>>
>>> On 02/05/2016 10:54 AM, Cinque Port Computers wrote:
>>>> Hi Israel,
>>>>
>>>> Installed into VB and observed the following; all noted previously, I
>>>> think.  Getting better.  :)
>>>>
>>>> 1) OBI progress box still in centre of screen being covered by other
>>>> pop-up boxes during install.
>>> For me this is not the case.
>>> The progress is on the bottom.
>>> I am using VB in 16.04
>>> Not sure what is causing this issue for you... :(
>>
>> In the previous version I had this problem with the progress window, but
>> it works correctly for me in this version. I'm testing in my Toshiba (so
>> a real computer).
>>
>>>> 2) Tick box to confirm final media to install to (pop-up with all the
>>>> red in) is not visible but present (I believe Nio has mentioned this
>>>> before).
>>> True, this is either GTK or Numix bug.  I have not gotten any response
>>> on G+ I will try on github
>>
>> The modified texts helps getting around this bug.
>>
>>>> 3) Tick area for setting keyboard map once installed (in my case from
>>>> 'us' to 'gb') is not visible but present.
>>> But you now can configure your locale, and does it work persistently??
>>
>> Yes, I can use the new feature to configure the locale, and it works,
>> but it is very slow to scroll to Swedish with the down arrow. I found no
>> quick scroll or hotkey method (page down helped a little). Please advice
>> to to manage this window in an efficient way!
>>
>>>>
>>>> Cheers,
>>>> JackT.
>>>>
>>> There is a bug in firstrun, that should be fixed in the upstream.  I
>>> will wait to rebuild the baseOS again until we have all the kinks out,
>>> as using export XZ_OPT=${XZ_OPT-"-e9"} takes a looooooong time to
>>> compress... but it obviously was worth it, as well as the xz compression
>>> (instead of gz) for the squashfs.
>>> I think this would make the precise version somewhere around 500MiB
>>> (since it is the pre systemd version)
>>
>> Great that you find an effient compression method and made the iso file
>> within CD size :-)
>>
>> I think you need a good number-crunching computer for this task. If I
>> remember correctly, xz can use only one thread/processor, so you need a
>> fast processor.
>>
>>>
>>> -- 
>>> Regards
>>>
>>
>> -----
>> *Most of the issues reported below for the previous iso file are solved*
>> -----
>>
>> Den 2016-01-30 kl. 23:27, skrev Israel:
>>> Hi everyone,
>>> this ISO fixes a number of issues.
>>> I had a few setbacks with the pae kernel progress, so for now it has the
>>> same behavior.
>>> Grub is installed with progress, so it is a bit nicer now.
>>>
>>> ISOhybrid is enabled.
>>>
>>> It is a bit larger than earlier, but I will soon look into trimming down
>>> the live version.  This may mean I cut a few more things, but I *may*
>>> see if I can build the ISO as a non-systemd version just to save the
>>> extra 90 some odd MB it is adding with all the dependencies.
>>> I will check through my build again and see where the most space is
>>> coming from.
>>>
>>> zsync the latest!
>>>
>>
>> Hi Israel,
>>
>> Testing the current debian iso file (and comparing to the previous test)
>>
>> Den 2016-01-23 kl. 21:37, skrev Nio Wiklund:
>>>
>>> Hi Israel and the other developers
>>>
>>> I have some time now for testing, first in my Toshiba with an Intel i5
>>> processor and 4 GB RAM.
>>>
>>> Live:
>>>
>>> 1. You forgot to treat the iso file with isohybrid, so I had to do that
>>> in order to make it work when flashed with mkusb into a pendrive.
>>
>> It is a hybrid iso file now
>>
>>> 2. The information about the version according to
>>>
>>> uname -a
>>>
>>> is a little confusing (i586 or i686?). It does not handle more than 2 GB
>>> RAM well, which indicates a non-pae kernel. I will check it in my IBM
>>> Thinkpad with an Intel Pentium M later.
>>
>> The following command works in the installed system and might work also
>> via chroot:
>>
>> sudo apt-get install linux-image-686-pae
>>
>> After reboot I get the following response, which make me convinced that
>> it works:
>>
>> $ uname -a
>> Linux torios 3.16.0-4-686-pae #1 SMP Debian 3.16.7-ckt20-1+deb8u3
>> (2016-01-17) i686 GNU/Linux
>> $ free -m
>>              total       used       free     shared    buffers     cached
>> Mem:          3929        237       3691         30         18        128
>> -/+ buffers/cache:         90       3838
>> Swap:         4838          0       4838
>>
>>
>>> Installing:
>>>
>>> 1. At the basic OBI level: The window 'Select device - and continue' is
>>> too small (should be wider to accomodate all the text)
>>
>> I did not react to this - so I guess it is good now
>>
>>> 2. For some reason the SSD drive connected via eSATA is shown twice in
>>> this menu (minor bug but confusing).
>>
>> I did not react to this - so I guess it is good now
>>
>>> 3. There is no box to check in the 'Final warning' window. But you can
>>> click on the text 'Check this box ...' - maybe it should be changed to
>>> 'Click on this line to get it checked, if you are ready to go'. I think
>>> this is a zenity bug (or bad cooperation between zenity and jwm).
>>
>> This bug remains, also found by Jack. It might be hard to get fixed.
>>
>>> Installed system:
>>>
>>> 1. Apparmor init failed for me too.
>>>
>>> 2. A start job is running ... (waiting for 90 s: systemd problem?)
>>
>> I did not react to this - so I guess it is good now
>>
>>> 3. The tick boxes in ztweaks is missing, you have to click on the white
>>> space where it is supposed to be in order to activate it. The same
>>> zenity bug as before. In this case it does not work to click on the
>>> text, so it is really confusing.
>>
>> As I said before, we might skip ztweaks (until this bug can be fixed).
>>
>>> 4. The colours of the prompt in the terminal window are not green and red.
>>
>> This is a minor bug, but it is still alive.
>>
>>> 5. The panel icon for audio is bad, or should I say, the app behind it
>>> does not work correctly. I cannot open the mixer. I have to select
>>> preferences and show slider, otherwise it is mute. 'Open mixer' just
>>> flashes by - does not leave anything open.
>>
>> This looks good now
>>
>>> 6. The wired network works. I can use ssh. But I have to 'sudo ping'
>>> (maybe this is a feature of debian, different from ubuntu).
>>>
>>> Installing:
>>>
>>> 1. At the advanced OBI level: The checkbox is missing (the same zenity
>>> bug as before).
>>> IBM Thinkpad with an Intel Pentium MI did not react to this - so I
>> guess it is good now
>>> 2. I tried to install the PAE kernel this time. But
>>> uname -a
>>> and
>>> free -m
>>> seem to display the same as before. Are you sure that there is a
>>> difference - that different kernels are installed?
>>>
>>> 3. I tried again and installed a third time, (to a third ext4
>>> partition), making sure I installed the 'PAE kernel'. This time the
>>> installed systems grub did not display the previous versions as lines in
>>> the grub menu. But it worked to run
>>>
>>> sudo update-grub
>>>
>>> to get them.
>>>
>>> Still the same result from uname -a and free -m, so I think it is the
>>> same kernel all the time.
>>>
>>> Pentium M:
>>>
>>> *. The live system and the installed system work in the IBM Thinkpad
>>> with an Intel Pentium M, and the the kernel is a non-pae kernel.
>>>
>>> It seems that we get a non-pae kernel also when we select to install a
>>> PAE kernel (in computers allowing it). So [I am almost 100% sure] this
>>> is a bug.
>>>
>>> -o-
>>>
>>> There are some bugs, but considering the short time you have been
>>> converting ToriOS to debian, I think you have done a very good job. Most
>>> things work
>>
>> The Final Warning window of the advanced OBI level still says
>>
>> 'Check this box if ...' but there is no box. Please change the text to
>>
>> 'Click on this line ...' or something easy to understand in both cases
>> (of the basic and advanced OBI level). Now I think is bad English and
>> maybe hard to understand for the basic OBI level.
>>
>>>
>>> Best regards
>>> Nio
>>>
>> -----
>>
>> * 1. But I found a bug in the OBI-installer *
>>
>> It works well to install the PAE kernel. The non-pae kernel is kept, and
>> can be booted. Previous operating systems are detected and can be booted
>> too.
>>
>> It works twice. The first time at the basic OBI level, and once at the
>> advanced OBI level. But next time (at the advanced OBI level) I install
>> into another ext4 partition, the previous operating systems are not
>> detected and will not be available in the grub menu. I tried several
>> times, and it just refuses to work. (It makes no difference if I install
>> the PAE kernel or not, and no difference if I select autologin or not.)
>>
>> As usual the previous operating systems are found when I run
>>
>> sudo update-grub
>>
>> when booted into the installed system.
>>
>> * 2. And I know of a bug in mkusb *
>>
>> mkusb can make live-only [USB] drives, but must be changed in order to
>> create persistent live drives of ToriOS-debian. I understand that I have
>> to do it, but I don't know when there will be enough time for me to
>> focus on that programming effort, maybe sooner, maybe later.
>> -----
> 
> Hi again Israel,
> 
> The bug in the OBI-installer (not finding the previous operating
> systems) occurs as usual in my Toshiba, when I install into an SSD
> connected via USB 3,
> 
> http://www.toshiba.se/laptops/satellite-pro/c850/satellite-pro-c850-19w/
> 
> -o-
> 
> Now I have also tested in a desktop 'cube' with the motherboard
> 
> http://www.asus.com/Motherboards/M2NVM_DVI/
> 
> and an AMD Athlon 64 X2 Dual Core Processor 4400+
> 
> The same SSD was connected to this computer via eSATA. Here and now the
> previous operating systems are detected and can be booted.
> 
> So it seems that the bug is limited to cases when the target drive is
> connected via USB. As you remember, we discussed the same problem with
> the Ubuntu based ToriOS.
> 
> Best regards
> Nio
> 

... and it works in the Toshiba too, when I have moved the SSD into the
internal SATA bay.

Best regards
Nio