| 1 |
I would also suggest not putting this off until week 12 -- it takes time |
| 2 |
to work back and forth and make changes. Also, if you are working with |
| 3 |
other people, they are just as busy as you needing to crank out the last |
| 4 |
bits, and need whatever lead-time their schedules require. Last minuite |
| 5 |
requests always mess things up. |
| 6 |
|
| 7 |
On Aug 8 2022 7:39 AM, cat@××××××××.org wrote: |
| 8 |
> Okay I got you, sure! |
| 9 |
> |
| 10 |
> |
| 11 |
> |
| 12 |
> Aug 8, 2022, 13:18 by ebo@×××××××.com: |
| 13 |
> |
| 14 |
>> When you get the draft of step-by-step doc, sit down and do a clean |
| 15 |
>> install and follow the step-by-step to verify that it works as |
| 16 |
>> documented. Maybe if you have someone that can beta test things, they |
| 17 |
>> can verify the final docs. |
| 18 |
>> |
| 19 |
>> Anyway, that is all I meant. |
| 20 |
>> |
| 21 |
>> EBo -- |
| 22 |
>> |
| 23 |
>> |
| 24 |
>> On Aug 8 2022 6:58 AM, cat@××××××××.org wrote: |
| 25 |
>> |
| 26 |
>>> Yes! I have week 12 fully dedicated to writing documentation |
| 27 |
>>> actually. |
| 28 |
>>> I have written some notes but I'll also use my daily blogs to remind |
| 29 |
>>> me of what to write about :) |
| 30 |
>>> |
| 31 |
>>> What do you mean by "and verification that they are really up to |
| 32 |
>>> date", like so the older docs are really up to date? For the |
| 33 |
>>> PinePhone |
| 34 |
>>> Pro in particular I'd like to do something like this |
| 35 |
>>> https://wiki.gentoo.org/wiki/PinePhone but with a lot more |
| 36 |
>>> information. |
| 37 |
>>> |
| 38 |
>>> |
| 39 |
>>> |
| 40 |
>>> Aug 8, 2022, 02:33 by ebo@×××××××.com: |
| 41 |
>>> |
| 42 |
>>>> I'm not one of your mentors (I am a former GSOC intern and mentor; |
| 43 |
>>>> which explains why I am still on the lists). |
| 44 |
>>>> |
| 45 |
>>>> All of this sounds like wonderful progress. The one thing I have |
| 46 |
>>>> not seen you mention is documentation, quick-starts, and |
| 47 |
>>>> verification that they are really up to date. Spending a week or |
| 48 |
>>>> two on those to polish then up would be a wonderful addition. |
| 49 |
>>>> |
| 50 |
>>>> EBo -- |
| 51 |
>>>> |
| 52 |
>>>> On Aug 7 2022 8:01 PM, cat@××××××××.org wrote: |
| 53 |
>>>> |
| 54 |
>>>>> This week I've spent most of my time getting Gentoo musl to run on |
| 55 |
>>>>> my |
| 56 |
>>>>> PinePhone Pro and packaging mauikit apps. I have also done some |
| 57 |
>>>>> minor |
| 58 |
>>>>> testing on KDE applications and some other miscellaneous things. |
| 59 |
>>>>> |
| 60 |
>>>>> In my proposal I had initially planned on porting KDE applications |
| 61 |
>>>>> and |
| 62 |
>>>>> making them build on Gentoo musl the following two weeks, and then |
| 63 |
>>>>> making sure test suites run the week following that. But I've |
| 64 |
>>>>> already |
| 65 |
>>>>> gotten kde-apps-meta installed and the programs are working well |
| 66 |
>>>>> for |
| 67 |
>>>>> the most part. Though I haven't enabled every use flag for the KDE |
| 68 |
>>>>> apps and I haven't ran every test suite so there's definitely work |
| 69 |
>>>>> to |
| 70 |
>>>>> do left, just not 3 weeks work. So instead I asked Sam if I could |
| 71 |
>>>>> spend some time working on getting Gentoo musl to run on my |
| 72 |
>>>>> PinePhone |
| 73 |
>>>>> Pro with Plasma Mobile as a side project. |
| 74 |
>>>>> <!--more--> |
| 75 |
>>>>> |
| 76 |
>>>>> Starting with the PinePhone. The Gentoo install itself was pretty |
| 77 |
>>>>> smooth and I didn't run into any major issues. But I really got |
| 78 |
>>>>> stuck |
| 79 |
>>>>> on some other more low level and non-Gentoo bits. |
| 80 |
>>>>> The first issue I ran into was when I installed a new kernel onto |
| 81 |
>>>>> it. |
| 82 |
>>>>> Apparently a single developer called Megi does most of the PPP |
| 83 |
>>>>> kernel |
| 84 |
>>>>> development so I installed his kernel. Confusingly his development |
| 85 |
>>>>> branch is called "orange-pi-5.x" and it took me some time figuring |
| 86 |
>>>>> that out :D. Anyways, the compilation itself was straight forward, |
| 87 |
>>>>> and |
| 88 |
>>>>> the defconfig _almost_ worked well. I stole the bootloader |
| 89 |
>>>>> configuration from PostmarketOS, rsync:ed the kernel + dtbs, and |
| 90 |
>>>>> then |
| 91 |
>>>>> changed some relevant parts in the bootloader config. Sadly the |
| 92 |
>>>>> phone |
| 93 |
>>>>> did not boot, and there was no output to be seen on the screen :/. |
| 94 |
>>>>> After reading the wiki I found out that I could connect via serial |
| 95 |
>>>>> through the headphone jack. I used an RS232-to-USB adapter and |
| 96 |
>>>>> soldered it onto the internal wires of a 3.5mm cable. For output |
| 97 |
>>>>> this |
| 98 |
>>>>> did work, but when connecting TX to also get input, the output just |
| 99 |
>>>>> got messed up and I couldn't read it. I tried to debug this and |
| 100 |
>>>>> also |
| 101 |
>>>>> seeked help from others, but ultimately couldn't get it to work. |
| 102 |
>>>>> Luckily the only thing I needed was output because I saw in the |
| 103 |
>>>>> bootlog that EFI stub was missing from the kernel. Enabling that |
| 104 |
>>>>> and |
| 105 |
>>>>> generating an initramfs made the phone boot! |
| 106 |
>>>>> |
| 107 |
>>>>> Then I started emerging some packages, and even though I had the |
| 108 |
>>>>> charger plugged in, the phone completely discharged after a while. |
| 109 |
>>>>> After that the phone did not want to boot and I tried all kinds of |
| 110 |
>>>>> things, like booting from SD card, reflashing bootloader, trying |
| 111 |
>>>>> different cables and nothing worked ... It turned out that the |
| 112 |
>>>>> bootloader (Tow-boot) had a bug that made the phone not charge |
| 113 |
>>>>> after |
| 114 |
>>>>> the battery as emptied, and I needed to boot it into a special mode |
| 115 |
>>>>> holding a button with a sim card opener. |
| 116 |
>>>>> |
| 117 |
>>>>> After that I set up distcc with cross compilation, and there came |
| 118 |
>>>>> the |
| 119 |
>>>>> second issue. "__aarch64_cas4_sync undefined symbol". I asked Sam |
| 120 |
>>>>> and |
| 121 |
>>>>> he said it probably was a distcc issue. Because of me not wanting |
| 122 |
>>>>> to |
| 123 |
>>>>> run into this again I tried the aarch64-gentoo-linux-musl-emerge |
| 124 |
>>>>> wrapper instead. This worked for the most part, but I had trouble |
| 125 |
>>>>> with |
| 126 |
>>>>> copying over my phones configuration to /usr/.../etc/portage. |
| 127 |
>>>>> I then learned about the ROOT, SYSROOT, and PORTAGE_CONFIGROOT |
| 128 |
>>>>> variables. Toghether with sshfs I could easily emerge packages for |
| 129 |
>>>>> the |
| 130 |
>>>>> phone on my PC without using something slow like qemu-user, nice! |
| 131 |
>>>>> I emerged a lot of packages like this and noticed that the program |
| 132 |
>>>>> dispatch-conf did not honour the variables. This was easy to fix |
| 133 |
>>>>> and I |
| 134 |
>>>>> PR:ed it here https://github.com/gentoo/portage/pull/881. |
| 135 |
>>>>> |
| 136 |
>>>>> I have also created a lot of ebuilds for Mauikit apps, these are |
| 137 |
>>>>> cross |
| 138 |
>>>>> platform KDE applications that look great on smaller devices like |
| 139 |
>>>>> phones. https://github.com/gentoo/kde/pull/910/commits. They also |
| 140 |
>>>>> work |
| 141 |
>>>>> great on my PC. |
| 142 |
>>>>> |
| 143 |
>>>>> All in all I've spent most my time this week working on the |
| 144 |
>>>>> PinePhone, |
| 145 |
>>>>> and the two following weeks I'll do a lot of testing for the KDE |
| 146 |
>>>>> applications. |
| 147 |
>>>>> |
Archives