02:03 Feijstag: Something happened with my Xorg , not obvious as per logs. Where i switched to wayland, and it runs very fine. w00t.
02:26 Feijstag: Well, I am improving the protocol for writing new filesystems, it had some shortcomings before. Not so cranky anymore, I turned into silent mode, i got to accept that there are no ways to describe what I do(better to just code), i had very big success on compression, i am already targetting bugs and shortcomings. And web certificate signing requests for free certificates, i found very many.
02:26 Feijstag: I tried very many proxies but stunnel plus tinyproxy worked towards the end. Google among many signs webcertificates for free for their cloud service. I got a lot of momentum and handle/grip on my ways , it seems my heart is freezing i can not believe that i was actually showing some correctness on those ideas, Testing testing and testing, but early alpha is scheduled somewhere, but i am
02:26 Feijstag: already satble ways accessing the data which is awesome, i had many doubts before but now i have faith in what i do, but it's tremendously sophisticated to descibe those methods. One silly mistake on manipulating sign and inversion and data goes wrong, but a huge set of tasks ahead. I introduced a compensation protocol, where the aim is to finally merge the compensation into more accurate
02:27 Feijstag: protocol. So long story short, i have been away for some time and will be so all this year, cause i am coding. But MXlinux wayland and all KDE Plasma, is just fantastic how it runs on all my hw already, my ideas get more attention in certain other areas. Good luck for everyone I am very sorry as to what i might had caused indirectly to you, i am dirty brick, i was wrong for so long time, but
02:27 Feijstag: yet close in absolute general view.
03:30 Feijstag: The protocol improvements is in general view you have two different signs, the remainder is off a plus and one copy of sub operand needs to subinvert the targeted element , currently i get 4+57+57 which is where 4 is 118-4, so four times the value, but now when i put the 37 little bit higher cause 37+37=74 118-37=81 cause 118-74=44 is bigger than 37, such small bug fixes on protocol, then
03:30 Feijstag: finally stretch the spectrum with one more protocol fix to yield a single value right away. Then move to offset calculations, and decoder encoder without multiply and with multiply for the bank calculus. Snippets is there are no triplets anymore on machine words, but 32th or 64thlets. And massive amount of data can be stored. But now i go to away mode. I post the snippets later, it's one add
03:30 Feijstag: and several and adds colasced into sub, the correction is only possible if some conditions are forces like told, 44 can not be higher than 37, or things can collide. Small fixes but a lot of work ahead. So to end here, the final protocol improvement is global state based, because on the add twice of storage would be needed on small fixes, so fill can be only half a word, but if youd, state
03:30 Feijstag: with dummy access on both operands theyd cancel without overcommit. So i log out now, i look at logs all the time.
03:48 ccr: well, he was absent for a while ..
21:20 karolherbst: zmike: image_copy_buffer doesn't seem to properly work for 16 bit formats, do I have to do some adjustment based on the pixel size?
21:24 karolherbst: mhhh if I pass an explicit stride it works
21:25 karolherbst: ahh yeah..
21:25 karolherbst: I'll leave a comment
21:28 karolherbst: now it works :)