BGAFC Vendégkönyv

Σ: 1 post

TCH (statz)

#1, Főfasz (10443)

5892 | #50d9 | ^ | Idézet | Sat, 17 Oct 2020 13:07:04 +02

188.143.*.*

*.pool.digikabel.hu

@Carlos:
Regarding servers: Maybe the Internet Archive would be a possibility?
https://archive.org/details/cdromimages?and%5B%5D=ppc&sin=

Otherwise I have to think of something.

Perhaps, but as we see, it's not fully reliable.

@Carlos:
Ah and yesterday I also found the regular LFS version for POWER (non-IA). It is from the same guy who does CruxPPC.
https://github.com/tarmiste/lfs-non-ia

Hey, great, thanks! I've added 8.4 and 9.1 of LFS-Non-IA to the list.

@Carlos:
If this will be a candidate for Raptor is in the stars. Because POWER10 will probably use some kind of binary blobs. On Github they have discovered some hints.

I don't really understand this part. What kind of binary blobs? Is there some kind of software, burned into the CPU?

@Carlos:
These should be the slides from the POWER10 presentation:
https://regmedia.co.uk/2020/08/17/ibm_power10_summary.pdf

Hey, thanks, much better. I don't really like videos, because they "demonstrate" stuff in tens of minutes, what could be written in a simple blogpost and could be read in a few minutes.

@Carlos:
Here you can find Fedora 15 to 30:
https://puppet.fedoraproject.org/repo/pub/archive/fedora-secondary/releases/

And here Fedora 7 to 13:
https://archives.fedoraproject.org/pub/archive/fedora/linux/releases/

Unfortunately, no new images found. What is now misses from the list is not present under these links either.

@Carlos:
I am surprised that F14 was not available for PPC?

Neither Fedora 13. When they switched to systemd, they've skipped two releases with POWER archs. I don't know why though.

@Carlos:
And you can remove the download links of Frugalware Linux. Unfortunately they are all dead. Archive.org has only saved the directory.

I think i still keep them for awhile, just to preserve the original paths in the URLs, because maybe we will find another mirror one day, if we look up those paths.

@Carlos:
This can happen with the password. The good news is, you noticed it. Will they check your work and discuss it with you?

I don't know if they ever write back or not. (Regardless of the password.)

@Carlos:
If so, it might be a good idea to ask them if they have discovered the password. But the shot could backfire. Depends on the guys.

My younger brother asked exactly the same, and i answered him the same: since they have to set the password to make the stuff work with their own MySQL access, they have to notice it.

@Carlos:
But all in all, did the work go well? I keep my fingers crossed.

Thank you. Aside from this, i say it went fairly well. Though i had problems with optimizing the query what produces the main view, because due to the braindead structure, it was slow as hell. (A few million entries only and took up 3-5 secs to retrieve the data.) To say, the database they have given to me was badly constructed is mildly put. (It's some kind of test database from the MySQL team.) Erroneously crafted PRIMARY KEYs, missing indexes, duplicated entries which made denormalization a nightmare and i could go on. I've done the most of the project in less than two days, but spent four days with optimizing the main view's query. More experienced SQL guys had a hard time with it, when i asked for some help. But in the end, i've decided, that if i don't know if i am allowed to change the erroneous part of the DB or to denormalize it, then i will provide these two operations as optional. So, now it works without the correction and the denormalization with a fairly decent speed (except if you try to order the view either by the salary which has 3 million multiplexed records in a separate table, or by department which needs to JOIN the department table through the also multiplexed half million record long department number table), but there is the option to correct the indexes and denormalize the database which will yield an 5-50x speedup. (Roughly, depending on the queries.)

However, they might refuse it, because the instructions told me, that i have to display the view via the help of jQuery. Now, formally i complied, since i pull the data for the view from the server via jQuery's AJAX...but that is a meagre one line of jQuery code, the rest is done without any kind of stupid JS framework. :]
But seriously, i've done the whole thing in less than 7 kB (even if i do not use my own ~300 byte AJAX library instead of jQuery), with documentation and without minifying my source, while jQuery is roughly 90 kB without documentation and with minifying...and then i would need a table plugin too, which can weigh up to MEGABYTES of JS code and yet, they can filter by or sort for one column at a time, while my few kB large view can filter by or order for all columns, simultaneously and prioritizedly... Please do not misunderstand: i am not bragging, i am complaining. Any decent webdeveloper could do most of these stupid frameworks' "features" with a few short functions, yet we are using hundred kB large "librares" to do it. This is how the web2 bloatware works and spreads. And if you dare to object (pun intended), then you are a laggard heretic, who tries to hinder this - so called - "progress" and you should be burned at the stake which is elevated from printed FORTRAN codes. :P

Eh, sorry for the rant, but i am fed up with the ways software is developed nowdays...