Daily (tech related) news diet

[chwe]

You can't read every interesting news blog, annotation etc. about stuff related to tech, linux whatever is in this field. Open a new thread for all those stuff you might share with others here ends that also nobody reads it cause it's spread though various threads. So why not share them all in one thread? 

 

Idea:

Share a link to the article in this thread with a short teaser what the article is about and maybe some opinions why you like the article. It should not be 'post link only without any description what the article is about'. 

 

So, I'll start with the first one (not that recent, but I think it's still worth to read):

http://kroah.com/log/blog/2018/02/05/linux-kernel-release-model/

 

A blog entry where Greg Kroah-Hartman (maintainer of the stable kernel branch and subsystems e.g. USB) describes how kernel developments works, he also brings up some background information about kernel development, its history, stable branch, LTS branch (rules for LTS patches) and updates in terms of security (he also explains once again why Linus hates 'security patches' ). 

The reason I like this blog entry: It's written in a way that even a non kernel dev understands the context and IMO he splits background information and his own opinion (I hate opinion mixed with facts, no matter I do this sometimes too - it's hard to avoid it). 

[chwe]

 

Today one, for the people who are lazy to read stuff, you can just listen...  (record quality is a bit low 480p which makes it hard to read stuff on the slides)

 

It's a talk given by Ulrich Drepper (you might know him as the 'strident' maintainer of glibc with some well known rage outs where people had issues with his behaviour)

Quote

It's working fine everywhere but this carp architectures. I'm not going to make the code perform worse just because of Arm. Providing your own copy of that file if you care.

 

He talks about modern CPU/GPU architectures, their limitations in terms of accessing memory, speed and how this affects an efficient way of programming them. His communication style might be bad during the time of the glibc maintainership but in this talk he shows that he can explain things the way that even someone like me which hasn't a clue about low-level stuff can learn something about it. 

 

I hope you'll enjoy it.  

[chwe]

This time a less technical one.. 

https://www.linuxjournal.com/content/why-you-should-do-it-yourself

More some sort of a 'motivation talk' to do things on your own instead of buy a pre-baked solution. I may post this one in the future in the support related subforums to encourage people to try fix (simple) tasks on their own (with some hints) rather than asking for a 'step by step tutorial' for their *random feature*....  

[chwe]

This time, not a blog post, not an article but a discussion which came up on LKML after syzbot reported an error in xfs. Sometimes developers are a bit annoyed about syzbot reports due to, it only reports bugs but fixing is then 'the communities' part of the work: You should go through the whole discussion:

https://lkml.org/lkml/2018/5/21/649

 

And just for fun:

https://lkml.org/lkml/2017/3/31/641 

[chwe]

hmmm Microsoft buys GitHub....

https://blogs.microsoft.com/blog/2018/06/04/microsoft-github-empowering-developers/

I'm not sure how I should comment this..  Maybe with their last statement?

Quote

Together we will continue to advance GitHub as a platform loved by developers and trusted by organizations.

 

[Tido]

brrr, what about https://about.gitlab.com/

[Igor]

 

2 hours ago, Tido said:

brrr, what about https://about.gitlab.com/

We are ready to go https://gitlab.com/armbian  It was working extremely slow and took ages to register. I guess everybody wants to escape the monster ASAP 

[zador.blood.stained]

Looking at gitlab performance it may be easier to host a simple Git server ourselves, though issues and PRs would become a problem.

Regarding performance BitBucket may be a better alternative.

 

But in any case it's too early to look at this, GitHub still works fine for us.

[chwe]

6 minutes ago, zador.blood.stained said:

Looking at gitlab performance it may be easier to host a simple Git server ourselves, though issues and PRs would become a problem.

Regarding performance BitBucket may be a better alternative.

You can host your own Gitlab (community version, what this means performance wise should be tested first).. As far as I know Debian hosts their own GitLab server and you can allow that people can use their GitHub credentials to login on your own GitLab server. But, this was just a small news thread and not a call that we should move away from GitHub..  (I think we have bigger issues than the Git* service we use).  It's good to have gitlab.com/armbian reserved but this shouldn't be a 5 seconds-"Microsoft is evil, let's move to GitLab"-decision...

It's not that long ago that microsoft announced that they use git intern: https://www.infoq.com/news/2017/02/GVFS 

hope for a 'git push remote master' of their kernel... 

 

As long as google doesn't announce that they bought the Linux Foundation, we shouldn't have major issues..  

[JMCC]

Run for your life...!!!

[chwe]

14 hours ago, JMCC said:

Run for your life...!!!

to late, now these guys are in power..  (okay, to be fair, he retired in 2014)

 

 

but hey, he loves developers so - nothing can go wrong...  

[chrisf]

5 hours ago, chwe said:

As long as google doesn't announce that they bought the Linux Foundation, we shouldn't have major issues..  

Don't worry, Google is only a gold member of the Linux Foundation.

Microsoft is a Platinum member...

[umiddelb]

9 hours ago, Igor said:

We are ready to go https://gitlab.com/armbian  It was working extremely slow and took ages to register. ...

seems that you're not the only one

[balbes150]

That's bad news. The worst thing, they were going to implement their "tools".

[chwe]

8 hours ago, chrisf said:

Google is only a gold member of the Linux Foundation.

Microsoft is a Platinum member...

As long as there's a grumpy finnish guy ranting when he doesn't like code, there's 'some' safety.. 

 

From microsofts point of view, this could be a smart move.  Don't change that much in the first one or two years in the front-end so that nobody cares anymore and then start to change things slightly...  Somehow similar to what Facebook does with whatsapp..  

GitHub as a service was never FOSS, it is/was just the easiest way to share your repos with others (you don't need a own git server, you don't need to care about traffic, load balancing for bigger repos, server security etc.). IMO, if you want to monetize it you've to link the backend data with portals like LinkedIn etc. I could not understand how companies came to the idea that outsourcing their git servers to private repos on github could be a good idea. You may save some bucks, but when proprietary code/software is your business model, there's no way to store it 'somewhere else'. Maybe that's me being paranoid, but I don't trust 'free/cheap' cloud services for stuff that should stay secret. But it seems to be a general trend to store everything in clouds, having every software as a web-service instead of dedicated programs on your PC (this can work if you host those services on your own server but your CFOs bonus gets smaller if the 'fixed costs' are too high ).

 

 

Todays article:

 

A 'short' blogpost about a guy getting a 'fitness tracker' as part of his welcome present for joining a new company. He decided to open it and describes his efforts to reverse engineer it up to the point where you can use the display, some digital pins, the accelerometer and the BLE module with a 'adjusted' arduino IDE. So, for around 13$ (from what I saw on Aliexpress) you get a battery powered thingie with a small OLED, accelerometer, BLE and a button. Based on a nRF51822 with  256 kB flash, 16 kB RAM. Links to the needed GitHub repos are linked in the blogpost (better hope that they didn't delete their accounts.. ).. 

 

https://rbaron.net/blog/2018/05/27/Hacking-a-cheap-fitness-tracker-bracelet.html

Have fun.

 

[Igor]

 

[chwe]

20 minutes ago, Igor said:



you probably missed those guys..

 

 after MS was there, they had to start from plain..  (not that MS was in charge to bring them to the cliff, but..)

[TonyMac32]

17 hours ago, chwe said:

As long as there's a grumpy finnish guy ranting when he doesn't like code,

I really identify with that guy's frustration, I took one of his rants and substituted sensor words for software words and no one could tell I hadn't written it.  I'm really extremely pleasant here (comes from being the least knowledgeable among the devs, I at least try to shut up and learn)

 

Skype was never valuable in the US, now they rebranded Lync as Skype, which in my opinion made it crappier (there is nothing professional about Skype)

 

I can't remember ever having a Nokia... Maybe one of my early button phones was one, the first I bought myself was a Sanyo Katana, an exceptionally solidly made device (better built than the phone it was a blatant ripoff of)

[balbes150]

Spoiler

 

Look carefully at what the MS buys and what becomes of them. Look who the real master is in MS. Then you will understand why they do it. Now they bought github, for several reasons.

1. Get full and uncontrolled access to the database with personal data of all developers with github. As a result, they are able to total control over any interested them, the developer.

2. Get full access to private source code. This will allow them to uncontrollably steal any code they need. Know all the internal features of any program and use for your benefit. I think they want to access all the encryption and data transfer algorithms that are used by independent developers (something that is not authorized to replace it and access protected information).

 

 

[Tido]

1 hour ago, TonyMac32 said:

now they rebranded Lync as Skype

I must use S4B (Skype for Business) at work, in combination with the Office 2016 AKA O365 = Beta Software. Life was great before.

However, I am wondering how much of skype they used for S4B, the alogrithm for voice was one of the great features of Skype as I understood.

26 minutes ago, balbes150 said:

Then you will understand why they do it

I fully agree with you. In a german forum I read, identify possilble patent violation, analyze other projects and copy those in a better way, reading your personal messages, are there upcoming or running projects in server-services that your valuable for MS and so on.

 

in other words: expect the worst, to be prepared

[chwe]

6 hours ago, TonyMac32 said:

I really identify with that guy's frustration, I took one of his rants and substituted sensor words for software words and no one could tell I hadn't written it. 

It's a good quality control instrument.. And we've hopefully some years until he retires.. 

6 hours ago, TonyMac32 said:

Skype was never valuable in the US, now they rebranded Lync as Skype

Lync... The companies version of MSN messenger?  We used it only to ask if everyone is ready for lunch...  For everything 'serious' I prefer mail.

 

After some (morbide) fun about GitHub (doesn't mean that the discussion should stop), this time a more technical one for those interested in batteries and the chemistry behind it. A short synopsis about Li-Ion batteries and recent research in this topic. Battery research is somehow the new 'hot topic' in material science at the moment (last ~3-4 years from now) and more or less every one or two months I spot a paper where they claim that their approach should be 'the next big thing' for better durability, increased energy density etc. Keeping track on it is impossible without access to some literature behind expensive pay-walls. ACS Central Science published recently (2017) this short one (7 pages).

To give you a bit background, ACS Central science is 'somehow' similar two famous journals Nature (IMO more focused on biology) and Science (focused on physics) around the central science of the humanity (and it's obvious that this must be chemistry ). Their IF/JIF (impact factor) might be a bit lower than the two others but everyone has his own opinion towards IF as a benchmarking tool (those publishing often in high IF journals may see this different than people who don't).

I think the basics behind this article is understandable even when you don't have a degree in physics (battery science is IMO more related to physics than to chemistry) and may be interesting for those who may want to get a short insight into this.

 

https://pubs.acs.org/doi/ipdf/10.1021/acscentsci.7b00288

 

For those not interested in chemistry/physics, there's a nice write up about the concerns to implement Speck (a lightweight block cipher) into the kernel. As part of the 'ISO board' which rejected it he gives some explanation why they rejected it:

https://www.spinics.net/lists/linux-crypto/msg33291.html

but seems that he wasn't loud enough:

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=da7a0ab5b4babbe5d7a46f852582be06a00a28f0

[chwe]

https://gizmodo.com/spains-biggest-football-app-reportedly-turned-fans-into-1826728587

 

hmm, read it on your own.. there's no way in hell to explain this without being unpolite from my side.. 

[chwe]

http://www.worldofindie.co.uk/?p=682

 

Amazing pictures from analog transmission of photos from the moon and the story to recover them. 

[chwe]

The article describes a weirdo app I never heard about before.. It's some strange mixture of 'social media' mixed with an easy way to pay without credit card.. In it's default every transaction is public.. Funfact, they have some sort of a json string which can be accessed by everyone with a bunch of informations. 

 

https://www.theguardian.com/world/2018/jul/17/venmo-payments-app-default-privacy-settings-public-information

 

really? do people really want that their financial transactions are shared that way (not how much you spent, but conversation related to it, full name nickname and to whom you paid, including their data)? Do I miss a new hipster trend? Should I do this too?  And my question.. Why should 'public' be default in such an app.. well, I'm curious what others think.. 

[chwe]

Not really tech related but maybe interesting. 

http://www.pharmatimes.com/news/europe_recalling_some_valsartan_meds_on_safety_fears_1243435

 

One of the more famous drugs against high blood pressure, valsartan, gets now recalled cause it high amounts of N-nitrosodimethylamine (NDMA). NDMA is, as many nitroso compounds highly toxic (most likely this stuff is carcinogenic). According to this article (in german), it might be possible that those high amounts of NDMA might be there since 2014 due to a change of their synthetic route (I didn't check patents nor registration sources to be sure here). Seems that they didn't consider that their new process might produce it as a side-product and that the authorities didn't check this either (can't understand why, 'normally' something like this should never happen).

But, just to don't smash them down fully..  quantitative analysis of nitroso compounds isn't that easy, first this small one is probably not 'UV-active', so one of the most used methods (HPLC) wouldn't detect it. They are often thermal labile so the second one which is often used (GC) might end ind decomposition and if you have a setup connected with an mass spectrometer as detector, during ionization this thingie gets probably smashed too. So, if you don't realize that nitroso compounds can be formed and you don't develop a method to detect them, chances are high that you'll miss it later. It's not something you'll realize by luck after its. When the main patent of a drug expires, it's quite often that a lot of generics show up immediately. Some of them produce their compounds on their own, but also a lot of them buy the API from producer and press their own pills (and for the rest it's just from the original production plant where the company has a 'sub brand' selling their products as generics a bit cheaper as soon as the patent expires). IMO, the more companies are involved into the production of your pill, the harder it is for those companies to keep properly track (who is responsible for which part of QC). Mostly (also due to pharma is a highly regulated market) this works without issues, but as soon as it fails, the impacts are often tremendous, it affects a lot of people, and well nobody wants to get cancer from his high blood pressure pill. 

Will be interesting as soon as the reports show up how this could happen. 

[chwe]

https://www.washingtonpost.com/world/2018/08/06/amid-spat-with-canada-pro-government-saudi-twitter-account-shares-image-plane-flying-toward-toronto-skyline/?utm_term=.041b92b9bc28

 

Diplomacy done with 280 characters... Seems to work well between saudi arabia and canada.  Not really a tech-related, but IMO it shows that 'social media' might not be an appropriate tool for a govt to comment their own or politics of other countries... Or at least people using twitter on behalf of a govt should be well briefed what they share on twitter. You've no idea what kind of 'trolls' will step in and make the situation even worse. 

 

Background: Canadians Foreign Policy twitter account posted this one:

 

Saudi Arabia wasn't that happy with this statement, called back their ambassador from Canada and kicked out the canadian ambassador. A Twitter account 'somehow' connected to Saudi Arabia (seems not to be an official govt account), shared this one later:

 

claiming later that it should show that Canadians ambassador flies back to Canada (well, that's up to you to decide what it shows - you should not work with 'memes' when it comes to diplomacy, that can go terrible wrong)...  @martinayotte do you have some 'insights'? 

[martinayotte]

22 minutes ago, chwe said:

  @martinayotte do you have some 'insights'? 

Not really ... I've seen the news on TV yesterday ... Diplomacy is always a double-edge blade !

According to reporters, since there is not much economic trade between the two countries, the worst that can happen is that Saudi will block the current military trucks contract, which would be great anyway ...

 

[chwe]

On 8/7/2018 at 2:50 PM, martinayotte said:

According to reporters, since there is not much economic trade between the two countries,

seems that you have around 15k students from saudi arabia in canada.. So there's some business.. according to the media here, those get also called back.. Well, that escalated quickly.. 

On 8/7/2018 at 2:50 PM, martinayotte said:

Saudi will block the current military trucks contract, which would be great anyway ...

we've some old military trucks:

not that comfortable when you sit on the load area (there's a wooden surface and nothing else), they're a bit thirsty (something like 60-100l/100km) but they still work.. 

 

Once again one from the 'social media':

https://inews.co.uk/news/technology/facebook-hasnt-learned-from-its-mistakes-as-its-meetings-with-banks-show/

 

People start to realize how much data FB collects from them..  They may have to run their commercial once again on TV:

 

 

well I've an other one: F stands for fuck off, we don't care... 

[martinayotte]

3 minutes ago, chwe said:

seems that you have around 15k students from saudi arabia in canada.. So there's some business.. according to the media here, those get also called back..

I bet those students doesn't wish to return to their country by force, they will fill for "refugee status" ...

[Tido]

Long Term Support for Kernels:

Linux 4.4 - longer maintenance because of Android (sounds good for RockChip )

Linux 4.9 - 6 years

Linux 4.19 - 2 years

http://kroah.com/log/blog/2018/08/24/what-stable-kernel-should-i-use/