motiejus

Yes. All SO, but with our "internal" content.

Some of the most awesome engineers contributed a lot, which made it pretty high quality.

You could also ping a person when asking a question, that's usually more polite than pinging over chat with your question.

Some high-touch teams (e.g. software networking, with whom almost everybody has something to say/ask) set up their "how to ask us a question" to post on internal slack. Works great.

Edit: also keeps people more honest when asking questions - SO taught us to ask questions correctly. :)

Current Uber employee

Curious why nobody mentioned Turris series - they provide consumer routers with OpenWRT, and with upgrades.

Is there a written version/summary available?

I was veey excited when I found this at first; it had some rough edges when I tried the branch (it wasn't merged yet), but it's been a long time ago. I will give it another shot.

Finally the hosting server need not be trusted. :)

Is there a written version?

Nothing about licensing or patents. I assume the worst (read: unusable for small businesses)?

10+ years ago I worked in a small voip shop, where we had very high quality (low jitter), but low bandwidth connection. I researched many codecs of the time (2010-ish).

We liked speex, because it can be used "without strings attached". Also, I can choose the quality depending on the bandwidth. Although for low bandwidth g729 was better. Which we couldn't use because of royalties (but allowed myself to test it).

We chose alaw/ulaw when bandwidth was not a concern, and speex when it was.

Since it does not mention usability outside of google, I also find this comparison unfair or incomplete: if you are comparing a proprietary codec, compare it to g729. If you are comparing a codec to speex, it should be open/free.

Edit: grammar

It's a re-implementation of https://akkadia.org/drepper/SHA-crypt.txt in Go. Because Linux does not support shadow in bcrypt or scrypt. Also see https://access.redhat.com/articles/1519843.

Unfortunately, none of that is open source.

I do.

I implemented a specialized json-patch and tested it with gopter.

My colleague implemented an equivalent of `openssl passwd` in go and used gopter to test it (comparing outputs between his implementation and what openssl returns).

Needless to say, gopter found subtle bugs in both cases.

I came to property-testing from PropEr (erlang), and was pleasantly surprised how well gopter worked for us.

Can you recommend a specific book?

Anything similar to https://github.com/drduh/YubiKey-Guide for what you have just described?

I am willing to use what I can understand, backup and operate, and yubikey+gpg seems to be it because of this guide.

Anything practical for what you've mentioned?

Edit: for ssh and encryption.

I don't trust aws will do a good job with the fork. Can anyone tell me a FOSS project led by amazon that's not for accessing their services (boto)?

Now that elastic.co is going sideways, sphinxsearch (best search server experience I've had) non-foss since circa 2017, what are the good search server options for a small shop dealing wth geo search?

Edit/disclaimer: former aws employee

I use Turris Omnia, which is stock OpenWRT + their (open source) addons. You can configure VLANs like you mentioned, I do the same.

It is on the expensive side, but the hardware is beefy and you get vendor's support for OpenWRT out of the box.

I have replaced Hangouts, WhatsApp and Messenger front-ends on my phone and laptop with Element. I have been running the Matrix bridges for each[1] and am donating monthly to the creator. It's worth it.

So far the "network effects" have been coincidental: one small group of colleagues registered on matrix.org after one more colleague from our small circle turned out to have a home server. There are 7-ish of us now, 2 with their home servers (myself and the "other colleague"), and 5 with accounts on matrix.org.

To go "native", I see one lacking point with discoverability: I don't know of a way to discover my "contacts" whether they use Matrix without asking them first. Which is not true in any other messaging apps I have tried: Signal and WhatsApp use my address book and their phone numbers, Messenger is tied to my "friends". However, although Matrix allows entering phone and e-mail identifiers, I haven't seen an easy way to "find" them. Any pointers?

Overall, Synapse is easy to install and run. Took about two evenings to configure synapse + 3 bridges (whatsapp, messenger, hangouts).

[1]: https://matrix.org/bridges/

The Matrix clients I have tried have "call via audio" and "call with video" buttons available in convenient places.

> fwiw if you use the Jitsi widget, you can share the URL to your Jitsi room with anyone and they do not need Element or Matrix to join.

I would like my grandma to click a green "phone" button and answer my call. I would like my grandma to be able to issue a call for me.

> and the participants then need to agree on some server/software combo that provides the actual video conferencing.

Since the button is in a very convenient place, I would expect the integration to be a bit deeper than "hope the client understands the invite"; it's tricky to make the clients without a spec.

It would have been nicer if this were documented better/more obviously, so we know what to expect if we want video calls to work across different UIs.

I have set up synapse myself a few months ago for my personal domain. I am a happy user so, have been bridging my different chat applications through it.

There is one caveat: looks like "agreeing on how to do video" is not a Matrix standard; it's in the UI configuration or "widgets", which means only matrix/element combo is "properly" supported (with caveats).

Has anyone had success running video calls over Matrix without Element? Is there any movement on the protocol side?

Extending to this thread, will non-element applications have reasonable support for watching (video) fosdem?

Edited, thanks for the tip.

Fred Hebert (a known name in Erlang ecosystem) created a tutorial "awk in 20 minutes": https://ferd.ca/awk-in-20-minutes.html

20 minutes for that tutorial is enough to write non-trivial awk programs, and, least or most importantly, impress colleagues.

I will announce something like this soon:

- ui: joplin. - agpl, fully self-hostable. - you own your content (because joplin). - choose among free templates, or create your own. - templates will be similar or compatible to hugo, still tbd.

Optional for paying customers:

- sync via webdav to my service. - custom domain. - backups, etc.

Once this starts generating money, I am planning to spend some of it to fund e2e per-folder encryption in joplin.

Disclaimer: I've been at Uber since early 2016, not the first layoff round I see myself in.

I have observed the opposite: when layoffs are announced, it gets shared internally, conversations die off within <0.5 day, and people get on with their jobs. 2017 was different though.

I just find it more productive to work on the technical problems, rather than to worry about things I cannot control. And this is quite universal in my environment.

I am quite sure Uber will restart hiring after the current crisis is over, I highly recommend this place.

OK, great! Clicking the main page[1], see the top of README:

"build error".

Wait, what? You just released it, it should definitely work! Clicking that icon[2], checking the failed build[3]:

ERROR: no certificate subject alternative name matches

requested host name `pear.php.net'.

To connect to pear.php.net insecurely, use `--no-check-certificate'.

Really? See what's in http://pear.php.net/

> The server running pear.php.net had a fatal hard disk failure and gets replaced by a new machine this week. Until the new machine is setup, this page is up to let you continue installing PEAR packages via the PEAR installer.

In 2015? Cool.

Connecting to the https website ...

% openssl s_client -connect pear.php.net:443 ... Certificate chain 0 s:/CN=mail.cweiske.de

So you're using CN=mail.cweiske.de for pear.php.net. I don't even know what to say.. Well, happy hacking!

[1]: https://github.com/php/php-src [2]: https://travis-ci.org/php/php-src [3]: https://travis-ci.org/php/php-src/jobs/94372493

Pretty good size: 214K (on amd64 Linux):

  $ ls -lhd hello
        -rwxr-xr-x 1 admin admin 214K Oct 26 09:53 hello
  $ ldd hello
	linux-vdso.so.1 (0x00007fff5f5ff000)
	libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007f10611e7000)
	libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007f1060fe3000)
	libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f1060c39000)
	/lib64/ld-linux-x86-64.so.2 (0x00007f10614f6000)

(edit: formatting)

I saw your presentation about project FiFo in Erlang User Conference 2014. I asked you what does Joyent think about it, you said they're cool with it (gave you a t-shirt?). At that time, however, Smart Data Center was not open source.

Do you still work on FiFo, even with SDC being open? What are the main its selling points compared to SDC?

Coreboot has great documentation and community. I once "corebooted" my X60s with a picture of my family and contact details in case the laptop is lost. The laptop was never lost, but my wife loved the picture of the _very_ early boot stage.

It's a great project from user's perspective: healthy community, very interesting, quick-reward project to play with.

If you have a spare compatible thinkpad, I highly encourage you to try it out. :-) IIRC, there is a patch from me in coreboot -- it was so easy to prepare and submit one.

Docker replacement with emphasis on tool reusability. Docker made incremental builds right, and I love the idea. However, it reimplemented the whole tooling around it instead of using what's in there.

It will support FreeBSD/ZFS/Jails and Linux/LVM|BTRFS|ZFS/LXC on the first release. My goal is to re-use existing tooling as much as possible.

I am writing it in Haskell.

I never tried Docker (looking around now), but the approach is extremely dirty for a few reasons.

1. Process supervision should be handled in Docker (or something that is designed to do supervision), not in the `while true` loop. Idea on hanging on *.log files is terrible. Maybe there is a reason to stop the container when application stops?

2. There must be a better way to handle docker logs of multiple programs. If not, run both instances in different containers.

3. The thing serves configuration files to arbitrary clients that ask for them. It doesn't even log multiple downloads of the same key (though serves the private keys over SSL, which makes it hilarious on the purpose of SSL here).

Point 3 is openvpn specific and can be acceptable for example (though I am still lost why bother with SSL). However, points 1 and 2 show how to seriously misuse Docker. Either Docker or the setup is flawed (I suspect the latter).

Please do not take this article as an example how to do things in Docker. There must be better way, in more or less every step.

Microsoft Office is the last obstacle preventing my wife (and my mother coincidently) from moving to Linux. They both tried Ubuntu, liked it, but Windows VM just for office was PITA. And they switched back because of the office suite.

License of the toolchain?

This is exactly what I built SoundPatty for, but for (much) smaller scale. We needed to capture certain robo-calls incoming to our call-centre, via GSM. Link: https://github.com/Motiejus/SoundPatty

Of course, this program is only a small part of the solution. But it works very well for matching the patterns in a sound record.

Unfortunately, I am European, and cannot participate. Feel free to use it, though. License is MIT.

I can't see /bin/sh, and I haven't heard of posix shell implementation in Perl. Am I missing something?