Do they have to include equity/bonus?

No they don’t have to include equity and bonus in their range. I asked for a range for base and the base salary they ended up giving me was higher than what they originally quoted. What they quoted was also lower than what I saw on levels.fyi so I would just assume anything a recruiter says is bs.

How is this compatible with the law? Especially at a large company with very defined pay bands!

(I would guess it is not)

Unless you mean an internal recruiter, I just assume anything a recruiter says is either an outright lie, or if you want to be charitable, merely mistaken.

It's 2020 and googling a lodash command still takes you to the lodash home page where you have to drill down all over again to find the command you want.



Clicks link:

just had company give me a survey with 500 words on it and asked me to check the words that described me.

kill

me

now

that might be one of those BS tests. like they want you to think outside the box and tell them to **** off bc this is some BS. I dunno, proly your best shot.

This is actually the right move either way:
* test is fake to see if you'll (correctly) put up a fight: win
* test is real and they think it's important: win because **** these guys

lol ya thats the other thing. you proly dont want to work there.

anyway, heres a really weird message I just got on ****ing LinkedIn.

like wtf, I should proly report this dude to linkedIn right? this is def a scammer phishing thing. I guess I will jsut report it to my company. eh, proly just do nothing. but this is a scam of some sort right? note that I work a really big company whos data would be like super mega important.

I've gotten similar messages via email a couple times, but for AWS or Azure instead of Oracle. Usually it's a super green sales guy who's basically doing cold calls to (apparently) try to drum up cloud spending. I just ignore them.

i asked for feedback for a final interview i did a month ago and then got ghosted on and they are now like "ACTUALLY, WE HAVE NOT OFFERED THE JOB TO ANYONE BUT WE WANT YOU. CAN YOU SIGN THE CONTRACT TODAY? WHY WON'T YOU SIGN THE CONTRACT TODAY? SIGN THE CONTRACT, MOTHER ****ER".

any place willing to hire me seems to be basically insane.

Yeah probably wouldn’t go there tbh

https://twitter.com/foone/status/122...370355200?s=12

Pretty funny bug that came from multiple people copying an incorrect solution from Stack Overflow.

Crazy how often that's the case.

Still butthurt eh?

SQL question: when using the REPLACE() function, is there a way to count how many replacements are made? Or even just to determine if any replacements were made at all?

For context, I have to write a script to anonymize personal data. This is straightforward enough in most cases. We also store several “comments” fields in our database, though. Those could also contain data I have to anonymize, so I basically just have to parse them and remove information when applicable.

The thing is, these columns rarely have a value. When they do, they rarely have any information I have to worry about. This will be rarely used and doesn’t really have to be performant, but it just bothers me on principal how inefficient it is. Right now I basically “update” the field by using the REPLACE() function to remove what I want. At the same time, I update some columns used for tracking purposes, so I’d also like to avoid making “fake” updates to those columns; that is also a pretty minor concern though, and I doubt anyone cares about it except me.

Anyway, if there’s a simple answer to my question at the top I’d love to hear it! Any other ideas are also welcome.

I don't think "replace()" is part of the SQL standard so the answer is probably dependent on what database you're using.

Is this a one time thing? Like you're anonymizing it once? Or something you run on some kind of interval?

If it's not a one time thing:
Why not anonymize the data on the way in, i.e. never store sensitive data in the first place?

If it is and you care: make use of whatever internal programming language your database has. I have found that usually the built in functions are awful and if you don't want to use the built in language then you're going to do a lot of redundant stuff. Like... do a replace on one column but also do a find on the same column to deduce if the replace did anything, etc. It's super common.

Do you need to update the tracking columns anyway? Anonymizing a column doesn't really count as a real update in my book.

Just happy to share the hate.

I'm overhearing my coworkers struggle mightily to get a docker-image-hosted web app working on a second dev's machine. But it works on the first dev's, so weird! I hear stuff about port 8080 and a lot of the same issues that always seem to come up with running a web app locally in docker.

I can't decide if I should get involved. But I'd lay odds this is a problem that doesn't need docker at all.

Even for stuff that doesn't "need" docker I still prefer docker. It is really nice not to have to maintain all the node nonsense and npm modules for our website's front end on my machine just so I can run it locally. Even more so because there is more than one UI that I might want to run.

If they're muttering about port 8080 then it's probably just the standard conflict that something else is using the port they want to use. You'd have the same problem without docker.

I'm sure it's great if you use docker all the time. But when you hand a docker-based app over to some shop where no one uses docker, it can be a pain for them to get working. Generally the instructions tend to leave out some steps and assume a medium-level working knowledge of docker concepts. It either works immediately or you're in for some googling.

My coworker found some instructions for deploying a node app to Elastic Beanstalk via a docker container. So now we have (what seem to me) unnecessary docker steps in our build process - all because the person who wrote the example he found loves docker. But I'm not going to go in and try to rework it to remove docker.

It really kind of depends on how well it's architected too - most of our stuff you literally just need to install docker and then run a command, and it will do the rest. In some cases you will need to do a "docker login" if you need to pull something from our private repo. But that's pretty much it. We recently hired someone and I led him through getting a dev env set up - it used to be a full day job with a wiki that was pretty much always wrong, and now it's a 10 minute deal.

Do you run docker commands locally like docker-compose run app npm test, or whatever, or do you go into a terminal inside the container to do everything? I find the lag annoyingly slow between running a docker command and getting it to execute, otherwise i would use it for everything

docker-compose run would spin up a container from scratch and then run your thing - that's going to be slow, yeah. I generally do docker-compose up to get whatever containers I want going up, and then "docker run" to run something inside the containers. If I want to do something more interactive then I run "docker run IMAGENAME bash". For stuff I do fairly often I have a bunch of bash aliases.

The other major use case is that I started distributing tools via docker containers. I can't link to any of our stuff because you need to be docker-logined to our artifactory, but as an example

docker run --rm=true eclecticiq/cabby:latest taxii-discovery --path https://otx.alienvault.com/taxii/discovery

This will run "cabby" from a docker image. Yes, you can pip install cabby and then run it, but the chance of it working first time from docker is basically 100%, and the chance of it working universally as a python package is high, but maybe only 80% - competing package versions, maybe needing to set up a virtualenv, issues installing or not having root permission to install packages, etc. It's guaranteed not to cause any problems or clashes pretty much.