It's not Microsoft curly quotes?

I was 99% sure it was when I saw it, but I was trying to think if there could be a weird reason for it. Really can't think of one.

BTW you should prefer the use of $() to backticks
http://mywiki.wooledge.org/BashFAQ/082

I do do that, although they are essentially the same. I try to avoid anything running in a subshell because of its issues with set -e and my scripts' heavy reliance on this convention.

edit: wait is $() and ` ` even a subshell? I thought it was, but as I typed that I'm not sure anymore.

Basically the only command I even use that would take backticks is:

often used in conjunction with:

There's no situation I know of where backticks are better than $(). There are some cases where it doesn't matter but in general it's just cleaner.

googled it, and yes, backticks and $() are executed in a subshell. i psych myself out so much on **** I already know.

Its time for everyone's favorite ****ing day: node js lts jumps from v8 to v10. And everything breaks

What is it with the tech industry's allergy to version 9?

iirc, Node's versioning system treats even and odd numbers differently, so jumping from 6 -> 8 -> 10 is expected (forget what the odd numbers mean)

Odd number node versions means literally no one uses it as no dependencies will work with it. I don't know anyone who uses current version of node.

Good example of how SO sucks - https://stackoverflow.com/questions/...ce-a-linebreak

The first answer is completely wrong, yet there it stands. Second answer has more points but is still second for some weird reason. Also it's actually correct - but gets called out for not 'substantiating' - whatever the hell that means.

Gotta give up your hate for SO suzzer lol

I'm really digging C# and Visual Studio. I've come up with a workflow that has a parent solution for all our lambdas. The parent repo has the solution file and supporting llbs. But each lambda is a git submodule with its own repo. The lambdas also use serverless.com framework to deploy during development - but devs can also use the AWS SDK for Visual Studio (PC-only though ).

SourceTree is smart enough to detect changes to any submodules and ask you if you want to commit those as well as your commit to the parent repo.

Seems like it's going to work pretty well to a) let a new dev pull in the entire environment and all the lambdas w/o a ton of wonky scripts, b) make it easy to quickly change a lib and a few lambdas and check it all in and c) make it easy to fix one lambda and commit it to start the CI/CD process - w/o triggering anything else.

git submodules can get kind of wonky, and I'm sure there will be some messes to untangle. But overall I think it will be worth it.

and for keeping most employee salaries secret.

and for refusing requests for raises. It's hard to claim your pay is no longer competitive if you don't know what your peers make.

I was chasing down a really evil bug today and ended up on SO, with top rated answers not even answering the question being asked, and it ends up going basically unanswered.

I was dealing with an insane bug caused by a subshell. Subshells are the devil I am convinced.

a) there's nothing wrong with subshells. Honestly they're the best way to handle cases where you want to avoid polluting your main namespace. You can set variable and change directories and then when it's done, poof, it's all gone.

b) as you noted the other day, any time you run any command in the shell and want to use the output, you're going to use a subshell.

right, but with the convention I have designed my test harness and scripts, with the set -e/trap errors - stuff failing in a subshell will (correctly) not exit the script, just the subshell.

I handle this by collecting PID's of my background jobs, waiting for each of their PID's individually, and collecting return statuses of each, erroring out if I need to.

However, unknown to me, an older change (not by me) had a background job called by my harness, which called my test scripts also as background jobs. So, my test scripts were correctly catching and reporting errors, and the harness was correctly waiting for their PID's and catching error statuses - but not exiting the main script, because this was all happening in the sub-subshell. So I was pulling my hair out for about an hour because the things I was seeing seemed absolutely impossible. Initially I thought it was a synchronization issue until I noticed some other weird things, like script variables not being set when they should have been. That's what alerted me to the subshell issue.

it was kind of embarrassing even though it wasn't my fault - there is a lack of faith in this system I am doing by a few people, so they are looking for any excuse to say "see, it doesn't work." Actually the person that alerted me to the bug has been aggressively trying to break it, which is fine by me, because it alerts me to serious bugs like this. But still a little stressful.

I'm worried at this point because I have been so aggressive in taking on new tasks that I have ownership of too many systems, because I'm starting to find that I've got my fingers in 40 things now and my work is piling up. That's theoretically fine with me because i HATE having nothing to do. But this week I really noticed that I might be taking on too much.

You should leave a comment to let them know - they like that.

If you don't have enough points to leave a comment, leave an answer instead - with a snarky note about how you can't leave a comment - they really like that.

jmakin,

Just keep doing exactly what you are doing and pay no mind to the people you work with. They generally sound toxic and are working at a place that you are considering whether it's time to dip after only months as your first software job.

Continue to improve and learn as much as you can.

I just don't want to put myself in a situation to get extremely bitter about doing twice as much work as my peers while making probably ~20% less, like I suspect is happening.

The productivity range for developers will be astounding everywhere. And 3 months is probably too early, but when you are confident that you are productive and useful, get yourself a big raise. When you think you are indispensable, get a huge one.

20% of what they are making is nothing on the grand scale.

What is important is gaining the experience and skills to make exponentially more than them. Everytime they are being lazy/useless and making you do the work, they are actually giving you an opportunity to improve and get better. You should desire to do more than them. Not just 1.5-2x more productivity, push for 5-10x+.

Take all the work you can get, and if they sit around on their asses, **** em. Gimme gimme the knowledge and the skills, let them kick back and drive themselves into the ground.

Stop focusing on the short term and the toxic people around you. Think about where you will be in 5-7 years from now. Knowledge and experience are 90% of what matters right now.