Based on the three companies I have experience with they all have scripts that merge code from master or a stage branch into a release branch.

One company had a release branch for each stage where the script would allow force push for dev and qa envs and clean merges only for sandbox and prod.

The other two use a release branch only for prod and sandbox and everything else takes a branch babe as an argument to the script

Hmmmm - that's an interesting approach. So everything is still triggered by merges to a branch right?

At these companies there was a mix of merges triggering deployment, scripts run by hand that handle the git stuff and deployment, and scripts run as cron that handle git stuff and deployment.

You could definitely set it up the way you're thinking though.

My first attempt at managing expectations as a PM:

My boss/product manager:

“Oh i know we ship soon but we need this new feature that this project totally wasnt designed for and hasnt ever been tested”

Me: “ok that sounds infeasible but i’ll look into it”

Look around, get some more knowledge, learn that yes, it’s infeasible. Tell my boss in detail why it’s infeasible.

Boss/product manager: “just have (new dev in way over his head) look at it.”

“Ok, but like I said, it is looking pretty infeasible. I think we need to ship as is and work on a better solution later, since there’s no time pressure for this feature.”

Shipping in 24 hours, get an email from Product manager CC’ing everyone and an advisor that used to work on this:

“Hey, so we need to ship this tomorrow and really need to get this (impossible feature) working by end of day.”

Mother****ers are you not hearing me?

Instead of using language that sounds like it is your opinion it is unfeasible give them numbers. “This feature will take x stories or y points and we only have z sprints before the ship date. Given our current velocity the soonest we can have that feature is in r sprints.” The estimates are still most likely going to be subjective, but at least they are defensible and make it sound less like you are disagreeing with people and more like you’re trying to accommodate them.

Be careful, project managers often take the fall in these spots.

We don’t do any of that crap. I tried earlier on but I just didn’t take. Suzzer I have plenty of places to point the finger if it fell on me and without some moves I made in the last few weeks this easily would’ve been delayed a month and I could prove it.

Just show them this

omg

Ya lmao i’m totally getting bus tossed. **** it.

efforts to ramp up and hotfix right up to the deadline of course failed

now without me doing an all nighter it won't ship tomorrow

they finally realized it was infeasible and now want to spend 2 weeks and completely refactor the entire thing. Like bottom up.

I was saying this a few weeks ago, maybe I should've been louder. Now I am trying to tell them that yes in an ideal world we would refactor and stop throwing these bad hacky fixes at a hacked together codebase that none of us wrote and likely was never working properly - for single use cases nonetheless. But two to three weeks, and just one fresh dev - it's likely impossible. I am at a complete loss right now, lol.

At my old job the PMs were always contractors, so they could be blamed and let go if the project went bad. If the project went well they get to stick around for a while. If they get like 3 in a row that go well then maybe we hire them to be a "program manager", which I'm pretty sure is just a title because "project manager manager" sounds silly.

I think my plan of motivating my CSS dev by going in and mangling her stuff is starting to pay off. She's starting to say stuff like "Oh you don't need to work on that, I'm almost done with this."

I use that same method at home for washing dishes. Do a crappy job a few times and my wife now tells me not to do it.

I'm sure they would never turn that around on us for like blowjobs or doing taxes or something.

Yeah, my wife makes me give all the blowjobs.

So thankfully our scrum leader and the consultants came to their senses. Really they claimed it was all a misunderstanding. When our business leader ie project owner heard about it in the retro she was not too happy.

Anyway I still hate scrum. What do you guys think of team tasking ? Or sprint reviews?

I am good with sprint planning and retro to an extent. It's 3 hrs of meeting but that's ok bc I think there's value. But 3 more hrs for tasking and review is just horrid. 6 hrs of meetings over 2 days is just grueling and obv no dev work gets done.

I agree - sprint-planning, standups, major retrospectives - after that it's diminishing returns.

Maybe there's value once you're already super agile and humming along towards legitimate continuous delivery. But there's definitely harm in trying to be too agile too fast.

We ran tasking as part of grooming, 1 hour a week, 5 minutes a story to determine size and task breakdown was usually done by a lead. We used it as a way to monitor progress on a story and was what the burn down chart was based on. The goal was to have enough stories tasked and pointed for the next sprint. Having it done allowed sprint planning and retro to happen in the same ~3 hour sprint planning meeting.

Is GIS a good specialty? It seems fairly complicated, which seems like that would make it good.

Today, in how not to do your tech:
- sign up for account on my new healthcare provider's website
- use my personal gmail as my username
- after sign up, it tries to log me in but can't
- I try to log in with my user/pass, it fails
- I get their automated email with "thanks for signing up, your username is blah" and notice my username is missing the "m" in "gmail.com"
- I say "oh, okay, I guess that's the problem" and try to log in with that username (no m)
- that successfully logs me in (my security questions show up), then it tries to take me to their internal website...and that fails, saying bad user/pass.

My guess would be that one of their internal systems has some kind of character limit on usernames that another system does not, and I got put into these two different systems with two different usernames, such that they cannot communicate with each other about my identity.

I spent half an hour on the phone with these idiots (because their website has no electronic support options listed) and got absolutely nowhere before I had to leave the room I only booked for 30 minutes, which I have to do to talk to them because their phone support ends at 5pm PT, weekdays only. They couldn't even ****ing reset my password without hitting an error (which obviously I did on my own, but phone support's script doesn't have room for "hi, I'm a programmer and I already did all the **** that you're about to tell me to do").

We spend proly 20x the time going over cards so that everyone understands them. It's brutal

Is it really true that bash has no out of the box way to pass named parameters from a parent script to a script it's calling? That seems crazy.

Right now I am adding this to any of my scripts that are called by other scripts:

Beautiful language. So expressive. I have no idea what it's doing but it works.

Don't you just call it the way you would run the target script from the command line?