|
freeasinbeer posted:Do you have a ton invested in the Jenkins ecosystem? If not I’d look at using something else. Jenkins is a nightmare. What about Jenkins is a nightmare? We use it for everything at my company
|
#
?
Jun 23, 2018 01:25
|
|
|
|
| # ? May 12, 2024 22:41 |
|
|
freeasinbeer posted:Do you have a ton invested in the Jenkins ecosystem? If not Id look at using something else. Jenkins is a nightmare. I gently caress to cum
|
|
#
?
Jun 23, 2018 07:40
|
|
|
I just volunteered to do scrummaster duties at work along with my regular development work. How hosed am I?
|
|
#
?
Jun 23, 2018 14:20
|
|
|
Slimy Hog posted:What about Jenkins is a nightmare? We use it for everything at my company I wouldn't say that there is one big fatal flaw in Jenkins, so much as a whole lot of small issues and annoyances that build up over time. I once helped transition a team from Jenkins to Teamcity, to bring it in line with the client's standard tools and processes, and found myself really liking the latter better.
|
|
#
?
Jun 23, 2018 14:35
|
|
|
bvj191jgl7bBsqF5m posted:I gently caress to cum What?
|
|
#
?
Jun 23, 2018 14:48
|
|
|
Pollyanna posted:What? He fucks to cum
|
|
#
?
Jun 23, 2018 14:51
|
|
|
Phobeste posted:He fucks to cum And that's his investment in Jenkins. I mean, maybe there's a plug-in that works with one of those robotic fleshlights to reward good commits that he developed?
|
|
#
?
Jun 23, 2018 15:04
|
|
|
Now that I think about it I think it's a glib way of referring to the sunk cost fallacy.
|
|
#
?
Jun 23, 2018 15:35
|
|
|
BurntCornMuffin posted:And that's his investment in Jenkins. the emacs plugin that speeds up a usb connected vibe with nesting depth except for lines of groovy code or number of plugins
|
|
#
?
Jun 23, 2018 16:50
|
|
|
Chron posted:I just volunteered to do scrummaster duties at work along with my regular development work. How hosed am I?
|
|
#
?
Jun 23, 2018 16:53
|
|
|
I legit enjoy management work. I enjoy being a manager, and it's always weird to me how much people here despise it. 
|
|
#
?
Jun 23, 2018 17:30
|
|
|
I'm constantly amazed how much harder management has been than being an individual contributor. The amount of things that are completely out of my control that I have to handle and take responsibility for and the compromises I have to make day-to-day often makes my head spin.
|
|
#
?
Jun 23, 2018 17:33
|
|
|
I don't know how much people outright hate management. I think people hate having to manage and develop. Hell, I'm not really in the position myself but I even find if I have one of those days with a bunch of meetings that I switch to extrovert mode and find it hard to sit down and bang on something. When that instead gets scattered across the whole day, there goes any chance of me working on something alone.
|
|
#
?
Jun 23, 2018 18:49
|
|
|
Chron posted:I just volunteered to do scrummaster duties at work along with my regular development work. How hosed am I? If you can keep people on track for 15 minutes each day, you're good.
|
|
#
?
Jun 23, 2018 22:25
|
|
|
Rocko Bonaparte posted:I don't know how much people outright hate management. I think people hate having to manage and develop. Hell, I'm not really in the position myself but I even find if I have one of those days with a bunch of meetings that I switch to extrovert mode and find it hard to sit down and bang on something. When that instead gets scattered across the whole day, there goes any chance of me working on something alone. Some people are better suited than others. I don't want to become a people manager, it's going to be terrible for everyone.
|
|
#
?
Jun 24, 2018 15:50
|
|
|
Rocko Bonaparte posted:I don't know how much people outright hate management. I think people hate having to manage and develop. Hell, I'm not really in the position myself but I even find if I have one of those days with a bunch of meetings that I switch to extrovert mode and find it hard to sit down and bang on something. When that instead gets scattered across the whole day, there goes any chance of me working on something alone. Talk to your manager, and look into blocking some time on your calendar if you need to dedicate some time to coding to improve your quality of life. When your own dev time is limited, you need to pick your battles a bit and be active in your time management, otherwise a rogue product/pm team will soak up every last second of your free time. I'm in a role where I'm mostly managing, but still have some opportunity to be hands on in terms of tool building - i.e. local environments for dev work, content tools, little scripts to help with setup or teardown, testing tools, etc. I pretty much keep a few open windows in mornings/afternoons for meetings, so if project management/etc need some time to discuss anything short on notice, there's still some availability. My time is best spent working with lead developers and keeping them unblocked and moving forward on their work and their team's work - so when I'm writing any code, it's to give the senior teams some helpful tools to make THEIR lives better. If my teams are happy, I have less hands on the critical pieces of work, and I don't have to worry about being the one blocking progress. Don't make all of this fall upon your own shoulders. Keep a schedule, work with your management to address concerns around time and the best usage of your time, and work actively to improve. Time spent in meetings is usually pretty ingrained in your team culture, so it may take some time, but it's something you can have active control over when you can communicate and get buy-in from your management chain.
|
|
#
?
Jun 24, 2018 19:02
|
|
|
Peggle Fever posted:Talk to your manager, and look into blocking some time on your calendar if you need to dedicate some time to coding to improve your quality of life. When your own dev time is limited, you need to pick your battles a bit and be active in your time management, otherwise a rogue product/pm team will soak up every last second of your free time. Are you me? Because this is what I do in software management too. My reward is that I get to go into work not dreading the day and/or not have everything be on fire all of the time.
|
|
#
?
Jun 25, 2018 16:24
|
|
|
Man I really thought the shift from Java to Scala was going to be easier. Slick is turning my brain into goo
|
|
#
?
Jun 25, 2018 19:45
|
|
|
ChickenWing posted:Man I really thought the shift from Java to Scala was going to be easier. Slick is turning my brain into goo Wtf does Slick have to do with Scala? Anyway, the switch from Java to Scala wasn't easy for me too. It took me a couple months before I knew what I was doing in Scala. And the only way I managed it was by being surrounded by very capable Scala developers. The different way you need to think about functional programming as compared to imperative really takes some practice, it isn't at all comparable to switching between to imperative languages. I've heard of a couple people who gave up halfway, and I can understand that, if you didn't have an environment as supportive and helping as mine.
|
|
#
?
Jun 25, 2018 20:15
|
|
|
ChickenWing posted:Man I really thought the shift from Java to Scala was going to be easier. Slick is turning my brain into goo We have a legacy codebase full of slick 2 and it's soured me to ORMs entirely. I love me a scala but when it goes bad, it goes bad.
|
|
#
?
Jun 25, 2018 21:04
|
|
|
Boss is refactoring our error process for what I think is the 6th time. If I had to care, I'd be so irritated because the thing he's all up about is the result of his constantly changing directions. I feel such internal peace because I do not give a single solitary gently caress.  I'm also laughing because he constantly complains about how there is never enough time and the team needs to sacrifice their personal lives to get the needful done, but he's taken up about 3-4 hours today lecturing the team on things he's "found out" and new patterns and whatnot he's insisting on now. Shirec fucked around with this message at 21:48 on Jun 25, 2018 |
|
#
?
Jun 25, 2018 21:36
|
|
|
Carbon dioxide posted:Wtf does Slick have to do with Scala? Slick is the most popular Scala SQL access library. It is also not an ORM, which didn't really sink in for me for the first couple days of work. Clanpot Shake posted:We have a legacy codebase full of slick 2 and it's soured me to ORMs entirely. I love me a scala but when it goes bad, it goes bad. We are using slick 3 and it's a microservice so luckily I don't have anything bad, it's just that the service's repository layer was definitely designed by someone who didn't understand transactions (doing select queries transactionally  ) and who doesn't seem to have looked up slick practices at all (based on what I've read about slick, the repository layer is just for storing queries and returning database operations to be executed at the service layer ??????????????????????????????????????)brain made of goo send help
|
|
#
?
Jun 25, 2018 22:57
|
|
|
ChickenWing posted:Slick is the most popular Scala SQL access library. It is also not an ORM, which didn't really sink in for me for the first couple days of work. Ah, okay. I haven't interacted with SQL directly from Scala. When I googled Slick some sort of jQuery thing turned up, which is why I got confused.
|
|
#
?
Jun 25, 2018 23:43
|
|
|
Shirec posted:Boss is refactoring our error process for what I think is the 6th time. If I had to care, I'd be so irritated because the thing he's all up about is the result of his constantly changing directions. I feel such internal peace because I do not give a single solitary gently caress. What are the patterns?
|
|
#
?
Jun 26, 2018 00:22
|
|
|
Rocko Bonaparte posted:What are the patterns? Let me see if I can hit some major ones. This is mostly the Node back end stuff. We have our back end, besides our messaging tier and whatnot, broken out into APIs that are very simple and the management classes that do the complicated stuff and interact directly with the database. For the first 3 months of re-working the back end, we were only doing the management class portion and their unit tests. Pattern 1: all calls to data management should be able to handle 1 - n number of things, regardless of what call it is. That meant any error patterns needed to tell you what number of the record it was that had the error. So errors would have arrays of JavaScript error objects. Pattern 2: all calls to data management, besides a get all, should only handle 1 thing at a time. The error patterns mostly were not updated to reflect this, and some of the older management classes were not changed at all, due to time constraints. The unit tests got more complicated and were taking longer and longer to run (and thus debug). Pattern 3 I was not as involved with as I was writing a separate server thing (that also had several rewrites that was the big thing my boss never stopped terrorizing me about). It mainly had to do with taking out a lot of validations we had required so it could be more flexible and re-writing a lot of function structure. Old pattern 1 and 2 stuff still remains throughout. I did create the official error logging though during this period and was told that I had to take in "an error object, a code, a message, a string, or nothing" and log it to our database. APIs were being written and both 200 and 400 responses could return errors, based on what got tripped over. Pattern 4: Supposed to be the new "gold standard". Deeply deeply validated, had rigid controls over all the failure points and had both optional transactions and non-transaction versions of all "public" methods. Error methods were still JavaScript objects but also had manual "codes" added to them that had supposed patterns (which caused problems because I'd have code 2003 fail before code 2004 and there was not a logical way to change it). Started failing because database returned date times such that it couldn't be validated down to the millisecond. Here is where unit tests became "individual" because boss got mad that it took him so long to run the tests. Got mad at us for not bringing it up sooner. Pattern 5: Similar to pattern 5 except only happy path. Error patterns then were abandoned again. Pattern 6: What they were discussing today. I mostly ignored them, but now they are breaking stuff into validation and errors. One will be the manual stuff that is thrown back, the others will be actual legitimate errors. I was writing the front end to handle this and was told to expect validation only in 200 responses, and it will have a code and message property. 400 responses will be "mysteries". This could also be pattern 7 because maybe that changed again Shirec fucked around with this message at 00:47 on Jun 26, 2018 |
|
#
?
Jun 26, 2018 00:45
|
|
|
Shirec posted:Let me see if I can hit some major ones. This is mostly the Node back end stuff. We have our back end, besides our messaging tier and whatnot, broken out into APIs that are very simple and the management classes that do the complicated stuff and interact directly with the database. For the first 3 months of re-working the back end, we were only doing the management class portion and their unit tests. Has he ever heard of exceptions?
|
|
#
?
Jun 26, 2018 00:54
|
|
|
Rocko Bonaparte posted:Has he ever heard of exceptions? Oh sorry, I'm probably using using these terms wrong? I think we use them interchangeably but it might also be due to my lack of experience  He generally doesn't like to think of anything other than very rare JavaScript errors happening and maybe something else really really rare.edit: Now I'm thinking I'm dumb Shirec fucked around with this message at 00:59 on Jun 26, 2018 |
|
#
?
Jun 26, 2018 00:57
|
|
|
Shirec posted:edit: Now I'm thinking I'm dumb
|
|
#
?
Jun 26, 2018 01:08
|
|
|
Shirec posted:Boss is refactoring our error process for what I think is the 6th time. If I had to care, I'd be so irritated because the thing he's all up about is the result of his constantly changing directions. I feel such internal peace because I do not give a single solitary gently caress. It's so beautiful, watching someone transition into the best part of a bad job.
|
|
#
?
Jun 26, 2018 01:26
|
|
|
Shirec posted:I feel such internal peace because I do not give a single solitary gently caress. 
|
|
#
?
Jun 26, 2018 02:11
|
|
|
Shirec posted:He generally doesn't like to think of anything other than very rare JavaScript errors happening and maybe something else really really rare. Ah, the old "the network is reliable" chestnut. You'll be surprised to learn that your boss is in fact an idiot.
|
|
#
?
Jun 26, 2018 03:58
|
|
|
Rocko Bonaparte posted:Has he ever heard of exceptions? He probably doesn't understand checked vs unchecked exceptions.
|
|
#
?
Jun 26, 2018 04:53
|
|
|
Shirec posted:edit: Now I'm thinking I'm dumb Everybody here is dumb. Your boss is dumber. I'm pretty sure I followed along with your boss' framewreck because I've dealt with this before. Let's see how I do (trigger warning): Shirec's boss posted:Pattern 1: all calls to data management should be able to handle 1 - n number of things, regardless of what call it is. That meant any error patterns needed to tell you what number of the record it was that had the error. So errors would have arrays of JavaScript error objects. quote:Pattern 2: all calls to data management, besides a get all, should only handle 1 thing at a time. The error patterns mostly were not updated to reflect this, and some of the older management classes were not changed at all, due to time constraints. The unit tests got more complicated and were taking longer and longer to run (and thus debug). quote:Pattern 3 I was not as involved with as I was writing a separate server thing (that also had several rewrites that was the big thing my boss never stopped terrorizing me about). It mainly had to do with taking out a lot of validations we had required so it could be more flexible and re-writing a lot of function structure. Old pattern 1 and 2 stuff still remains throughout. I did create the official error logging though during this period and was told that I had to take in "an error object, a code, a message, a string, or nothing" and log it to our database. APIs were being written and both 200 and 400 responses could return errors, based on what got tripped over. quote:Pattern 4: Supposed to be the new "gold standard". Deeply deeply validated, had rigid controls over all the failure points and had both optional transactions and non-transaction versions of all "public" methods. Error methods were still JavaScript objects but also had manual "codes" added to them that had supposed patterns (which caused problems because I'd have code 2003 fail before code 2004 and there was not a logical way to change it). Started failing because database returned date times such that it couldn't be validated down to the millisecond. Here is where unit tests became "individual" because boss got mad that it took him so long to run the tests. Got mad at us for not bringing it up sooner. "Let's make each of our functions a thousand lines long with all our copy pasted bullshit taken to its illogical conclusion. The if-else code we're copying and pasting after each of these calls is getting complicated so I'm going to add an error number variable. Instead of testing against magic error strings, we'll check against magic numbers instead that are probably randomly slapped somewhere as constants. We'll do this after each call. It's like a giant switch block. So robust! Whenever we change it, you just copy and paste it from my code everywhere. You're an idiot for not understanding that the 1000-code errors are totally different from the 2000-code errors, which are completely different from the 2100-code errors. What's the problem with code 3001 representing an HTTP 404?" quote:Pattern 5: Similar to pattern 5 except only happy path. Error patterns then were abandoned again. quote:Pattern 6: What they were discussing today. I mostly ignored them, but now they are breaking stuff into validation and errors. One will be the manual stuff that is thrown back, the others will be actual legitimate errors. I was writing the front end to handle this and was told to expect validation only in 200 responses, and it will have a code and message property. 400 responses will be "mysteries". This could also be pattern 7 because maybe that changed again quote:He generally doesn't like to think of anything other than very rare JavaScript errors happening and maybe something else really really rare. I recently had to deal with a fresh crew of REEWDASOC (Random Electrical Engineers Who Did A Semester Of C) with Python and they got into the typical trap of returning magic number codes from everything. This would take over the real return value for their stuff so they'd devolve into tuples and poo poo like that. The problem with that code is after each drat call, you then have to go through and interrogate the function to tell what happened. A 10-line routine becomes 200 with most of that code copy-pasted "if this stuff, return fail" junk. I eventually had one of them on a Skype meeting, opened up the drawing tool, drew a stick figure of them, the word "exceptions" and then a big heart around it. "This is you, this is exceptions, and now you're married." When something goes wrong, it percolates up to some place where somebody actually cares--or even simply can care--to do something with it. Hell if multiple things went wrong then you can create a ImPissedAtYourForAThousandReasonsException and put all the problems in there. Then the code just returns back what it should do when its running normally. Their response was to call me a mad scientist with crazy ideas.
|
|
#
?
Jun 26, 2018 05:21
|
|
|
Rocko Bonaparte posted:Their response was to call me a mad scientist with crazy ideas. Sometimes followed up with "Why is nothing improving, what is our specialist doing?" The confusion that a technical specialist can change an unwilling organisation is terrible to behold. "What do you mean you need mandate? Can't you just use your influence people using your powers of persuasion? What do you mean you do not have that skill? You said you were a Senior software engineer!" Last week I told the PO I was very happy with her as she was a never ending source of fun puzzles to solve. She did not understand or at least pretended not to. Yesterday there were four of us around a screen discussing a feature when someone said how nice it would be to have finalized specs before starting development. Then I said that would be fine but that changes to the spec would only be approved using a change request with the same level of approval. And that was the end of that.
|
|
#
?
Jun 26, 2018 07:00
|
|
|
My boss: You need to extensively test this software you wrote which sends out texts when cars are stolen, it's too verbose. Also my boss, on the previous friday: We need something that can alert us when cars are stolen and we need it today!
|
|
#
?
Jun 26, 2018 07:18
|
|
|
Keetron posted:Ah yes, the old "Let's hire a specialist to solve this problem we cannot solve ourselves and then dismiss all his solutions as crazy and continue doing things the old way."
|
|
#
?
Jun 26, 2018 11:35
|
|
|
Boiled Water posted:My boss: You need to extensively test this software you wrote which sends out texts when cars are stolen, it's too verbose. Your boss: gently caress your weekend, do this NOW! We are not slaves.
|
|
#
?
Jun 26, 2018 13:39
|
|
|
Rocko Bonaparte posted:Eerily accurate Wow, that's just.... uh, like spot on in most places? Like I think some things are almost verbatim quotes of how he explained his choices. The only plus to that is that the "codes" were supposed to be able to be coded against, like you said in a giant switch statement, but ALSO to get translated by our front end thing. Which blah blah, so much easier, but gently caress him, it was still me manually coding against each error statement to look up what the Spanish translation of it was. Now he's planning on coding a reporting system for our data processor, something I had asked and pushed for, but there wasn't enough "business value". Now there is that I'm leaving. Pollyanna posted:Your boss: gently caress your weekend, do this NOW! It's amazing how some places really don't agree with this statement. Salaried? Basically you're mine forever now!
|
|
#
?
Jun 26, 2018 15:04
|
|
|
We have a situation at work and I dont know how to explain it in a way that properly shows why its stupid. We have many Agents. An Agent can be Expired, Inactive, or Active. Inactive and Active represent their status with some optional feature we have. We get a (really crappy) CSV from a company listing which of their agents are still with them, and if theyre eligible for that optional feature. Our task, which has been in testing for like 2 months now, is to keep our records up to date based on this CSV they send us. Im of the opinion that we shouldnt be doing this at all and instead we should just give them an endpoint that lets them manage their agents directly on our platform because trying to keep our records in sync with their CSV is a pain in the rear end and has proven to be really touchy, brittle, and unreliable. Am I just being picky/lazy when I say it sucks, or is it genuinely a bad idea?
|
|
#
?
Jun 26, 2018 15:15
|
|
|
|
| # ? May 12, 2024 22:41 |
|
|
Pollyanna posted:We have a situation at work and I don’t know how to explain it in a way that properly shows why it’s stupid. Lol, yes, that sounds stupid. At the very least I hope there's some automation that sticks the csv into your database. If you're doing a process like that manually it's a waste of money and resources.
|
|
#
?
Jun 26, 2018 15:22
|
|
