|
The way the sentence read and trailed is amusing wrt it's content. The point of CI and CD is faster feedback and testing and catching problems ahead of time. Same reasons we pick strong and static type systems. I don't understand places that throw $2500 laptops at 30+ developers and manage to find a $500 box of crap for the system they will ALL use and possibly get blocked on constantly.
|
# ? Aug 1, 2017 14:38 |
|
|
# ? May 13, 2024 11:01 |
|
Vulture Culture posted:I think all developers should have the fastest CI money can buy, and painfully slow testing servers that randomly drop, duplicate or mis-order packets, lose connections, and stop responding to requests at odd intervals https://jagt.github.io/clumsy/
|
# ? Aug 1, 2017 16:04 |
|
This, but it also has a dying hard drive that occasionally takes upwards of a minute to service random I/O requests
|
# ? Aug 1, 2017 20:08 |
|
Vulture Culture posted:I think all developers should have the fastest CI money can buy, and painfully slow testing servers that randomly drop, duplicate or mis-order packets, lose connections, and stop responding to requests at odd intervals I resent the implication that our software doesn't have enough problems with its own intrinsic failures when run in an otherwise perfect universe.
|
# ? Aug 1, 2017 23:56 |
|
I've been at my current company for near 5 months as a senior dev. I feel like I've been able to contribute to the team so far, but have had a lot of questions getting up to speed with the current system since it's so huge and has a lot of legacy code. My tech lead is a nice guy, but I'm not sure of what he thinks of my skills. Today we had a monthly 1-1 meeting. We chatted about bs for an hour or so, and when we were wrapping up, he brings up the fact that our department is getting a new boss and to make sure I come in on and leave on time and that he was telling that to everyone, not singling me out. Then, he asks me for a copy of my resume, whenever I can send it over, saying that since he was away when I was interviewed/hired, he never actually saw it. I think this is weird. Anyone ever had this happen? I bring up him mentioning coming in on time because it seemed like he was saving the 'bad' or serious things for the end of the meeting and if he wanted a copy of my resume or talk more about what I did at my last job, why not just ask me during the hour we were bsing? I know it could mean nothing, but it's kinda made me paranoid and not sure if I'm worrying about nothing and he in fact just wants to look at my resume because he's curious, or if it could mean I could be out of a job soon. Anyone have any thoughts?
|
# ? Aug 2, 2017 00:46 |
|
The end of the conversation's a natural place to put a request for action. Maybe the new boss asked for everyone's resumes so they could get a brief introduction to their new personnel.
|
# ? Aug 2, 2017 00:57 |
|
I have been asked for my resume by new bosses before. I just imagined that they just want to see what are everyone's experience, nothing more than that. Maybe their boss is asking them about"who in your team can do X". I never gave it more thought than that.
|
# ? Aug 2, 2017 01:00 |
|
The department is huge so the new boss definitely isn't asking for my resume, the current tech lead (who I report to) asked me for my resume. I just found it weird that I've been there for 5 months, working with him this entire time, and all of a sudden he's asking for it. I guess it's nothing but I just found it odd and have been thinking about it.
|
# ? Aug 2, 2017 02:08 |
|
You're going to interview for your own job, but on company time so
|
# ? Aug 2, 2017 02:29 |
|
Vulture Culture posted:I think all developers should have the fastest CI money can buy, and painfully slow testing servers that randomly drop, duplicate or mis-order packets, lose connections, and stop responding to requests at odd intervals I had to swap out our local office test system using decent machines stacked on someone's desk for the then very new and shiny corporate Jenkins system a while back. I can understand the desire for sharing resources for better utilization and freeing devs from sysadminning dedicated test machines, and it worked out in the end, but at the time the enterprise Jenkins was used for nice little dockerized webapps and I was throwing a big honking multiplatform C++ renderer at it. Slaves were constantly running out of space or disconnecting and our very cpu intensive path tracer was tested on virtualized dual cores that may or may not have multiple other clients at the same time, but that didn't matter because our performance regression graphs on the test reports were rendered in javascript and we had no script permissions on the new system for "safety" reasons. I gnashed my teeth a lot that sprint.
|
# ? Aug 2, 2017 06:24 |
|
Volguus posted:The easiest thing to do is to not handle anything. Everything is UTC, until the moment of display. When sending it back, convert to UTC in the client. Then simply not worry about it. Otherwise you'll go insane. Usually the trouble we run into is when the source data is just a date with no time. A lot of the date/time libraries only have a time stamp type and will treat that as 0:00 UTC which then might actually be the previous date when converted into someone's local time zone. Which is generally not what we want. Imagine displaying someone's birthday as the wrong date, for instance.
|
# ? Aug 2, 2017 12:01 |
|
smackfu posted:Usually the trouble we run into is when the source data is just a date with no time. A lot of the date/time libraries only have a time stamp type and will treat that as 0:00 UTC which then might actually be the previous date when converted into someone's local time zone. Which is generally not what we want. Imagine displaying someone's birthday as the wrong date, for instance. When you deal with "source data", you do what you have to do, no other choice. My advice was when you control the full chain.
|
# ? Aug 2, 2017 14:27 |
|
It's easy enough to say "just store everything in UTC", but there's a massive amount of overhead doing so when you're dealing with multiple layers and making sure the times come in and out the right way, it seems that it's overly difficult to pull off when dealing with multiple layers and frameworks. Especially when you have to go on a field-by-field basis to determine what specifically you want to format, like the birthday example. As far as I can tell, there are three cases: - Dates that shouldn't change with timezone (birthdays, historical dates, etc.) - Times that shouldn't change with timezone (set my alarm for 7am every morning - this should still be 7am if i fly to California, and should still remain 7am when DST rolls around) - Times that do change with timezone (this conference call with us and our UK affiliate is at 4 EST) or are used in actual math (how many hours has it been since <thing> started)? Sure I can store 3:45pm UTC in the database, but I still have to tell Dapper that all datetimes coming out are UTC, tell my mapping layers and serializers to do formatting and conversions, and get Knockout and Moment on the frontend to either localize or not localize the dates based off the above. As someone mentioned earlier, the tooling does seem to be bullshit because none of the above frameworks makes doing that easy. I've got it all worked out, but I do feel like someone living in black and white infomercial land, thinking "there has to be a better way!!"
|
# ? Aug 2, 2017 19:20 |
|
Programmers should start a lobby to abolish daylight savings time and time zones.
|
# ? Aug 2, 2017 19:30 |
|
fantastic in plastic posted:Programmers should start a lobby to abolish daylight savings time and time zones. Please and thank you.
|
# ? Aug 2, 2017 20:00 |
|
Swatch Internet Time
|
# ? Aug 2, 2017 20:06 |
|
then you will still have to deal with non-abolished and abolished DST regions forever and ever and ever
|
# ? Aug 2, 2017 20:09 |
|
curufinor posted:then you will still have to deal with non-abolished and abolished DST regions And/or any time and date comparisons across the moment of changeover.
|
# ? Aug 2, 2017 20:13 |
|
Just store it with the timezone that it was intended to be saved with, let the user pick, or assume it.
|
# ? Aug 2, 2017 20:31 |
|
MisterZimbu posted:- Times that do change with timezone (this conference call with us and our UK affiliate is at 4 EST) or are used in actual math (how many hours has it been since <thing> started)? The international appointment booking is a fun one because time zones change too. Not just daylight savings, sometimes a country picks an entirely different time zone (Samoa). You need to include where the event will be, so if the time zone changes it's still correct for local time. Probably better to include GPS coordinates too in case the country changes shape or stops existing.
|
# ? Aug 2, 2017 22:22 |
|
Abolish time zones in half the countries of the world, then reinstate them in a different, incompatible way 25 years later.
|
# ? Aug 3, 2017 00:15 |
|
I eagerly await the addition of leap time zones, and the complexities that will come with keeping track of them.
|
# ? Aug 3, 2017 00:22 |
|
suggestion: get actual humans to use unix timestamps
|
# ? Aug 3, 2017 00:43 |
|
suggestion: count down time till the unix epoch end at 2038
|
# ? Aug 3, 2017 00:44 |
|
suggestion: count down time till that mayan apocalypse date, so we'd be year negative 5 right now
|
# ? Aug 3, 2017 00:45 |
|
Time was a mistake. Users should have to pick from an enumerated list of approximate values. Nowish, later, and eh.
|
# ? Aug 3, 2017 01:21 |
|
Any attempt to simplify time further than what it is already is probably doomed to failure. It's already a rounded abstraction of the rotation of the earth and then the orbit of the earth. If we ever get off this rock then suddenly our concepts of hours and years will be broke for any non earth planet, and then the next level of confusing localisation will begin. (which is essentially what time zones are, localisation that you can't fob off as easily as language and formatting localisation)
|
# ? Aug 3, 2017 02:34 |
|
Maluco Marinero posted:Any attempt to simplify time further than what it is already is probably doomed to failure. It's already a rounded abstraction of the rotation of the earth and then the orbit of the earth.
|
# ? Aug 3, 2017 04:55 |
|
MisterZimbu posted:It's easy enough to say "just store everything in UTC", but there's a massive amount of overhead doing so when you're dealing with multiple layers and making sure the times come in and out the right way, it seems that it's overly difficult to pull off when dealing with multiple layers and frameworks. Case #2 - no conversion or manipulation occurs in the DB, backend, or frontend Case #3 - the DB queries and backend never touch the timestamp, the API exposes it as seconds UTC (an Integer), and any manipulation happens at the very last second by moment.js and nothing else. That's how I do it and it's pretty easy edit: part of the trick is using either String or Integer datatypes in the API. This prevents having to convert the timestamp into a Date object which is one possible layer where problems can happen. My Rhythmic Crotch fucked around with this message at 16:55 on Aug 3, 2017 |
# ? Aug 3, 2017 16:48 |
|
In case you guys haven't seen it yet : https://www.troyhunt.com/introducing-306-million-freely-downloadable-pwned-passwords/
|
# ? Aug 3, 2017 19:12 |
|
AskYourself posted:In case you guys haven't seen it yet : quote:Request denied by WatchGuard HTTP proxy. Oh for fucks' sake work firewall get over yourself.
|
# ? Aug 3, 2017 20:40 |
|
Munkeymon posted:Oh for fucks' sake work firewall get over yourself. My work just put up a new firewall and all the white papers I used to reference from anything tangentially games industry related are blocked. It's been a fun couple weeks of back and forth with IT.
|
# ? Aug 3, 2017 21:35 |
|
leper khan posted:My work just put up a new firewall and all the white papers I used to reference from anything tangentially games industry related are blocked. It's been a fun couple weeks of back and forth with IT. I'm remoting into my home desktop to because SA is blocked under the category 'Tasteless' which I'd have a hard time arguing with, so I haven't brought it up
|
# ? Aug 4, 2017 13:58 |
|
Munkeymon posted:I'm remoting into my home desktop to because SA is blocked under the category 'Tasteless' which I'd have a hard time arguing with, so I haven't brought it up I’m getting close to drastic measures like that. Office has only one Ethernet port per desk. The 802.WHATEVER security will kill the port if more than one MAC address is scene from it. Pinging from the Ethernet to a Mac-mini on my desk on wifi was giving ~120ms MEAN latency. Oh and I’m working on a low latency networking App with 3 machines 2 of which need to be on wifi. They offered to move the mini to the server room but I need it’s monitor as well.
|
# ? Aug 4, 2017 14:23 |
|
Hughlander posted:I’m getting close to drastic measures like that. Office has only one Ethernet port per desk. The 802.WHATEVER security will kill the port if more than one MAC address is scene from it. Pinging from the Ethernet to a Mac-mini on my desk on wifi was giving ~120ms MEAN latency. Oh and I’m working on a low latency networking App with 3 machines 2 of which need to be on wifi. Can't you put a cheap NAT router in front of it?
|
# ? Aug 4, 2017 16:33 |
|
Munkeymon posted:Can't you put a cheap NAT router in front of it? It'd need to authenticate against AD to even get a DHCP address and probably be a fireable offense if discovered. I think instead I'm going to grab some USB ethernet dongles and make a P2P network. Still probably a violation of policy but not to the point of allowing unauthenticated devices on the corporate network.
|
# ? Aug 4, 2017 17:01 |
|
Yes, that's totally not insane at all. How does management react when you bring that up?
|
# ? Aug 4, 2017 18:47 |
|
The solution is to use your workstation as a proxy server on a virtual bridged network and force your child systems to use the host as a proxy machine. This is how a lot of people run VMs on their local machines with a single IP and no MAC address besides the one assigned to their host machine is seen on the physical network. Now, someone can oftentimes tell if you're using NAT by analyzing the source ports for certain behaviors, but this tends to have a high false-positive rate without checking against OS types. For example, Windows only uses a certain range of ephemeral ports and Linux has a different set, and if SNMP detects your workstation is Windows but is seeing a lot of ephemeral ports overlapping with Linux, you're going to be found. Also, macOS has VNC built-in via Screen Sharing. It can be a bit flakey admittedly (seems to get stuck and I'll need to wait for the daemon to reset or whatever), but it beats nothing.
|
# ? Aug 4, 2017 23:16 |
|
Or you could just work whilst at work, instead of posting on these here forums
|
# ? Aug 5, 2017 03:03 |
|
|
# ? May 13, 2024 11:01 |
|
My Rhythmic Crotch posted:Or you could just work whilst at work, instead of posting on these here forums Regression testing with 0 defects on my track and 1 week before my last day. Leaving at 2pm. But making it to 2pm is a challenge, heroic even.
|
# ? Aug 5, 2017 03:41 |