The Official Status Thread
-
@cartman82 said in The Official Status Thread:
The installer IS pretty neat. I just don't understand why the complication with graphical UI, when 99% of users will run servers.
Quite a few commercial applications have graphical installers (“because they can!”). Examples include the Oracle DB and ANSYS's FEM/CFD solvers. If you're doing anything other than running them as toys, you won't care about running those codes with a GUI attached (and will probably want the memory assigned to running the business end of those packages).
-
STATUS:
Contractor is handling the deployment of customer's website on the brand new Centos VM I have handed him over:
-- OK, Cartman, you can switch over the website domain now.
-- Oh, you are done already?
-- Nope, but I figure, the domain won't switch over immediately.
-- Umm... I think... I'll leave that in your capable hands. Call me if you need anything.(steps away slowly)
-
@cartman82
I mean, depending on the domain's TTL settings, he's not entirely wrong.But man, is he a moron.
-
@cartman82 here it is:
- customer need a functionality X
- create branch for X
- code X
- ship it (no test and no merge)
I have many versions of the same things here.
-
Status: I just got out of a long meeting where I was in the shittiest chair in the conference room. Any time you leaned back on the backrest, the chair would lower.
-
@Polygeekery makes it easy to duck out of the discussions!
-
@cartman82 said in The Official Status Thread:
TIL:
Buying a dedicated server for virtualization is like getting a new shelf. You might not think you need all that space, but a month later it's full to the brim and you're strapped for space again.
Especially when the memory requirements to run are no longer in the 500mb range...
-
@Jaloopa said in The Official Status Thread:
@cartman82 said in The Official Status Thread:
what does that do with stack traces?
Keeps it as it was. If it was
catch (Exception e) { throw e }
the stack trace would be from the rethrow
Yeah, I eliminated that version, it looked like @Jaloopa 's before.
-
@groo said in The Official Status Thread:
@cartman82 here it is:
customer need a functionality X
create branch for X
code X
ship it (no test and no merge)I have many versions of the same things here.
Unfortunately, customers tend to expect that their special snowflake version will also receive all the updates and new features from the main branch.
-
@Polygeekery said in The Official Status Thread:
I was in the shittiest chair in the conference room.
How much did you wind up giving up?
-
@Jaloopa said in The Official Status Thread:
@cartman82 said in The Official Status Thread:
what does that do with stack traces?
Keeps it as it was. If it was
catch (Exception e) { throw e }
the stack trace would be from the rethrow
I'd put money its used for error handling!
STATUS all going well. Waiting for the piano drop!
-
Status: I love how Windows 10 does not have an intuitive way to create a local account during install. You have to tell it to join an Active Directory domain, which doesn't actually prompt you to join AD but instead allows you to create a local account with a note that "You can actually connect to AD sometime later."
-
@mott555 said in The Official Status Thread:
You have to tell it to join an Active Directory domain
Where do you even get this stuff? I mean, you're right about the "create a local account" being unintuitive, but if what you said were true, think of the trouble all the home users would have?
The people who brought you "press Start to shut down" have given us "choose to create a Microsoft account, and then say 'no thanks--I want a local account'".
-
@FrostCat This just happened about 10 seconds ago. I guess I should have taken a cell phone video.
The funny part is I actually wanted to join it to AD, but it said "Nope, this is labeled 'Join Active Directory' but it actually means 'Create a local account'. Joke's on you, haha!"
-
@Lorne-Kates said in The Official Status Thread:
I built a wall. Apparently now Mexico owes me $600 CAD
You just need to add little ladders. And GI-Joes. Or (and?) LEGO people.
-
@Lorne-Kates said in The Official Status Thread:
Windows 10 broke moving-a-window-with-the-keyboard.
Huh? I just moved Notepad 1/2 off the top of the screen. Of course, hitting enter or alt-tab snapped it back. (I think I'm missing the start of this "thread")
-
@cartman82 said in The Official Status Thread:
STATUS:
My morning ritual at work:
- Breakfast
- Coffee
- Bulk-delete error notifier emails from inbox
1 and 2 look backwards. Unless you start with coffee at home...
-
@Magus said in The Official Status Thread:
Status: Got in at 7:30 as usual. No one was here by 8, not too unusual. No one is here yet, at 9:06 - is going on!?
Maybe they all live on my block? My power was out from 3:30a (when I woke to beepbeepbeep) to 2:25pm yesterday.
-
Status: ugh
-
@mott555 said in The Official Status Thread:
This just happened about 10 seconds ago. I guess I should have taken a cell phone video.
It's not that I don't believe you, but you seem cursed as far as Win10 goes.
I thought you pretty much always had to set up a local account before joining a domain, but I've always used Home/Pro editions.
-
@groo said in The Official Status Thread:
@cartman82 here it is:
- customer need a functionality X
- create branch for X
- code X
- ship it (no test and no merge)
I have many versions of the same things here.
This reminds me of a "motivational" poster we have here. Anonymized a bit:
while(TRUE) { code(); ship(); improve(); }
Someone noticed an important step missing and wrote in an improvement to the algorithm. Unfortunately, they didn't take their own advice, and created the semantic equivalent of
while(TRUE) { code(); test(); continue; ship(); improve(); }
Filed under: Hey, guys! What happened to our revenue?
-
@HardwareGeek said in The Official Status Thread:
This reminds me of a "motivational" poster we have here. Anonymized a bit:
while(TRUE) { code(); ship(); debug(); }
FFT
-
OMFG. I hate IT.
Especially software developers.
-
@FrostCat said in The Official Status Thread:
but you seem cursed as far as Win10 goes.
Of course, it's not a problem with Micro-Soft's software, it is @mott555 who is cursed.
It's not like software should act the same for everyone.
-
@TimeBandit said in The Official Status Thread:
it is @mott555 who is cursed.
You have no idea. I have accomplished nothing this week. First I was chasing a bunch of bugs in our product, only to discover there's nothing wrong with the product and my HP workstation is just a piece of junk which doesn't really like PCI/PCI-E add-in cards anymore. So I switched over to my (15-year-old) Dell workstation and spent a bunch of time getting it set up, only to have some required software fail two hours into an install for who knows what reason. Then I grabbed an unused PC from our lab, but it had no network connection and the Windows 7 NIC drivers from the manufacturer refused to install on Windows 7 because it's an "unsupported operating system" . Then I spent the day wiping that system and re-installing Windows 10 and am still doing updates and installs.
-
@Weng We have MSDN accounts you fucking morons, why are you discussing at the top of your lungs how to pirate Windows 10?
-
@Weng said in The Official Status Thread:
@Weng We have MSDN accounts you fucking morons, why are you discussing at the top of your lungs how to pirate Windows 10?
HEY WE CAN PIRATE WINDOWS 10 BY TAKING OUR MSDN SUBSCRIPTION AND THEN DRAWING A LITTLE EYEPATCH AND PEG LEG ON IT
-
Status: In a meeting discussing how we're going to mitigate incompatibility issues moving from .Net 4.0 to 4.6.1.
Saw this in our "application listing":
Here's a hint guys: VB6 is NOT .Net. At all.
At least someone acknowledged that.
Edit: And someone paid attention to my response! Yay!
-
Status: Spent this afternoon unfucking the network configuration of one of our mass spectrometers, all caused thanks to the “kind help” of the installation engineers who fitted a new Liquid Chromatography column to the system a few weeks ago:
- Guess what, our systems don't use fixed IP addresses (or rather they do, but it's a policy set in our DHCP management layer) so putting a fixed network config in won't work well.
- Not everything should be firewalled totally by default. We want to actually be able to pull data off this system (to our archiving systems) for some reason!
- Linked to that, the relevant part of the disk should be shared (because that's the most practical way, beating slapping an FTP server on, enabling webdav, or dicking around with running sshd on Windows).
- The password was changed. Except in the parts of the code that were trying to mount the shares, of course. Nobody had thought to tell them. (Or me before today.)
All stupid trivialities. All fixed. The
spicedata must flow!
-
@Tsaukpaetra said in The Official Status Thread:
how we're going to mitigate incompatibility issues moving from .Net 4.0 to 4.6.1.
End result of the meeting: "WE NEED TESTING OMG MANAGERS AREN'T TESTING WTFARGBLABARG!!!"
To the forum: Has anyone seen core application breaks between running a 4.0-compiled application running on a 4.0 installation versus running a 4.0-compiled application running on a 4.6.1 installation?
This isn't counting recompiling 4.0 applications to 4.6.1 (That I know there are some potential issues), this is running applications that haven't been recompiled to the one supposedly installed on the client.
-
@Tsaukpaetra We ran into some kind of library incompatibility between 4.5 and 4.6, fortunately we had source for the library and could recompile it for 4.6. I don't recall the issue because it's not a project I'm involved with.
-
@Tsaukpaetra said in The Official Status Thread:
Has anyone seen core application breaks between running a 4.0-compiled application running on a 4.0 installation versus running a 4.0-compiled application running on a 4.6.1 installation?
You don't want a theoretical answer — whether some shitty code could break — you want to know about whether that actually happens in the real case you're dealing with. Heck, bad code could break even if everything is supposedly OK because it is peeking at the version of the runtime and trying to do something “smart”. I've seen that sort of stupid in other places, it wouldn't surprise me to see it again. What matters is not whether some wiseguy has done a crime against programming, what matters is whether you've got one of those crimes about to detonate in a steaming pile of WTF on your desk.
There's no substitute for testing.
-
@FrostCat said in The Official Status Thread:
@mott555 said in The Official Status Thread:
This just happened about 10 seconds ago. I guess I should have taken a cell phone video.
It's not that I don't believe you, but you seem cursed as far as Win10 goes.
I thought you pretty much always had to set up a local account before joining a domain, but I've always used Home/Pro editions.
No he's right, but that only happens if you're installing from PXE IIRC. It's pretty obnoxious that clicking "Use an active directory domain" doesn't actually... Uh.... Let you join an AD domain...
-
@sloosecannon said in The Official Status Thread:
@FrostCat said in The Official Status Thread:
@mott555 said in The Official Status Thread:
This just happened about 10 seconds ago. I guess I should have taken a cell phone video.
It's not that I don't believe you, but you seem cursed as far as Win10 goes.
I thought you pretty much always had to set up a local account before joining a domain, but I've always used Home/Pro editions.
No he's right, but that only happens if you're installing from PXE IIRC. It's pretty obnoxious that clicking "Use an active directory domain" doesn't actually... Uh.... Let you join an AD domain...
Yeah. My perspective is that it should prompt for the credentials for a user allowed to join the domain, and if they worked, create that as the local user, and then proceed to Do the Needful.
Edit: It also happens on Enterprise installations (from the DVD). Not sure about Pro.
-
@dkf said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
Has anyone seen core application breaks between running a 4.0-compiled application running on a 4.0 installation versus running a 4.0-compiled application running on a 4.6.1 installation?
You don't want a theoretical answer — whether some shitty code could break — you want to know about whether that actually happens in the real case you're dealing with. Heck, bad code could break even if everything is supposedly OK because it is peeking at the version of the runtime and trying to do something “smart”. I've seen that sort of stupid in other places, it wouldn't surprise me to see it again. What matters is not whether some wiseguy has done a crime against programming, what matters is whether you've got one of those crimes about to detonate in a steaming pile of WTF on your desk.
There's no substitute for testing.
Hence "seen", not "Is it possible". I agree it should be tested, but the monkeys screeching on the conference call just blew up without recourse for how to actually get things done. I mean, that's what the whole idea of "pilot program" is about, and they seem to not have a clue...
-
@sloosecannon said in The Official Status Thread:
but that only happens if you're installing from PXE IIRC.
I didn't know Windows even supported PXE booting anymore. We'd looked into it because a certain project required it, but had to use Linux instead since supposedly nothing newer than XP Embedded supports PXE.
-
@dkf said in The Official Status Thread:
Not everything should be firewalled totally by default.
Ugh. I mean, ok, I get it, but some admins... I can't count the number of times when I argued for a box to get temporarily allowed full reign and then slowly locked down as we iron out the problems because SIP and RDP are finnicky bastards. No, we had to spend 6 hours sending emails back and forth to debug issues, instead of having it open to the world for 20 minutes and lock it down as we go. With only thing even listening being the PBX (with outgoing calls already being restricted to local network only) and SSH (with password authentication disabled).
Sometimes the paranoia just gets in the way of getting shit done...
-
@mott555 said in The Official Status Thread:
@sloosecannon said in The Official Status Thread:
but that only happens if you're installing from PXE IIRC.
I didn't know Windows even supported PXE booting anymore. We'd looked into it because a certain project required it, but had to use Linux instead since supposedly nothing newer than XP Embedded supports PXE.
Yep! PXE still supported, though if you're UEFI booting it gets slightly trickier.
-
@mott555 said in The Official Status Thread:
I didn't know Windows even supported PXE booting anymore. We'd looked into it because a certain project required it, but had to use Linux instead since supposedly nothing newer than XP Embedded supports PXE.
A co-worker spent ages on it, but he now has a nice little script that handles multiple Windows versions (we got everything XP - 10 on it IIRC) and multiple Linux distros, all from one menu. Getting the Windows selector working required some ugly hacks from what I heard, but it works smoothly now.
I'm pretty sure he mentioned releasing it at some point, if he does I'll try to remember to link to it.
-
@Tsaukpaetra said in The Official Status Thread:
To the forum: Has anyone seen core application breaks between running a 4.0-compiled application running on a 4.0 installation versus running a 4.0-compiled application running on a 4.6.1 installation?
This isn't counting recompiling 4.0 applications to 4.6.1 (That I know there are some potential issues), this is running applications that haven't been recompiled to the one supposedly installed on the client.Someone around here ( @Yamikuronue ?) was talking about something (Oracle?) that did something like that. It would throw an error and stop when it detected that the "wrong [higher]" version was running.
-
@boomzilla Oracle UPK, with the .net runtime. It'll have a bizzare crash upon loading. I'd treat all Oracle products as potentials for this; make sure the docs say they work with 4.6.1 and not just 4.0 generally
-
@Onyx said in The Official Status Thread:
I'm pretty sure he mentioned releasing it at some point, if he does I'll try to remember to link to it.
Please do! I need to update my setup for UEFI booting and it would be nice if it's a solved issue. At the moment I'm using iPXE, and my DHCP server isn't smart enough to serve the right version to EFI vs Legacy clients....
-
-
@lucas1 I's use the tools I gots
-
@Tsaukpaetra I think it is IPXE in our case as well. I don't know the details, nor the magic involved, but I'll ask him about it.
-
@Onyx said in The Official Status Thread:
@Tsaukpaetra I think it is IPXE in our case as well. I don't know the details, nor the magic involved, but I'll ask him about it.
Yeah, if you're ever curious, some of my stuff is even bootable over the net once you have that loaded. Just tell it to chainload from https://tsaukpaetra.com/iPXE (or, if https isn't enabled, http://tsaukpaetra.com:8000/iPXE )
Edit: NVM on the non-https location, apparently I borked something and it's no longer working...
-
@boomzilla said in The Official Status Thread:
Oracle ... would throw an error and stop when it detected that the "wrong [higher]" version was running.
I'm sure it's easily resolved by performing a simple magical incantation to a license file (following an exchange of a large quantity of , of course).
-
Status: ERMERGERD How do people deal?!?
-
Status: So, in other words.... we're so strapped for cash right now we can't pay for plastic cups anymore?!?
-
@Tsaukpaetra said in The Official Status Thread:
Status: So, in other words.... we're so strapped for cash right now we can't pay for plastic cups anymore?!?
No, that looks to me like
Management found out that we can pat ourselves on the back for saving the environment by literally doing no work, and save money doing it.