The Official Status Thread
-
Status:
This database diagram of a system I just inherited this week is a work of beauty:
I mean that seriously. It's complex as shit, but it describes a complex business process, and all the tables have reasonable names, and all the foreign key relations are actually specified in the database.
I've seen much, much worse.
This is the same development team that apologized to me over the bad state of their code, and I was like, "look, I pulled the repo, opened the solution, hit the 'play' button, and it all compiled and ran without me having to do any tinkering or shit. That puts you ahead of like 95% of legacy codebases I've been involved with."
-
-
@blakeyrat said in The Official Status Thread:
I think I had an emotion once.
When you discovered HyperCard ?
-
-
-
Status: Aaaand biology final in 15 minutes, and then I'm done for the year.
-
Status: Sober enough again after our work Christmas Lunch to consider coding. A good time was had indeed!
-
@izzion
Follow-up Status: The explosion was due to a configurable connection limit on the SSRS server and raising that limit seems to have resolved the kaboom. Now to cross my fingers that the job runs to completion in the expected amount of time...
-
@blakeyrat said in The Official Status Thread:
"That puts you ahead of like 95% of legacy codebases I've been involved with."
That puts them ahead of 100% of legacy codebases I've been involved with, to the best of my memory.
-
@heterodox However, that said it turns out:
- The Authentication database has no Dev or Stage version
- The Dev and Stage databases for the main functionality are on the same server and use the same logins
- The logins are SQL Server Connections and not domain logins
-
@blakeyrat said in The Official Status Thread:
@heterodox However, that said it turns out:
- The Authentication database has no Dev or Stage version
Yikes. (These seem to be system integrator problems rather than problems with the codebase, but still.)
- The Dev and Stage databases for the main functionality are on the same server and use the same logins
Fucking yikes. I've seen nasty mistakes happen that way.
- The logins are SQL Server Connections and not domain logins
Eh. Meh. They may have done that because they encountered insurmountable hurdles working through their provider to get proper service accounts. When faced with a stunning wall of bureaucracy, I've been known to take the pragmatist's way out as well. Not disagreeing that domain logins wouldn't be better.
-
@hardwaregeek said in The Official Status Thread:
@tsaukpaetra said in The Official Status Thread:
@hardwaregeek said in The Official Status Thread:
to drink cerveza.
It's not beer?
Referencing commercial where the tagline was "it's not beer. It's cerveza".
-
Status: I just want to go home. Today is a waste. But I got free lunch, and I get free dinner. Good, yes, we met our goals and released on time. But today has been all meetings.
-
Status: Dammit Windows, it'd have been nice to know that one of the files I was trying to copy was >4 GB before you copied everything else onto the removable device with the FAT32 filesystem, which I now need to reformat to NTFS...
-
I made a thing.
I took a Philco radio from the 1930's, put some amplified speakers into it and a Raspberry Pi with a DLNA renderer service.
Now I can stream music from my phone to it
I'll probably replace the Pi with a Chromecast audio for added convenience.
-
@blakeyrat said in The Official Status Thread:
@carrievs said in The Official Status Thread:
and I'm currently getting all soppy and emotional with nothing to distract me.
I think I had an emotion once.
Was it anger?
-
@jaloopa He didn't stop having it, though.
-
@anotherusername said in The Official Status Thread:
the removable device with the FAT32 filesystem
-
Status: Looking through the code for the project we're inheriting now.
The tests are long (Vegeta would be shocked at the line count), but admittedly the individual tests usually aren't. They have no asserts, though, which I disapprove of.
The code is loooooong. They didn't use enough classes. They don't seem to know how Linq works.
But... it's fixable. It was clearly written fast, so it's messy. It works, and the methods are mostly short.
It's a relief, seriously. At least this time, they just went ahead and used REST, instead of making something up. My mood has lightened slightly.
-
Status: As part of my coding a connection pool, I also added a bit more verbosity to errors (previously of course it didn't say anything but "Error running query".
Result?
[2017.12.20-00.01.57:273][ 15]DatabaseCommunicatorLog:Error: SQLCODE: -1 Query failed to execute command: [[INV_Saved_SyncBack] @PlayerID = 1135] The error was: 24000: [Microsoft][SQL Server Native Client 11.0]Invalid cursor state
So now off to figure out what that's supposed to mean...
-
@magus said in The Official Status Thread:
The tests are long (Vegeta would be shocked at the line count), but admittedly the individual tests usually aren't. They have no asserts, though, which I disapprove of.
As long as each test is clear and independent of all the other tests so you can safely run them in isolation or together, the wordiness is really not important. Non-isolated test cases are a total PITA that you only find out about when things start going wrong horribly while you're trying to fix something else, so please be proactive about stamping them out early as it saves you so much pain.
(This has caused me so much trouble over the years. Don't skimp on it.)
-
@tsaukpaetra said in The Official Status Thread:
this was in response to an hour long debate on buying gift cards with gift cards.
So, I just had to try it and see what's what.
Depends on the store. Also it depends on whether you're on the website, or face-to-face with a helpful manager.
-
@anotherusername said in The Official Status Thread:
@tsaukpaetra said in The Official Status Thread:
this was in response to an hour long debate on buying gift cards with gift cards.
So, I just had to try it and see what's what.
Depends on the store. Also it depends on whether you're on the website, or face-to-face with a helpful manager.
Yeah. I used a self-checkout kiosk. Went fine. But, taking the second card online, couldn't use it for an eGiftCard (option just wasn't available).
-
@dkf said in The Official Status Thread:
Non-isolated test cases are a total PITA that you only find out about when things start going wrong horribly while you're trying to fix something else, so please be proactive about stamping them out early as it saves you so much pain.
I always wondered why testing libraries don't have an option to run the tests in random order, which would expose these pretty quickly.
-
@blakeyrat said in The Official Status Thread:
@dkf said in The Official Status Thread:
Non-isolated test cases are a total PITA that you only find out about when things start going wrong horribly while you're trying to fix something else, so please be proactive about stamping them out early as it saves you so much pain.
I always wondered why testing libraries don't have an option to run the tests in random order, which would expose these pretty quickly.
I want a tool that will run parallel code not just in a random order but in every possible order. If the code does something different in one sequence of actions than another, there's a race condition in the program.
-
@anotherusername said in The Official Status Thread:
which I now need to reformat to NTFS
Can't you just do an in-place conversion? I seem to recall that being an option.
-
@pie_flavor said in The Official Status Thread:
@anotherusername said in The Official Status Thread:
the removable device with the FAT32 filesystem
It was formatted when I got it. It never occurred to me to even worry about it, and I never needed to until now.
@hungrier said in The Official Status Thread:
@anotherusername said in The Official Status Thread:
which I now need to reformat to NTFS
Can't you just do an in-place conversion? I seem to recall that being an option.
Not by right clicking the drive and using the "format" command. And it never even occurred to me to see if I could find a tool that would convert in place.
Only need to do it the once, and it took maybe 5 minutes extra to move everything that wasn't the partially copied folder into a temporary folder on the desktop and back again after I formatted it.
Anyway, the ability or not to do in-place conversion in no way excuses the fact that Windows couldn't just tell me right at the start that it'd be a problem.
-
@anotherusername I'm not sure but I think Teracopy would tell you beforehand. And you should be using Teracopy for all your file copying needs anyway, there's really no reason not to.
-
Status Dug out my old linear algebra knowledge to work out demographics for a fantasy setting I've been building.
n[ji] = N*g[j]*Sum(r[k]*C[ki],k)
The number of people who can cast spells of level
j
of classi
(n[ji]
, where[]
are subscripts) is given by the product of the total number of peopleN
, a weighting factor that depends on levelg[j]
, and the matrix product of the race make-up vectorr[k]
(what percent of people are of racek
) and the race-class correlation matrixC[ki]
. Most annoying part was normalizing the columns ofC[ki]
--not all races produce spell-casters at the same rate and the sum over each columnSum(C[ki],i)
should be the overall prevalence of spell-casters of that race.Side note--writing this makes me wish we had Mathjax support on this forum. Or is there a better way of writing equations that I'm just missing?
-
@benjamin-hall That's great, but still nobody's gonna read it if you don't get rid of Jar-Jar Binks
-
@blakeyrat said in The Official Status Thread:
@benjamin-hall That's great, but still nobody's gonna read it if you don't get rid of Jar-Jar Binks
It's a D&D setting I've been running games in for several years now. And Jar-Jar Binks is banned, along with gnomes.
-
@tsaukpaetra said in The Official Status Thread:
So now off to figure out what that's supposed to mean...
Apparently you're not supposed to try running two different statements from the same handle (i.e. I didn't re-allocate the statement and just splatted in a new one over the current one).
I think I could probably get away with just resetting the statement cursor, but decided it would be safer just to create a whole new one.
In any case, initial implementation is done, and cursory testing seems to indicate it's working. Tomorrow we shall stress it!
-
Earlier status: I love mornings that start with blue screens, followed by updates.
-
@pie_flavor said in The Official Status Thread:
@tsaukpaetra said in The Official Status Thread:
@pie_flavor said in The Official Status Thread:
@tsaukpaetra I was not responding to that post. I was responding to this post:
@tsaukpaetra said in The Official Status Thread:
@pie_flavor said in The Official Status Thread:
@tsaukpaetra You're asking us why you're doing this?
No, I'm asking why it's another source of unnecessary complexity in my life.
I mean, I could simplify my life so much if I just didn't have any electronic (note: not necessarily no electric) devices except my mobile phone, but, what point would that prove?
I don't understand? What is normally?
Well, start by not reinstalling Windows until you (a) have a serious problem and (b) know that only reinstalling Windows will fix it.
I haven't done a maintenance Windows reinstall since XP/Win2003. And I ran Vista continuously the entire time it was current. And 7 from 2009 until 2016. All on a single install.
Bitrot is dead, if it was ever real in the first place. My current thinking is that it wasn't, and it was just shitty hardware combined with shitty software combined with being a shitty know it all kid.
-
@weng said in The Official Status Thread:
@pie_flavor said in The Official Status Thread:
@tsaukpaetra said in The Official Status Thread:
@pie_flavor said in The Official Status Thread:
@tsaukpaetra I was not responding to that post. I was responding to this post:
@tsaukpaetra said in The Official Status Thread:
@pie_flavor said in The Official Status Thread:
@tsaukpaetra You're asking us why you're doing this?
No, I'm asking why it's another source of unnecessary complexity in my life.
I mean, I could simplify my life so much if I just didn't have any electronic (note: not necessarily no electric) devices except my mobile phone, but, what point would that prove?
I don't understand? What is normally?
Well, start by not reinstalling Windows until you (a) have a serious problem and (b) know that only reinstalling Windows will fix it.
I haven't done a maintenance Windows reinstall since XP/Win2003. And I ran Vista continuously the entire time it was current. And 7 from 2009 until 2016. All on a single install.
Bitrot is dead, if it was ever real in the first place. My current thinking is that it wasn't, and it was just shitty hardware combined with shitty software combined with being a shitty know it all kid.
I have noticed my system running noticeably worse since the latest Windows update (I frequently get full system stutters and games crashing with no error messages now), but that's not something that I could fix by reinstalling Windows unless I installed an older version (or a newer version if they fixed the problem).
-
Status: Flabbergasted. Somehow, connecting to this Centurylink's 2.4 GHz network causes the network card to crash. At least, that's what it appears, the device gets disabled in Network Adapters for a second and then Windows claims it couldn't connect. Other 2.4 GHz networks (and the 5GHz side of the same device) work fine.
I bounced the interface on the modem and it seems to be back to normal, but I've never seen this behaviour before.
-
Status: Got my parents a super-wizzy upscaling Bluray player for Christmas. Then remembered their sound system was a bit old so rang them up to innocently inquire what inputs it had.
Does it have ARC?
No
Does it have Optical In?
No
What does it have?
It's got two little red and white posts.One Amazon order for an S/PDIF DAC and some RCA leads coming up.
Edit: I hope their TV has HDMI in...
-
@blakeyrat said in The Official Status Thread:
The Dev and Stage databases for the main functionality are on the same server and use the same logins
Why is this bad?
-
@izzion
Job died because another job conflicted with it, but it was churning at more than 200% the rate of the production job for 12 continuous hours without hitting the connection limit. victory
-
3 more days of this job. It's really hard to get out of bed now
Would it be bad if I just napped all day at the office?
-
@bb36e Would your productivity significantly decrease?
-
@cartman82 it means if I accidentally deployed a dev or stage version with the prod connection string, it'd work and not fail. Then I'd have untested code interacting with production data.
-
@blakeyrat said in The Official Status Thread:
@cartman82 it means if I accidentally deployed a dev or stage version with the prod connection string, it'd work and not fail. Then I'd have untested code interacting with production data.
Ah, dev and staging are on the same server as PRODUCTION. That makes more sense.
It's still a tiny chance (unless you are manually copying settings during each deployment), but I acknowledge it'd be better if production is away from dev.
-
@ben_lubar said in The Official Status Thread:
I want a tool that will run parallel code not just in a random order but in every possible order.
Unless you have a stupid quantity of hardware, that'll slow things down a load as the number of permutations grows really fastβ¦
-
@benjamin-hall said in The Official Status Thread:
n[ji] = N*g[j]*Sum(r[k]*C[ki],k)
nji = N Γ gj Γ β(rk Γ Cki, k)
Or something like that.
-
@dkf said in The Official Status Thread:
@ben_lubar said in The Official Status Thread:
I want a tool that will run parallel code not just in a random order but in every possible order.
Unless you have a stupid quantity of hardware, that'll slow things down a load as the number of permutations grows really fastβ¦
Running two operations in parallel with about the same number of operations in each is only O(n2), so adding another thread would make a much larger performance impact than adding another operation.
I'm assuming the analysis tool is statically able to know which pieces of code never interact in any way.
-
@dkf I agree totally, but I haven't had time to read through a 13k line file to examine how well the tests run in isolation yet.
@blakeyrat said in The Official Status Thread:
I always wondered why testing libraries don't have an option to run the tests in random order, which would expose these pretty quickly.
I normally run them in parallel, and consider that to be about good enough.
@ben_lubar said in The Official Status Thread:
I want a tool that will run parallel code not just in a random order but in every possible order. If the code does something different in one sequence of actions than another, there's a race condition in the program.
I kind of wonder if you could trick Intellitest into doing that...
-
@cartman82 said in The Official Status Thread:
Ah, dev and staging are on the same server as PRODUCTION. That makes more sense.
Oh yeah. I guess I typed that idiotly. It made sense in my head.
@cartman82 said in The Official Status Thread:
It's still a tiny chance (unless you are manually copying settings during each deployment), but I acknowledge it'd be better if production is away from dev.
They hired me to produce quality software, not pull Evel Knievel stunts with their code.
-
Status: Sad, somehow:
-
@magus XSS in 3...2...