VikingofRock posted:

I just started working somewhere that uses perforce and I'm having trouble squaring it with this kind of work flow. E.g. today I was writing some tests when I noticed that some existing code was broken, and then it turned out that the fix strongly suggested a small refactor. In git I would split that into 3-ish commits (something like fix + add tests for bug, refactor, add new tests). I'm not sure what the "right" perforce workflow is, though. Should I be splitting this work into multiple workspaces? Are large changelists just considered more acceptable?

VikingofRock posted:

I guess my more general problem is what do I do if I start working on feature B, and then I realize that it depends on feature A which is unimplemented and in the same file? Is it possible to submit the two features as two separate changelists, without throwing away my work on feature B?

Sorry if I'm getting some of the terminology wrong or overlooking something obvious; I've only been working with perforce for about a week.

gonadic io posted:

I'm making a twitter app. All calls to twitter have to include both app credentials and user credentials (if you're acting on behalf if a user).

This is fine when I'm running it on my own comp, and would be fine with a server model where the user clicks an oauth thing on my page, gives me their user creds and I make the requests.

However let's say I wanted to make it clientside. The trouble is, whether I did it via a script on a page or via a browser plugin, some user would be able to access the app creds and then flaunt their rate limit to get the app banned.

Is there a way to solve this? There are other browser plugins that block people on twitter, I wonder how they handle this.

Sapozhnik posted:

Also make sure you use TypeScript and Parcel in your project as well.

gonadic io posted:

I'm doing my first async http server/client/db thing in rust and am finding it pretty painful honestly: https://github.com/djmcgill/de-list-server

just the process of getting used to the hyper/tokio setup, and it doesn't help that the library I proofed out the sync stuff in hasn't updated it's async interface in a while so i'm having to reimplement it. i'll probably end up contributing my changes back as a pr i think but it's annoying to have to do

e: also i should have just used an error library (or just ignore all errors) from the beginning instead of ending up with poo poo like Box<(dyn std::error::Error + Send + Sync) in a bunch of places

Beamed posted:

This was 2008-2011

Wheany posted:

Phobeste posted:

c++ metaprogramming is incredibly good but unfortunately it requires a lovely perfectly smooth brain. marble-like, hard as a rock. this is enlightenment

TheFluff posted:

I don't seem to have made the difference between a query builder and an ORM clear. A query builder has an API that maps very closely to SQL statements - the goal is to manipulate a SQL AST. It's a metaprogramming tool, focused on generating programs (queries). An ORM on the other hand has an API that focuses on abstracting away the SQL in favor of an API that focuses on models and their relationships. It's a mostly declarative data modeling tool, based on the assumption that if you describe the models (tables) and the relationship between them then it's straightforward to generate a fairly limited set of functions to traverse those relationships. This assumption obviously does not hold, and in fact breaks down pretty much as soon as you see a many-to-many relationship. You end up restricted to an incredibly small subset of SQL - the forms that the ORM knows how to translate into simple model relationships.

I've used both and the ORM style is extremely limiting while not really making things any easier than the query builder style. Most query builders also come with things like model classes and that sort of thing too. Like, seriously, if your SQL toolkit can't even represent a semi-join without falling all over itself, it's loving useless. Even Django lets you do that, although it's awkward and I really wish I could have a real query builder instead of their janky poo poo.

Space Whale posted:

The only reason this threadshit got checked in was nobody was doing code reviews, or, the people at $JOB when it was fully startup mode honestly didn't understand this - none of them.

Which scares me.

Plorkyeran posted:

these sound like things that could result in a crashed plane if you gently caress up enough

Space Whale posted:

Oh, very much so.

Space Whale posted:

So basically just loving rip the thread

Wheany posted:

i spent all day making a progressbar for our webapp's bootstrap script loading the actual ~1 MB (compressed) js and css data. i develop it in firefox and when i try it in chrome, the loving browser reports the amount of uncompressed data in its progressevent, so the progress bar hilariously breaks way out of the right side of the containing div. edge worked correctly. ie 11 didn't work because i used arrow functions (and probably would have broken on promises next)

web "development"

HoboMan posted:

guy here likes to inert into the db via stored procedures where there is only one augment which is an xml blob of whatever you are updating/inserting.
it skips having to do an explicit mapping and lets the same proc handle any number of items, but it still feels wrong despite not having anything specific to point to.
what terrible thing am i missing?
or is this fine?

e: the xml blob is automagically parsed and the data properly put in the appropriate columns, not just jammed in the db

Share Bear posted:

is there a good generalized guide to writing docs that i could provide most high-school level english writing programmers i work with? i know this is a skill that people need to work on to improve

i've been going with saying "who (is this for, made it), what (is it for), how (to build and run it, add features), and why (was this made, what problem does it attempt to solve)"

we've only had how to build and run it up until recently

CRIP EATIN BREAD posted:

stream of consciousness style thats impossible to follow

florida lan posted:

https://developer.mozilla.org/en-US/docs/Web/HTTP/CORS

the


really should be followed by


it may well be some other reason but i am astounded by how many people who work with web apps have just somehow never heard of cors and have never looked at dev tools to see the big angry 'hey you should check out cors' warning that pops up when its constraints arent satisfied.

Sapozhnik posted:

Certain attacks were always possible through malicious IMG tags or post forms or combining the above with URLs that cause http to happen on non http ports.

New stuff that adds additional potential attack vectors beyond what already existed has to prove remote consent first. This happens through same origin policy or cors for, well, cross origin requests, or the websocket handshake to prevent arbitrary JavaScript from generating arbitrary TCP traffic to arbitrary hosts.

Why does a websocket handshake begin with something that looks like an http get? Because a malicious IMG tag could already generate those against arbitrary ports on arbitrary hosts long before even JavaScript was a thing.

Jabor posted:

You do realise that the person who wrote a website might not be the same as the person accessing it through a web browser, right?

HoboMan posted:

nginx:
i imagine it's supposed to be said like n-gin-x (engine x) but i can't help but read it as n-ginx (en jinx)

VikingofRock posted:

Java question: I have a few large-ish functions, each of which I would like to call with two (possibly later more) different possible input types. The implementation of these functions would be largely the same when called with either input type, but there's just a small method or two in each function that would be different for each type. I know I could implement this roughly like this:

Java code:

void <T extends Baz> doOneThing(T input) {
  // lots of code...
  if (input instanceof Foo) {
    ((Foo)input).doSomething();
  } else if (input instanceof Bar) {
    ((Bar)input).doSomethingElse();
  }
  // lots more code
}

void <T extends Baz> doAnotherThing(T input) {
  // similar strategy here...
}

However, this feels pretty inelegant. For one thing, it doesn't scale well to more functions / more types. For another thing, it doesn't behave nicely if I pass it the wrong types: at best I can just make it fail at runtime. So what's the best strategy here?

A little more info: while Foo and Bar both extend Baz, Baz as a whole is very broad and most of it's subclasses won't work with these methods. I do not directly control Foo and Bar, so I can't e.g. make them both implement an interface which does what I want.

Blinkz0rz posted:

this thread gets into dumb slapfights about which language or ecosystem or dependency management tool is better but it's all garbage and we should be talking about how we're doing our best to make our day-to-day lives better while we stack absurd amounts of paper

HoboMan posted:

pl thread is for language slapfights, this therad is for commiserating and or panicking about the terrible code we are all making GBS threads out (which is all code)

that said, shaggar is still right

HoboMan posted:

ok, again i have not had to design poo poo before.

how to restfully have an endpoint that returns a resource filtered by another resource?

eschaton posted:

the iPhone developer tools were and are good, hope this helps

CRIP EATIN BREAD posted:

they've had a ios simulator since like forever. its not a true vm but it runs your code.

eschaton posted:

what the gently caress is wrong at these places that things like a taxi ordering app wind up 2.3 MLOC?

Corla Plankun posted:

I have a $500 stipend for "educational materials' at my computer-toucher job. is there any better way to spend it than oreilly safari?

i tried a free demo of safari for a couple of weeks and it seemed very good but i grew up poor so i have tremendous anxiety when making any purchase over 16 USD

Progressive JPEG posted:

no that's 0 in 100

it's for saving money on rent, nothing else is even a consideration

animist posted:

the worst data format i've had to work with recently is json-ld, which is the result of the semantic web people getting their greasy hands on json

code:

{
  "@graph": [
    {
      "http://schema.org/description": "The Empire State Building is a 102-story landmark in New York City.",
      "http://schema.org/geo": {
        "@id": "_:b1",
        "http://schema.org/latitude": "40.75",
        "http://schema.org/longitude": "73.98"
      },
      "http://schema.org/image": {
        "@id": "http://www.civil.usherbrooke.ca/cours/gci215a/empire-state-building.jpg"
      },
      "http://schema.org/name": "The Empire State Building"
    },
    {
      "@id": "_:b1",
      "http://schema.org/latitude": "40.75",
      "http://schema.org/longitude": "73.98"
    },
    {
      "@id": "http://www.civil.usherbrooke.ca/cours/gci215a/empire-state-building.jpg"
    }
  ],
  "@context": "https://w3id.org/security/v1",
  "signature": {
    "type": "EcdsaKoblitzSignature2016",
    "created": "2019-06-07T03:23:16Z",
    "creator": "ecdsa-koblitz-pubkey:020d79074ef137d4f338c2e6bef2a49c618109eccf1cd01ccc3286634789baef4b",
    "signatureValue": "IJY25idmG+Yvy9kB57yNB6WNRU9jLtb8Vv6BEEIus9mcQTIpyPr25H3tx876WA7c1DXl8Wcx7TPvOSovKQ3HbyE="
  }
}

yes, that is RDF embedded in JSON. it's also signed with the bitcoin elliptic curve because they're pushing this to the blockchain people for some godforsaken reason

quote:

EXAMPLE OF A DESCRIPTION

code:

 <script type="text/javascript">
var fo = new FlashObject("http://google.com/flash/preview-player.swf",... 
</script>

eschaton posted:

but what is a build system really but a whole bunch of terrible programming

Ciaphas posted:

i'm searching almost 15 year old C memory here. do execv() and friends do anything to the process space's heap? or is that platform/impl dependent

i'm chasing a memory corruption bug that only appears when this library emergency-restarts, and the final call is just a plain old forkless execv. i've not seen this before, and i think it's causing an uninitialized pointer elsewhere to look initialized after the image is loaded and starts running, if that makes sense

code:

      execv(g_hMotionWnd, "\\hard disk\\boamotion", gMotionSettings->argv);

(it also seems like a "clever" way to restart a program, scare quotes intended, but idk on that one, might be standard procedure)

Ciaphas posted:

I've not encountered this professionally, somehow, so I'll just ask


when you've got a bug like this that you think you've put in enough defensive code to prevent relapse, BUT for whatever reason you can't reproduce the original bug in the first place, how is my time best spent going forward

- find the drat bug come hell or high water, doesn't matter how unlikely it is to produce in the Real World
- code defensively, put in traces, keep an eye out, and hope like hell it doesn't happen again while I get some other work done


I guess this is more a 'depends on your manager' question, huh

jit bull transpile posted:

I'm not usually one to make fun but


isn't this guy the one who casually admitted to farming his job out to a team of cutrate Brazilian programmers without telling his employer? did he really think he could run that con on Amazon?

Blinkz0rz posted:

i'll be the terrible programmer and say that i honestly don't understand what the practical case for an uninitialized guid is in c#

is it just so you can preallocate a collection of them and defer the call to the rng to when you need to use them? if so that feels really weird. i have to assume the clr introduces enough of an overhead that the performance impact would be negligible...

Blinkz0rz posted:

but why is that something you'd want to do in the first place? it feels like a weird consideration for something whose existence and use relies entirely on randomness

like, have System.Guid() create a default v4 and have an alternative constructor that accepts null and spits out all zeroes or something or like System.NullGuid() or some poo poo

or take it a step further and don't even have a guid type; just have a bunch of guid helper methods that spit out strings

cinci zoo sniper posted:

his code:



posting as images because clownflare blocks sqls lol

Blinkz0rz posted:

a uuid should be a string generated by static methods depending on the type of uuid with other static methods to convert them to their other representative types

Finster Dexter posted:

oh a billion new posts over the weekend, which argument did we re-hash this time?

1. spaces/tabs
2. vim/emacs/sublime/vscode/notepad++
3. p-lang x is dumb/not dumb
4. all of the above

gonadic io posted:

Every programmer ever sees some dumb thing which isn't behaving the ways they want (usually for reasons they don't understand) and smugly says to themselves that they could do better. For fun reading check out the history of noisebridge's elevator

E: https://www.noisebridge.net/wiki/Elevator

akadajet posted:

lol if you get onto this thing

Sapozhnik posted:

is rust language server still hot steaming garbage

gonadic io posted:

Yes. For as long as this issue: https://github.com/rust-lang/rls/issues/352 is WONTFIX it will not be usable compared to jetbrains'. Rust analyser aka RLS 2.0 does fix it, but doesn't have IDE integration yet (and they might not build it, and instead use its lessons for a real RLS 2.0? https://ferrous-systems.com/blog/rust-analyzer-status-opencollective/)

MononcQc posted:

In any case, all the recent major incidents from FAANG-style orgs were all blamed on "config changes" which points towards a shift in how people work around their dev pipelines towards less-monitored and manually-controlled roll-outs of complex features where rather than going through a code process for everything, you write code that can be tweaked through config, deploy that, and then do the old-style prod-fiddling via configs that are dynamic, untested, unmonitored (in comparison to code). Then when it blows up, you go "well, we need to make configs safer and testable" but it feels more and more that configs are a workaround or consequence of developer tools and pipelines with lovely ergonomics.

Bloody posted:

i want a type in c# that i can use like a sql table. like i want some Collection<TIndex1, TIndex2, TValue>. like a dictionary with another index. is there anything im missing in the stdlib like this or a usable nuget package for this or do i get to roll my own bullshit

CRIP EATIN BREAD posted:

I've never met a person asking about a K/V store that had a workload that even came close to requiring something other than a postgres table with the default database settings.

Captain Foo posted:

EIEIO

Captain Foo posted:

EIEIO

Captain Foo posted:

EIEIO