marstato

general recent conversations

Discussion between SuperJMN and marstato

Imported from a comment discussion on softwareengineering.stac...

marstato

Aug 28, 2021 12:04

re. immutability of other information: things like types or whether control flow is correct is information that is very unknown at the stage when only the AST or a first level IR exists. Here is an example from one of my projects (one fine day it wants to be an actual compiler, but its far from that right now):

The project currently has two IRs. One is a 1:1 representation of the AST, the other concerns itself mostly with validating all the language rules. This is a class from that second IR:
https://github.com/tmarsteel/compiler-fiddle/blob/master/src/compiler/binding/expression/BoundInvo…

marstato

Aug 28, 2021 11:53

thanks for the feedback <3 :)

marstato

Aug 28, 2021 11:52

@SuperJMN unfortunately, i don't see a way to go immutable for forward references. You have to create the referenced thing bevor creating an object with an immutable reference. That stops working if you have cyclic references :/

Very cool project, though!

marstato

Aug 27, 2021 10:58

Right, that is the more general thing. I was going for cyclic because these are the special kind of forward references that cause the trouble,

marstato

Aug 27, 2021 10:57

@SuperJMN you're welcome :)

marstato

Aug 27, 2021 09:58

What are getting instead?

marstato

Aug 27, 2021 09:58

i edited your question. Is my edit accurate? If yes i have an answer for you :)

marstato

Aug 27, 2021 09:58

Hm, but you already said that you look up from a table of definitions, but the definition is not what you want a reference to. But what else is it? Do you create another object from the definition? If yes, why not place that one in a table as well?

marstato

Aug 27, 2021 09:58

What is your question? I guess you have something like Call(Procedure: "Proc1") and want Call(Procedure: <Reference to Proc1>)?

Discussion on question by hackydude: …

Imported from a comment discussion on workplace.stackexchange....

marstato

Apr 2, 2020 12:37

I totally agree to what Jonathan said. But what light does this shed on you as a Teamplayer? Someone who, behind the backs of the team, makes deals with management that make everyone else on the team look bad. I'd not tell that story in an interview.

Discussion between Oren Milman and ma…

Imported from a comment discussion on security.stackexchange.c...

marstato

Feb 10, 2019 13:01

thanks for challenging me :)

marstato

Feb 10, 2019 13:01

Okay, you convinced me. If we ignore attackers with 0-days in the TPM, i think its safe. I edited the answer.

marstato

Feb 10, 2019 08:43

@OrenMilman ping

marstato

Feb 10, 2019 08:42

Should we change the answer? If so, how?

marstato

Feb 10, 2019 08:41

Okay, I agree to that. Only having the APK is not enough. Only out of curiosity: would it be possible to send the signed attestation of another device?

marstato

Feb 9, 2019 16:51

I have no detailed knowledge of ARM TrustZone. AFAIK TrustZone programs can be read and updated. Even if the signature key is inside a TPM, the problem shifts from between phone and server to between processor and tpm.

marstato

Feb 9, 2019 16:51

@OrenMilman see the 10 immutable laws of security: #1 "If a bad guy can persuade you to run his program on your computer, it's not solely your computer anymore.". Lets assume there is a plethora of root-level exploits for android devices, especially unpatched older ones (e.g. stagefright). I don't see how android root would not get access to that key.

marstato

Feb 9, 2019 16:51

@OrenMilman No, it can't be trusted. Again: the server that verifies the attestation can only verify the data it receives from the phone. The server has no way to tell whether that data is a result of an honest security check or something made up by a malicious app. Like literally: the exact same sequence of 1s and 0s could have been produced by both a valid, clean client and a malicious one. It's all eyewash. At best, it gets harder for an attacker. But there is always a way.

The Water Cooler

General chit-chat for workplace.stackexchange.com. Feel free t...

marstato

Nov 6, 2018 19:31

ok, i'll give it a go today or tomorrow. Thanks very much for the assistance :)

marstato

Nov 6, 2018 19:19

@DarkCygnus how much detail would be needed to answer the deescalating part? I want to avoid the individuals being identifiable from the question because i am not directly involved.

marstato

Nov 6, 2018 19:13

i'll also remove my comments

marstato

Nov 6, 2018 19:13

aboslutely agree

marstato

Nov 6, 2018 19:13

it's really two separate questions

marstato

Nov 6, 2018 19:13

i'll remove the details then

marstato

Nov 6, 2018 19:13

@DarkCygnus Thats exactly what i did and what made @JoeStrazzere highly doubt the legitimacy of my question

marstato

Nov 6, 2018 19:12

okay, i can see it fit into the "communication issues" area

marstato

Nov 6, 2018 19:07

@DarkCygnus and that would be on-topic in workplace.se?

marstato

Nov 6, 2018 19:04

@enderland how do you come to the conclusion that i am aksing "how to lie to my employer?"

marstato

Nov 6, 2018 19:03

and whats the TWP format, btw?

marstato

Nov 6, 2018 19:01

Rephrasing the question now does not really make sense IMO

marstato

Nov 6, 2018 19:01

I am totally fine with just removing the comments.

marstato

Nov 6, 2018 19:01

Now, how do we make that question useful?

marstato

Nov 6, 2018 19:01

@DarkCygnus also absolutely agreed. In this specific scenairo though, i dont think it would have made a difference. There's a serious conflict going on....

marstato

Nov 6, 2018 19:00

@DarkCygnus i agree. And it was phrasings like that i was hoping to see in the answer.

marstato

Nov 6, 2018 18:59

fact also is: when you call in sick in germany and you are obliged to given a rough estimate on the duration of your leave (possibly stating "i dont know, will see the doc ASAP and come back to you")

marstato

Nov 6, 2018 18:57

@DarkCygnus it's not; but it can be taken as an implication for asking effectively that

marstato

Nov 6, 2018 18:56

but the answers given perfectly adress my curiousity as to whats ethical for an employer to do in that situation

marstato

Nov 6, 2018 18:56

I agree, the question as posted originally does not help much in the first aspect

marstato

Nov 6, 2018 18:55

so really, there are two reasons i asked. I wanted to deescalate. And also, thinking about the problem, i wondered what options an employer has. In germany you can really easily get into trouble for contacting sick employees for work-related topics

marstato

Nov 6, 2018 18:54

hi guys

Discussion between candied_orange and…

Imported from a comment discussion on softwareengineering.stac...

marstato

Nov 2, 2018 07:32

Also, to be clear, I am not talking a out the determinism from physics science; That's a whole other story. To me, in programming, determinism is limited to the scope of the language specification; completely independent from whether the code is ran on a silicon chip or by my brain when reading the code. Change in a global variable can definitely be deterministic. What the electrons in the chip do, is not (and it's out of scope for this discussion).

marstato

Nov 2, 2018 07:23

That is an interesting an helpful property as opposed to a function that depends on state the caller cannot (or only with huge efforts) control, like a DB. This is where the term deterministic comes in to help. Whether you like that or whether you find it useful is absolutely irellevant.

marstato

Nov 2, 2018 07:22

Now how do we call this function? Pure is rules out because it is not referentially transparent. It also depends on and modifies state, but we can control all that state. So we can safely say that, if you invoke this function several times with the same state and the same inputs, the state will always change in the same way and the return value will be the same.

marstato

Nov 2, 2018 07:18

Again, as to side effects and threading: assume this code int foo(int x, *int y) { *y = x + *y; return x; }. The return value depends only on the inputs. But it has side effects. Those depend on the inputs only, too. Is that a problem for threading? Sure, if you pass a shared pointer for y. But it's totally fine in a multithreading context if you pass a stack pointer.

marstato

Nov 2, 2018 07:10

@candied_orange in the comments you literally said "you cannot classify a function based on the equipment it is run on" but then go ahead an insist that our definition of determinism is flawed because calculation on a CPU consumes energy and emits electromagnetic waves. Your reasoning is at least as inconsistent. An I hope it's only your willingness to learn that's showing here. Just because determinism is not useful for you, you cannot just redefine it??

Discussion on answer by Ewan: Should …

Imported from a comment discussion on softwareengineering.stac...

marstato

Sep 26, 2018 13:47

@Ewan "slow" - are you serious? How in the world can a stack unwind be slower than a round-trip to the DB via the network. Please stop this performance religion hogwash. -1

Discussion on answer by Cort Ammon: W…

Imported from a comment discussion on softwareengineering.stac...

marstato

Oct 12, 2017 12:33

@PieterB where does the question mention physical transport of the information? The tags say data and storage. I agree that 8 times is BS, as i realised 2 minutes after posting it; got it correct in my answer. I also agree that, in the context of physical information transport using electrical circuity my answer makes no sense.

marstato

Oct 12, 2017 12:33

This does Not even answer the question. 4 Symbols per Digit means 8 Times the amount of information compared with because for aby given number of digits, Base 4 can encode 4^n states whereas Base 2 encodes 2^n. Also, the networking stuff is offtopic. I didnt edit the answer because it has many upvotes.

Discussion on question by Gusdor: Wha…

Imported from a comment discussion on softwareengineering.stac...

marstato

Jun 27, 2017 11:35

If all the new stuff is baffling them you can try to go step by step. Do one thing at a time and move on to the next only after the previous one has been fully understood.

Discussion between Brilliand and TSar

Imported from a comment discussion on softwareengineering.stac...

marstato

May 26, 2017 13:43

anyways, having skilled workers on a project makes all sorts of things easier, regardless of what approach you take