DDraceNetwork - developer

good idea

00:12

but very expensive

00:16

or maybe not

@coffee, now that I updated to master we can work on it :)

04:49

My server is running 24/7 now, will probably make some improvements/ changes next week

fokkonaut

[ddnet/ddnet] Pull request opened: #1851 Fix blockZ and infectionZ

I talked to coffee (creator of both mods) and he said it is okay to use zoom on the mods, also: eye emotes work there so the eye wheel is also supported. I have updated both mods to ddnet master so they have all the new stuff

bors[bot]

[ddnet:master] 2 new commits

5c280a3 Fix blockZ and infectionZ - fokkonaut 898e6ca Merge #1851 - bors[bot]

06:27

bors[bot]

[ddnet/ddnet] Pull request closed: #1851 Fix blockZ and infectionZ

C++ classes are really quite lacking for reliable software

13:08

s/$/ development/

13:09

A failure in the constructor is very hard to communicate up the stack

13:09

Exceptions/decoupling error handling from error detection, is also quite an unwise pattern

13:11

OOP in general is quite unwieldy...

having 50 manual error checks is also unwieldy compared to a single catch

50 manual error checks that know the context of the error and are easy to spot if missing

14:19

exceptions can travel all the way up the stack if you forget to handle some and it's not so easy to know from a glance what exceptions can arise from calling a function

what exactly does %current% do as entities background? Does it allow to see the quads layers that are placed before the game group or all other quad layers too?

It uses the current map as your entities background

14:22

Same rules as for any other map too

14:22

Nothing in front of game is displayed

14:23

Thats why i told you before, it doesnt work for your servers with Movable objects. You would need to play both; one layer behind and one in front of game

@qshar you can set whatever bg you want

14:25

With this

@n000b sorry that was absolutely not helpful

ye, thats sad. Is it hard to implement such option to display all quad layers with entities?

14:26

bg means behind the game layer - so it doesn't fit

no

14:26

shouldn't be hard

this is very simple

Can I ask anyone to create such pull request? Or advice me where to take a look (I found something in background.cpp, bit still not sure)

C++ ppl: "exceptions should only be used in exceptional situations" Also C++ ppl: "Only way to fail in a constructor is an exception"

14:36

@qshar: What are you trying? Maybe I can give you a pointer

I'm trying to make quads layers visible when you play with entities. It should help many pros to play maps with moving objects using entities

14:44

In my mod there is a map configuration that says that the map is using this modification, so it can be used to enable this by default. Or in the background entities setting can be set to something like %quads% to see all quads layers of this map

Ryozuki

[ddnet/ddnet] Pull request opened: #1852 Fix decimals on pause/spec

Fixes #1846

@Ryozuki nice duplicate

is it already there?

16:06

xd

16:06

i didnt see it

https://github.com/ddnet/ddnet/pull/1848

Fix deci seconds not updating in spectator/pause (fixes #1846) by ...

Before this was set in RenderDDRaceEffects(), but that function is only called if m_pClient->m_Snap.m_pLocalCharacter exists, which isn't the case in spectator/pause.

dont worry i remove it

my fix is better, because the timer itself has nothing to do with the DDRaceEffects, calling that agin would make no sense

Ryozuki

[ddnet/ddnet] Pull request closed: #1852 Fix decimals on pause/spec

why u open a issue if u fix it urself tho?

16:07

its useless

i opened the issue, the though mh i could try to fix it

16:07

then fixed it

16:07

not useless

I just read a piece about how to check errors in a std::ifstream properly, much more convoluted then I imagined

16:40

Did you know that errno is only set if badbit is set? Not if failbit is set

12pm

[ddnet/ddnet] Pull request opened: #1853 Render all entities on non ddnet servers

Fixes #1741

https://fedoramagazine.org/introducing-fedora-coreos/

Benjamin Gilbert

Introducing Fedora CoreOS - Fedora Magazine

Introducing the first preview release of Fedora CoreOS, a new Fedora edition built for running containerized workloads securely and at scale.

@Learath2 @heinrich5991 @ChillerDragon So i´ve just tried the new power brick nothing changed...

So my raspi is dead, actually this piece of sh.. never worked...

Hmm, it's pretty rare that a raspi is DOA

19:53

@moris unlucky

cruzomat: Might want to try another cable

20:04

https://arstechnica.com/gadgets/2019/07/raspberry-pi-4-uses-incorrect-usb-c-design-wont-work-with-some-chargers/

Raspberry Pi admits to faulty USB-C design on the Pi 4

"I expect this will be fixed in a future board revision," says co-creator.

@Learath2 the solution to that is: don't do nontrivial work in constructors

@heinrich5991 then you can have an object in invalid state

can you give an example?

e.g. I have a socket class, if it can't acquire a socket from the OS, it's left in a broken state

then you can add a private constructor that just takes the fd and a static member function that creates a socket

I don't like compensating for the shortcomings of languages

Just throw an exception from the constructor: https://isocpp.org/wiki/faq/exceptions#ctors-can-throw

I already do have a static optional<T> create()

only destructors are bad, see the next question in the FAQ

I don't like the exception pattern to begin with, but yes, it's the only proper solution in C++

with static member functions and trivial constructors, you basically mirror what rust does

I think the constructor should be private tbh

20:13

in that case atleast

20:13

The thing we are trying to avoid is that state where an object is constructed but isn't yet valid

But yeah, I also get pissed when exceptions are thrown all the time in regular usage

20:14

Because when I notice some exception that's not traced properly, I use gdb and "catch throw"

that works? ^^

yes, very cool

20:15

then I get a backtrace for every exception

I don't really like decoupling the errors from their context

So you want to handle potential failure in every possible location separately?

@deen I mean it's how I've been going about it for quite some time now, a lot of programs do that

Yes, and sometimes it makes sense. But if some kind of error has the same handling all the time, why repeat yourself?

20:17

then people used the "goto failure" pattern and that's what exceptions give you but more generalized and safer

20:17

And what if you forget to handle an error one time?

Well is it really safer?

In our case we have a component with lots of code by dozens of developers

20:18

and by having a catch(std::exception) around the central calling place we can make sure that no exceptions will leave

goto's can't get out of the functions bounds

20:19

Besides if exceptions even travel up a couple frames it becomes really not obvious how to handle them

20:19

As you lose context

20:20

They are indeed very useful when throwing things together, a very easy to work with pattern, but when you want to make sure things happen a certain way, I think it's harder to read

at least catch and log them is good for us already

Well tbh, what do I know, you work on software that needs to be much more reliable then what I write

Nah, that's not a good argument in my favour

20:21

Using exceptions for error handling makes your code simpler, cleaner, and less likely to miss errors. But what’s wrong with “good old errno and if-statements”? The basic answer is: Using those, your error handling and your normal code are closely intertwined. That way, your code gets messy and it becomes hard to ensure that you have dealt with all errors

https://isocpp.org/wiki/faq/exceptions#why-exceptions

Was reading around today, find if anyone else thought along my line and found this http://250bpm.com/blog:4

i remember that post

Well I guess it depends a lot on how well versed you are in a language, I find it's clearer with the if(!foo()){ handle_error(); } but that might just be bias on my part

20:24

rather the pattern then the language, to be more precise

I guess that also applies to any readability argument, experience factors in heavily and thus the argument is quite subjective