DDraceNetwork - developer

i have a question about bam (not ddnet related specifically)

00:21

sec

00:21

00:21

but clang is installed so im confused

00:22

configure.lua should be the exact same one tw uses

i also didnt got clang to work back then

00:26

is there no cmake alternative?

i dont know much about cmake but this repo isnt really mine so i dont know if i could write one for it

00:30

ill just try gcc

00:31

wat gcc is installed already

00:31

how do i tell bam to use gcc

CC=gcc CCX=g++

00:31

as env flags

thanks

or CXX

still says global error

00:32

hrmm

00:32

can i dm you so i dont flood this channel

u need to delete the configuration.lua probs first

o

config.lua is it

ahhh yes it says clang in there

00:33

ill change it

00:33

or i can remove it and it will generate for my env?

00:34

ill just change the value to 'gcc' and see if that works

00:34

oh hey that worked i think

00:35

hah, worked

00:36

thanks a bunch

@Deleted User

Deleted User

I didn't really made it, but what's broken about it?

can't change the size on the fly in gl3

TsFreddie

can't change the size on the fly in gl3

strange works for me xd

ddnet-maps

[ddnet-maps:master] 1 new commit

a6f556c A Retro - ddnet-maps

TsFreddie

[ddnet/ddnet] Pull request opened: #3818 Remove a Dangling Declaration

Don't know why these are there

Checklist

[ ] Tested the change ingame
[ ] Provided screenshots if it is a visual change
[ ] Tested in combination with possibly related configuration options
[ ] Written a unit test if it works standalone, system.c especially
[ ] Considered possible null pointers and out of bounds array indexing
[ ] Changed no physics that affect existing maps
[ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/dd...

def-

[ddnet-web:master] 1 new commit

d89fcef Add 10 € donation by PlantKnight - def-

bors[bot]

[ddnet:master] 2 new commits

796035a Remove dangling declaration - TsFreddie ba6759f Merge #3818 - bors[bot]

08:03

bors[bot]

[ddnet/ddnet] Pull request closed: #3818 Remove a Dangling Declaration

10:26

one day ill parse the map format using parser combinators like this

Someone, please fix the parsing of /times "number name". The behavior is incorrect if we want to query a number ID. For example, /times "950" will give the finish history of last 950 fin(which is the same to /times 950) but we want to query the finish history of player "950"(which is a chn guy now named "rhino"). Now pure number ID is unable to query finish history but names with a blank can.

21:40

/times "ano :3" behaves correctly.

21:45

Also need to check if other commands work correctly with quotes

@Night_L the parsing is "correct" in its own way

21:52

quoting doesn't change the way parameters are treated

@Night_L /times 950 1

21:53

will show you the last finishes by player named 950

In general you can use /help times to find out stuff like that

I see. But default behavior is inconsistent between /times "name with a blank" and /times "pure-number". If this is designed to be, it is somekind confusing. (edited)

indeed

I would agree this is not a designed calling of the command (but it just, doesn't fit)

22:15

This might be a out-of-consider case in the pipeline of parsing. IDK

it's the weird choice the first person to implement it made

In essence the command is both times ?s[name] ?i[numbertoskip] and times ?i[numbertoskip] when name is. When name is just numeric it's a parser conflict

to treat numbers differently from names

22:16

when numbers can obviously also be names

I don't like it either but I doubt we can change it without someones mothers cousins uncles brothers bind breaking

Learath2

In essence the command is both times ?s[name] ?i[numbertoskip] and times ?i[numbertoskip] when name is. When name is just numeric it's a parser conflict

There is quotes when querying ID

22:17

or quotes doesn't work as I expect??

Quotes don't imply string. Maybe they should? I'm not completely sure of the implications if we do that

I know now.

22:18

But, /times "12 3" doesn't make sense unless "12 3" refers to a ID instead of [numbertoskip]

"12 3" should work as expected

I wouldn't be so sure of that

I mean, numbers won't contain blanks inside themself.

"12 3" will make it evaluate as one string argument

22:21

unless the person who made the command considers ' ' numeric I can't see it doing times ?i

so /times "123" is not that useful since /times 123 will work ofcuz..

nah, 12 3 is going to be an offset

22:22

as is 123abc

22:22

if I read the code correctly

22:22

pResult->GetInteger(0) != 0

22:23

and that's straight up atoi

22:23

praise C's shitty integer parsing functions and that we use them

Learath2

Quotes don't imply string. Maybe they should? I'm not completely sure of the implications if we do that

Now I am rly confused by parsing haha. I suppose "abcd" implies abcd is treated as an ID no matter what abcd is. The reason is quite simple: quoting a number doesn't make sense, we only quotes ID since there are names with blanks

22:27

Do we have anything else in commands other than "s t r i n g" and number?

are you interested in why things are the way they are?

22:29

I agree that the current syntax is bad

22:29

I think that problem appeared because the /times function tried to stuff two possible parameters into one

what happens if a name contains quotes

quotes are used to separate parameters

22:31

you can do "\""

22:32

that is used for the name that is a single doublequote

im pretty sure there was a name u couldnt check /times of

heinrich5991

I think that problem appeared because the /times function tried to stuff two possible parameters into one

Probably yes. But things can be fixed if we suppose quoting a number directly refers to a string, a name. So things will become quite simple :< (edited)

Night_L

Now I am rly confused by parsing haha. I suppose "abcd" implies abcd is treated as an ID no matter what abcd is. The reason is quite simple: quoting a number doesn't make sense, we only quotes ID since there are names with blanks

The callback function knows nothing of quotes. The console parses that. So quotes only enforce that no matter what, the things within the quotes will be treated as one parameter

Learath2

The callback function knows nothing of quotes. The console parses that. So quotes only enforce that no matter what, the things within the quotes will be treated as one parameter

k. So not able to fix it directly, we should seperate /times into two commands i guess

22:42

Or just make it legacy and nevermind

22:42

Anyway, I now know the problems behind. NVM

Also, the part that does this parsing knows nothing of the concept of names

22:42

I wanted to teach it but just haven't had the time to hash out a plan

I guess one could stop accepting offsets in there

22:43

I feel like that would be the cleanest fix

@heinrich5991 but how do we replace times ?i[numskip] then?

22:44

a new command?

/times heinrich5991 5

/times 5 is what I was concerned about

oh right, it returned recent finishes

22:44

new command?

22:45

/recent

Yeah I think that's the cleanest

22:45

or maybe /finishes

ah didn't know it supported escape sequences

22:47

though I could have expected that based on the fact that motd does (edited)

the motd has its own treatment of escape sequences

22:49

in the weirdest way possible

22:49

the character sequence \n is translated into a newline by the client(!)

ohh interesting

Learath2

The callback function knows nothing of quotes. The console parses that. So quotes only enforce that no matter what, the things within the quotes will be treated as one parameter

em, I think the console parser could parse /cmd "some thing" 123 into cmd ; "some thing"; 123. So the pipeline gets information of quotes. Just do not break quotes without escapes and split by blanks. (Well, this might need an automata to realize..)

any recommendations for simple web servers

22:55

I might give cobalt.rs a try

what do you want to build?

lynn

any recommendations for simple web servers

cobalt.rs seems to be a static site generator not a web server

well it also serves the files

Well it seems to have a live preview thing like most new frontend frameworks have provided by tiny_http. I wouldn't host anything like that in production no matter how small

Night_L

em, I think the console parser could parse /cmd "some thing" 123 into cmd ; "some thing"; 123. So the pipeline gets information of quotes. Just do not break quotes without escapes and split by blanks. (Well, this might need an automata to realize..)

I might be able to deal with it with some lexer. After my thesis defence :<

Learath2

Well it seems to have a live preview thing like most new frontend frameworks have provided by tiny_http. I wouldn't host anything like that in production no matter how small

nginx is not that difficult to set up and is small, especially trivial if you are hosting a static website

Night_L

I might be able to deal with it with some lexer. After my thesis defence :<

The problem isn't that it's unsolvable. The problem is whether we want it solved. This seems to be more of an issue on the person implementing the command trying to pack too much into one command

23:05

Quotes imply string would also fix this without messing with the parser possibly breaking dozens of things

nginx gives me heartburn

23:09

it's really going to be a single page with links / IP addresses. https://teehou.se

23:09

unreachable for now

just bite the bullet and learn it, it's a skill useful for a lifetime

23:12

for a static website it's absolutely trivial as I said