<rocketmagnet> hi everyone, when i creat a hashtable from string with a certain max lenght, what hash functino should i use ?
<VLetrmx> elichai2, right
* MarcinWieczorek (~marcin@dynamic-78-8-18-159.ssp.dialog.net.pl) has joined
* database2 has quit (Ping timeout: 245 seconds)
* lsm513 (~lsm5@93.170.51.236) has joined
* AmR|EiSa has quit (Remote host closed the connection)
<computerquip> rocketmagnet, sort of depends but FNV is a solid fast and general purpose hash.
* gloomy (~Gloomy@109.57.146.149.mobile.3.dk) has joined
<VLetrmx> elichai2, might be worth mentioning that Linux does have a mechanism that will let you effectively override functions, this is LD_PRELOAD, so depending on what platforms you're running on that might be an option.
* jamesbvaughan (~jamesbvau@2804:2728:101:83e:8524:92f3:d39b:c2fc) has joined
* lsm513 has quit (Ping timeout: 250 seconds)
<elichai2> VLetrmx: hmmm, and there isn't anything similiar in windows?
* AmR||EiSa (~AmR-EiSa@197.166.97.10) has joined
<VLetrmx> elichai2, Don't know sorry, I know almost nothing about Windows
<elichai2> yeah I get it, I don't use linux either but don't want to break compatibility just because it's easier
* jamesbvaughan has quit (Ping timeout: 252 seconds)
* royal_screwup21 (3309aae6@gateway/web/cgi-irc/kiwiirc.com/ip.51.9.170.230) has joined
<elichai2> VLetrmx: btw, if you're interested: https://github.com/rust-bitcoin/rust-secp256k1/pull/100#issuecomment-469002386
<badSophia> void sayHi(char name[])  <— why char name[] is not array?
<badSophia> i can’t understand it
<badSophia> :(
<computerquip> badSophia, because legacy C syntax says that's a pointer.
* form24 (~form@2804:2728:101:57d:95f6:6792:a36c:473f) has joined
<badSophia> i think that there is a reason
<badSophia> :(
<badSophia> i think that it is related with declaration?
<badSophia> computerquip: ?
<computerquip> No, you can't pass a C-array directly or by value.
<rocketmagnet> how to avoid collisions with hashtables ?
* Spinoza_ has quit (Remote host closed the connection)
* Spinoza_ (~Spinoza@97-91-132-136.dhcp.stls.mo.charter.com) has joined
<computerquip> rocketmagnet, hash algorithms are usually *good enough*. There's always a small chance you're going to collide at some point in your lifetime.
<fstd> lol
* fstd makes a mental note never to use a data structure written by computerquip
<VLetrmx> elichai2, worth mentioning that LDPRELOAD can't help you there since you want to override during compilation and LD_PRELOAD happens at runtime.
* badSophia has quit (Quit: badSophia)
<murii> Is there any problem if I declare a variable like this in a header file: char **a; then in a .c file I do: a = some_other_var ?
* Vante has quit (Ping timeout: 245 seconds)
<computerquip> fstd must not realize that unless you use a cryptographic function where a collision hasn't been found yet, you're always at risk of a potential collision.
<elichai2> VLetrmx: really? lol Ok, I'll delete that comment :)
<computerquip> But it's not practical to use SHA256 for a hash table algorithm.
* form24 has quit (Ping timeout: 257 seconds)
<murii> Qt Creator tells me this: no previous extern declaration for non-static variable. I just want to be sure everything is alright
<fstd> computerquip: have you ever actually implemented a hash table?  you seem to have some fundamental misconceptions there
<computerquip> fstd, like what?
<koollman> computerquip: even if a collision hash has not been found, you are always at risk of collions. by design
<koollman> if you had no collision risk, it would be a fantastic compression algorithm to hash any data :)
* wigust has quit (Quit: ZNC 1.7.2 - https://znc.in)
<fstd> which is why you need a way to deal with the collisions.  computerquip: for starters, i don't think people use cryptographic hashes for hash tables (but even if they would, a means for dealing with collisions would be a good idea)
* stefanc_diff has quit (Ping timeout: 268 seconds)
* BCMM has quit (Quit: Konversation terminated!)
<fstd> the last thing you want is data structures that can "lose" elements merely because someone deemed it unlikely enough
<fstd> rocketmagnet: the usual approach is to select a 'bucket' of elements with your hash, function
<fstd> essentially a list of entities stored that have the same hash
<koollman> you have to deal with collisions, because it's a hash table. unless you're creating a perfect hash table, but then it's a pretty different problem with know input
<fstd> to lookup, you find the bucket by hash, then find the desired element by comparing keys or somesuch
<koollman> *known
* CrystalMath (~coderain@reactos/developer/theflash) has joined
<computerquip> Not sure how any of that conflicts with what I said but sure.
<fstd> well you essentially said not to worry about collisions becuase they're unlikely enough, or did I misunderstand that?
<computerquip> You did.
<computerquip> But whatever.
<fstd> feel free to clarify then
<computerquip> You've already done that after insulting me for no reason.
<fstd> after taking another look, it doesn't seem like that's what you originally had in mind
<fstd> also where's the insult?
* georgios (~georgios@unaffiliated/anaphaxeton) has joined
* Wetmelon has quit (Ping timeout: 245 seconds)
<computerquip> Not sure what you expect when you start rebuking me in less than 30 seconds before I can finish with what I was saying.
<computerquip> I didn't know I was on a fucking timer here.
* johnflux has quit (Ping timeout: 255 seconds)
<koollman> computerquip: attention span on internet is like 5s. get to the point quickly ;)
* Psylent has quit (Quit: wheee)
* fstd shakes his head
<fstd> how would I even know that it was but an introduction to get to your real point
<fstd> assuming that's what you actually were going to do, that is
<fstd> might also be the usual getting-defensive-over-being-wrong-and-backpedaling..
* fstd shrugs
<computerquip> I guess I wouldn't want to work with you in a work place.
<fstd> trust me, the feeling is mutual
<computerquip> I hope nobody does frankly.
<fstd> you hope nobody wants to work with me?  uh, okay
<rocketmagnet> how can i compute a hascompute function if i know the exaclt lenght of my strings ?
<fstd> says more about yourself than it does about me, tbh
<computerquip> Actively trying to shit on people as fast as you can, not a very professional outlook there guy.
<fstd> if you're frequently wrong, you need to learn to not take it as a personal offense when you get called out
* wigust (~wigust@178.71.234.40) has joined
<fstd> otherwise you might end up in the place you hope me to be in -- becasue nobody likes to work with those people
* mowcat (~mowcat@cpc105070-sgyl40-2-0-cust220.18-2.cable.virginm.net) has joined
<computerquip> fstd, I wasn't even wrong which is the problem.
* mowcat has quit (Max SendQ exceeded)
* mowcat (~mowcat@cpc105070-sgyl40-2-0-cust220.18-2.cable.virginm.net) has joined
<computerquip> You went out of your way to shit on me for no reason.
<computerquip> Am I wrong?
<computerquip> Why would I work with someone who does that?
<fstd> i just got the impression that you're trying to give advice on something you're not familiar with, which is often a bad idea
<computerquip> So your first instinct is to insult them?
* Vante (~Vant@host87-250-15-075.etth.mark-itt.net) has joined
<computerquip> Not, "what do you mean by that" or, "I think that's wrong" but to insult them.
* Me4502 has quit (Read error: Connection reset by peer)
<fstd> where's the insult?  [again]?
<fstd> i didn't say "what do you mean by that" becasue it seemed perfectly clear what you meant by that
<computerquip> You said you would never use a data structure written by me not to mention the other comments.
<fstd> how is that an insult?  and what other comments?
* gist (~jack@88.130.155.248) has joined
* Vante has quit (Client Quit)
<computerquip> "fstd: computerquip: have you ever actually implemented a hash table?" You then proceeded to be more offensive by claiming I was being defensive when the conversation was over.
* stefanc_diff (~stefanc_d@cpc97944-croy24-2-0-cust250.19-2.cable.virginm.net) has joined
<fstd> i asked that question not to make a point, but because i actually wondered about that
<fstd> maybe try to not take *everything* as a personal attack?
<fstd> that said, i've missed the 'But it's not practical to use SHA256 for a hash table algorithm' earlier
<fstd> so there's that.
<computerquip> It's not a personal attack until you start directing insult directly towards me
<computerquip> Not sure how that's hard to understand
<computerquip> When I start having to be defensive about what I'm saying before I can even finish with what I'm saying, you're not going to get the full explanation.
<fstd> we seem to have different ideas about what constitutes an insult
<computerquip> I would never use code from fstd because because he's probably never written in C before.
<fstd> computerquip: you need to resist hitting the return key then, when your follow-up message is so vital
<computerquip> ^ how the fuck is that not an insult
<fstd> computerquip: i don't feel insulted by it, for starters
* computerquip walks away because holy shit.
<fstd> ,next
<cantdide> another satisfied customer! NEXT!!
* RebelCoder (~RebelCode@62.208.39.206) has joined
* Mattiaslndstrm has quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
* talmobi has quit (Ping timeout: 240 seconds)
<dorp> computerquip: If somebody said: "UB is a myth, everything is safe" ... and I would respond: "I would rather not use a program written by somebody who thinks that" ... would you still consider that to be an attempt by me to insult that person?