The Crystal programming language | http://crystal-lang.org | Fund Crystal's development: http://is.gd/X7PRtI | Docs: http://crystal-lang.org/docs/ | API: http://crystal-lang.org/api/
<SamantazFox_>
Unhandled exception: operator does not exist: text = text[]
<SamantazFox_> trying without the parenthesis
<SamantazFox_> (and yes, I've removed
args: from the query_all call)
<SamantazFox_> Nope, another error without parenthesis
https://crystal-lang.github.io/crystal-db/api/0.10.1/DB/QueryMethods.html#query(query,*args_,args:Array?=nil)-instance-method, assuming this applies to the other methods as well
<SamantazFox_> In my case, I want to provide the array directly
im sure theres a way...
is there an easy way to tell if an optional dependency (like GMP or regex) is installed in order to fall back on a different implementation?
Though that's about using
.so files at runtime, so maybe there's an easier way of just checking if they're there.
yea, i would rather do that at compile time (and link statically if at all possible). runtime deps are nothing but pain. having random binaries break after an
apt upgrade is no fun.
Running into an interesting issue trying to work with libgit2 bindings - libgit2 passes the same payload of typeVoid* to each remote callback when cloning
so to intercept each callback from crystal, I have to pass an object containing all the callbacks (essentially a {} of String => AliasForCallbackFunctions) for each type.
Boxing the object and passing it through causes an invalid memory access error, whereas not passing it does not. I've been able to pass single boxed blocks for payloads so far, wondering what I might be doing wrong here.
@skinnyjames: I'm fully sure that I could help you. also willing to do so. but can't look into it today. also maybe missing some details
1 reply
why do you need to pass an object containing all callbacks? why not just pass a class instance with normal methods?
I'd think that it should be
@on_credentials_acquirw = block
if so, that can be a getter on some object
and then you just box that object and pass that as Void*
I'm currently noticing that it blows up if I pass
Box.box("hello") as the payload
so it is probably unrelated?
weird 🤔
how do you unbox? maybe there's just some basic mistake ¯\_(ツ)_/¯
how do you unbox? maybe there's just some basic mistake ¯\_(ツ)_/¯
I think the bindings might be messed up? (I'm definitely still learning
the thing is, if you box something, you need to store that persistently as an instance variable or whatever. so that could be what's missing
bindings could be messed up too, who knows
if you could publish the code in a way that one can just run it and see the crash, that'd be quite a productive way to try to solve this. though i probably won't get to it today
1 reply
Was trying to debug, so I commented out the specs
curious though what this means:
I feel like I'm misinterpreting it
int (*)(git_remote_completion_t, void *) in the context of a callback function.I feel like I'm misinterpreting it
life improvement of the day (for rbenv users): http://mattgreensmith.net/2014/12/25/speed-up-rbenv-init-via-background-rehashing/
not going to have as much learning resources due to it being newer, also might not be as smooth as experience if you're on windows
something like:
options = Grits::Cloning::CloneOptions.default
options.fetch_options.on_credentials_acquire do |credential, url, username|
# handle credentials
credential
end
Grits::Repo.clone("git@github.com:skinnyjames/grits.git", "./directory", options)
go to thread