<whyrusleeping>
in the namespace datastore, we just query the child
<whyrusleeping>
and then filter results coming out
<kevina>
I thought it was the mount datastore?
<whyrusleeping>
the problem is with the namespace datastore not being able to understand the mount datastore
<whyrusleeping>
if the namepace datastore has a namespace of /blocks
<whyrusleeping>
and its wrapping a mount datastore that has a /blocks mount
<whyrusleeping>
then we should be able to short circuit that, no?
<whyrusleeping>
the biggest issue is querying a namespace datastore querying
<kevina>
maybe
<whyrusleeping>
whoops, english
<kvda>
does ipfs.io accept api connections?
<kevina>
okay I will try to investigate
<kvda>
ie can you use py-ipfs-api to connect to it
<whyrusleeping>
kevina: yeah... i'm still not 100% sure
<kevina>
it has been two months since I last worked with the code
<whyrusleeping>
but i was noticing some weird stuff when querying on leveldb and blockstore stuff
<whyrusleeping>
kvda: theres a readonly api
<kvda>
ah ok i need write, thanks whyrusleeping
<kevina>
whyrusleeping: so are you _sure_ that both datastores are being queried?
<whyrusleeping>
kevina: not 100% sure no
<kevina>
so what where you noticing than?
<kevina>
I know there is a lot of unnecessary path rewriting, but I worked to optimize that, I didn't notice the behavior you describe of querying both datastores
<whyrusleeping>
hrm... then idk what i was having trouble with...
<kevina>
"idk" = i don't know?
tclass has quit [Ping timeout: 240 seconds]
<whyrusleeping>
correct
<kevina>
whyrusleeping: ok it sounds like there is nothing for me to do on this, should we close this bug, or wait until you figure out what the real problem is you where seeing?
<whyrusleeping>
Yeah, we can close it. Though the key mangling back and forth is annoying, its not a perf issue at this point
<whyrusleeping>
sorry about that
<kevina>
no problem
HostFat__ has joined #ipfs
<kevina>
I'm just glad I don't have to mess with go-datastore again, setting things up the last time was a major PITA :(
<whyrusleeping>
yeah... i need to figure out how to make gx workspaces a little easier
<whyrusleeping>
in the simple case you just `gx-go rewrite --undo PACKAGEYOUWANTTOWORKON`
<kevina>
this wasn't a simple case
<kevina>
to many things depended on go-datastore from go-ipfs that I had to do some ugly hacks
<kevina>
the details of which I forget :)
HostFat_ has quit [Ping timeout: 256 seconds]
wallacoloo____ has joined #ipfs
<whyrusleeping>
yeah...
<whyrusleeping>
what i'm thinking of doing as a better solution
<whyrusleeping>
is to do `gx install --local`
<whyrusleeping>
and then symlink the repo you want to work on into place under ./vendor
<whyrusleeping>
it *should* work as expected
<kevina>
whyrusleeping: I am not 100% following, but if you can spell out the details in a issue I might be able to help implement it...
<whyrusleeping>
yeah, i'm still not sure on the UX yet
<whyrusleeping>
i'll see about making an issue about it
<kevina>
okay
<pjz>
whyrusleeping: how goes 0.4.5 ?
<whyrusleeping>
going pretty well, got the migrations finished
<whyrusleeping>
theres a release of them up on dists
<whyrusleeping>
going to cut an rc1 tomorrow i think
<whyrusleeping>
then let that simmer for a few days
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
mguentner has joined #ipfs
cretz[m] has joined #ipfs
<cretz[m]>
newb dev question/thought...I was wanting to use go libp2p's kad dht impl to find peers for a possible tor onion-like routing system
<cretz[m]>
but it seems get closest peers has a limited peer set size and I see no way, externally, to clear out the peerset, would the ability to clear a peer set or provide a filter to lookups be welcome?
tmg has joined #ipfs
<cretz[m]>
(I need filtering to abide by some circuit path selection constraints, ala no peers in same /16's, etc)
<lgierth>
yes that sounds useful
<lgierth>
there's already an address filtering mechanism somewhere else in the code that you can reuse
<cretz[m]>
I saw it internally in the query
<lgierth>
e.g. /ip4/172.16.0.0/ipcidr/16
<lgierth>
it sounds useful to somehow wrap that around the peerstore
<cretz[m]>
true since that comes from the host object anyways
<lgierth>
also the openbazaar people have been doing a lot of work with libp2p and tor/onion
<lgierth>
(they built a libp2p onion transport)
<cretz[m]>
link?
<cretz[m]>
I will look that up
<lgierth>
eeh github/openbazaar/go-onion-transport i think
<substack>
kvda: most of my stuff lately uses yo-yo
<substack>
which uses hyperx and morphdom
<kvda>
ah yup, choo is based on that
<substack>
yes, a lot of stuff happened after I whipped up hyperx
<substack>
nice to see
<kvda>
:)
<substack>
I think choo seems like a good fit for much more fully-featured, traditional webapps with many people who need to coordinate on architecture
<substack>
I just don't often work on such projects, so I haven't had as much of a chance to evaluate it fully
<kvda>
yup very true
<kvda>
that's what it is, but it's still pretty tiny
<kvda>
so how is neocities still around, no-ads etc, are they doing the whole get users -> vc money -> bail thing
tclass has joined #ipfs
<substack>
no, it's incorporated as an oregon public benefit llc
<substack>
freemium, you can pay for a supporter plan and get custom domains, way more space
<substack>
it also backs up to ipfs automatically and kyledrake is very responsive to issues
<SchrodingersScat>
is there a file I can pin to save the climate?
ashark has joined #ipfs
<richardlitt>
SchrodingersScat: Hypothetically, there may be a hash that would save everything. We're still working on making the right algorithms to find that hash.
<SchrodingersScat>
42
<richardlitt>
SchrodingersScat: We think it's probably a multihash. Unfortunately, we don't know the multicode.
john3 has joined #ipfs
pfrazee has joined #ipfs
john2 has quit [Ping timeout: 256 seconds]
john4 has joined #ipfs
john3 has quit [Ping timeout: 240 seconds]
__uguu__ has quit [Read error: Connection reset by peer]
[0__0] has quit [Remote host closed the connection]
<__uguu__>
it's been like that for over a decade at least
<nicolagreco>
__uguu__: it wasn't that "official"
<__uguu__>
level of "official" is irrelevant
<__uguu__>
but... blah blah blah offtopic
wmoh has joined #ipfs
<SchrodingersScat>
so would a setup be having an rpi ipfs box with an external drive for .ipfs, then point the firefox/chrome addons to point to that local machine for ipfs access?
Nukien has quit [Ping timeout: 252 seconds]
realisation has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<__uguu__>
would an rpi have enough juice to handle that?
<SchrodingersScat>
what resource are you worried about?
<__uguu__>
cputime memory and disk io
<SchrodingersScat>
would it use more than a gig of memory?
Encrypt_ has joined #ipfs
Foxcool__ has quit [Ping timeout: 255 seconds]
Nukien has joined #ipfs
realisation has joined #ipfs
realisation has quit [Client Quit]
G-Ray_ has quit [Quit: G-Ray_]
cwahlers has joined #ipfs
espadrine_ has joined #ipfs
john2 has joined #ipfs
john3 has joined #ipfs
john1 has quit [Ping timeout: 260 seconds]
s_kunk has joined #ipfs
john2 has quit [Ping timeout: 252 seconds]
mildred1 has quit [Ping timeout: 276 seconds]
ygrek_ has joined #ipfs
realisation has joined #ipfs
Baughn has quit [Quit: ZNC 1.6.2+deb1 - http://znc.in]
Baughn has joined #ipfs
pcre has quit [Ping timeout: 255 seconds]
gully-foyle has joined #ipfs
bluetriangle3 is now known as vapid
<whyrusleeping>
Yeah, it does end up using more than a gig of ram right now
<whyrusleeping>
working on bringing that down
Encrypt_ has quit [Quit: Quit]
vapid is now known as arabs_did_911
<SchrodingersScat>
surprising, i didn't notice it using much
arabs_did_911 is now known as vapid
john4 has joined #ipfs
john3 has quit [Ping timeout: 264 seconds]
john has joined #ipfs
john is now known as Guest88753
tclass has joined #ipfs
john4 has quit [Ping timeout: 256 seconds]
hannes[m]1 has joined #ipfs
ligi has joined #ipfs
ligi has joined #ipfs
ligi has quit [Changing host]
jkilpatr has quit [Ping timeout: 240 seconds]
matoro has joined #ipfs
<kpcyrd>
SchrodingersScat: ensure you have swap configured on low memory systems
matoro has quit [Remote host closed the connection]
<SchrodingersScat>
ZRAM
matoro has joined #ipfs
<kpcyrd>
whyrusleeping: do we know what is eating all the memory?
<kpcyrd>
that'd be interesting
<whyrusleeping>
mostly yamux
<kpcyrd>
interesting
<whyrusleeping>
thats why i've added multiplex as an experimental feature in this next release
tclass has quit [Remote host closed the connection]
tclass has joined #ipfs
Encrypt_ has joined #ipfs
Encrypt_ is now known as Encrypt
tclass has quit [Ping timeout: 248 seconds]
tilgovi has joined #ipfs
jkilpatr has joined #ipfs
tclass has joined #ipfs
iovoid has quit [Quit: Iovoid has quit!]
airsickpayload[m has joined #ipfs
iovoid has joined #ipfs
Gergely[m] has joined #ipfs
tilgovi has quit [Ping timeout: 240 seconds]
tclass has quit [Remote host closed the connection]
<whyrusleeping>
kevina: is something actually wrong with initDefaultDatastore?
<kumavis>
whyrusleeping does http-api support this by chance?
<whyrusleeping>
kevina: it seems like the bug is in the Writeable function not cleaning up after itself
<whyrusleeping>
kumavis: yeap, all cli flags match the http api
<kumavis>
cool
<whyrusleeping>
(though there are a couple exceptions where args are clientside, such as ipfs get)
john1 has joined #ipfs
<kevina>
whyrusleeping: writable already should remove expect if the os.Create fails
<kevina>
initDefaultDatastore could just call flatfs.Create instead of creating an empty directory
Guest88753 has quit [Ping timeout: 260 seconds]
cyanobacteria has quit [Ping timeout: 240 seconds]
<whyrusleeping>
right, it could, but why is that file lingering around then?
<whyrusleeping>
that bit doesnt make sense to me
john2 has joined #ipfs
<whyrusleeping>
i do agree we can fix this by not calling writeable, but i want to try and figure out the issue with writeable as well
<kevina>
I don't know it should not be. I can fix it so that os.Remove(f.Name()) is always calls delete (even if os.Create fails) but that is all I can do
<whyrusleeping>
kumavis: wow, the fact that the go vendor stuff is 'vendor/packagename' and not 'vendor/src/packagename' is REALLY annoying...
<whyrusleeping>
oop
<whyrusleeping>
Kubuxu: not kumavis
<Kubuxu>
why?
john1 has quit [Ping timeout: 240 seconds]
<whyrusleeping>
because to install go-ipfs into a vendor directory, i cant set GOPATH=`pwd`/vendor
<whyrusleeping>
that puts it in vendor/src
rendar has joined #ipfs
rendar has joined #ipfs
rendar has quit [Changing host]
FrederickSilva has quit [Remote host closed the connection]
john3 has joined #ipfs
john2 has quit [Ping timeout: 252 seconds]
john4 has joined #ipfs
john3 has quit [Ping timeout: 276 seconds]
matoro has quit [Remote host closed the connection]
wallacoloo____ has joined #ipfs
jkilpatr has quit [Ping timeout: 245 seconds]
draynium has joined #ipfs
saintromuald has quit [Ping timeout: 264 seconds]
kresbach_ has joined #ipfs
<kresbach_>
ahh it's a wonderful day
draynium has quit [Ping timeout: 248 seconds]
rendar has quit [Quit: std::lower_bound + std::less_equal *works* with a vector without duplicates!]
lgierth has quit [Quit: WeeChat 1.4]
lgierth has joined #ipfs
Encrypt has quit [Quit: Quit]
tilgovi has joined #ipfs
galois_d_ has joined #ipfs
rmi7[m] has joined #ipfs
saintromuald has joined #ipfs
ianopolous_ has quit [Ping timeout: 248 seconds]
galois_dmz has quit [Ping timeout: 240 seconds]
kulelu88 has joined #ipfs
ianopolous_ has joined #ipfs
galois_d_ has quit [Remote host closed the connection]
galois_dmz has joined #ipfs
tclass has quit [Remote host closed the connection]
tclass has joined #ipfs
jkilpatr has joined #ipfs
tclass has quit [Ping timeout: 258 seconds]
slothbag has joined #ipfs
cemerick has quit [Ping timeout: 248 seconds]
ligi has quit [Quit: leaving]
draynium has joined #ipfs
palkeo has quit [Quit: Konversation terminated!]
wallacoloo____ has quit [Quit: wallacoloo____]
ashark has quit [Ping timeout: 256 seconds]
kvda has joined #ipfs
pfrazee has quit [Remote host closed the connection]
Caterpillar has quit [Quit: You were not made to live as brutes, but to follow virtue and knowledge.]