<asheesh>
Arguably an interesting Sandstorm app: applicant tracking software, http://www.opencats.org/
<asheesh>
I like the way this page contrasts self-hosted vs. the cloud version of this FLOSS package: http://www.recruity.com/en
<asheesh>
: D OK pseudo-collections are cool.
jadewang has quit [Remote host closed the connection]
<docaedo>
I'm interested in running sandstorm apps on OpenStack clouds - other than running the install script in a VM, has anyone looked at automating that specifically for openstack environments?
<asheesh>
No, but I'd love to see that happen.
<asheesh>
What would you need for you to consider that a success?
<docaedo>
Heh, not sure :) Maybe something as simple as a heat template - but TBH I haven't played much with it.
aldeka has quit [Remote host closed the connection]
ocdtrekkie has quit [Remote host closed the connection]
<asheesh>
Ooh, +1 to heat template
<asheesh>
People often think an app like Sandstorm is designed to run on mutiple machines, so even though heat is optimized for single-machine, it'd be OK in this case.
<asheesh>
Curious what you would expect to do for hostname!
<asheesh>
(and https)
<docaedo>
I was planning to spend some time tomorrow or Friday playing more sandstorm and getting myself a little more educated on all the sandstorm things (like hostnames, how the .. grains? are managed, etc.)
<docaedo>
I work on the app catalog we have (https://apps.openstack.org) and would love to figure out a way to make it easier for people to go there and end up with the bits they would need to consume sandstorm apps
<docaedo>
I'll speak up here if I have questions when I get time to play around with it - but I love what you're all building :)
<asheesh>
Awesome! : D
aldeka has joined #sandstorm
aldeka has quit [Ping timeout: 246 seconds]
|jemc| has quit [Ping timeout: 260 seconds]
simonv3 has quit [Quit: Connection closed for inactivity]
ocdtrekkie has joined #sandstorm
aldeka has joined #sandstorm
|jemc| has joined #sandstorm
jadewang has joined #sandstorm
bb010g has joined #sandstorm
[d__d] has joined #sandstorm
natea has joined #sandstorm
[d__d] has quit [Remote host closed the connection]
[d__d] has joined #sandstorm
simonv3 has joined #sandstorm
home has quit [Quit: Leaving]
kecolus has quit [Quit: Leaving]
simonv3 has quit [Quit: Connection closed for inactivity]
KCinJP has joined #sandstorm
natea has quit [Quit: natea]
isd has joined #sandstorm
home has joined #sandstorm
|jemc| has quit [Ping timeout: 245 seconds]
ocdtrekkie has quit [Remote host closed the connection]
aldeka has quit [Remote host closed the connection]
ocdtrekkie has joined #sandstorm
aldeka has joined #sandstorm
isd has quit [Quit: Leaving.]
<larjona>
Good morning TZAG dear sandstorm friends! Is there any place where I can see the changelog or summary of new features etc before upgrading an app? (or after!) I got several recent updates for Davros and not sure where to look to see the improvements
aldeka has quit [Remote host closed the connection]
ocdtrekkie has quit [Remote host closed the connection]
ocdtrekkie has joined #sandstorm
aldeka has joined #sandstorm
myguidingstar has left #sandstorm ["Anything in parenthesis can (not) be ignored."]
natea has joined #sandstorm
natea has quit [Client Quit]
natea has joined #sandstorm
jadewang has quit [Remote host closed the connection]
natea has quit [Quit: natea]
simonv3 has joined #sandstorm
|jemc| has joined #sandstorm
natea has joined #sandstorm
natea has quit [Client Quit]
natea has joined #sandstorm
natea has quit [Client Quit]
natea has joined #sandstorm
<ocdtrekkie>
larjona Check mnutt's CHANGELOG.md
<ocdtrekkie>
He doesn't always publish a release on Github, but the code and changelog is up to date.
<larjona>
fine! thank you
<ocdtrekkie>
Sandstorm packages now commonly do include changelogs, but the app store doesn't display them yet. Until it does, finding CHANGELOG.md will be your best bet.
<larjona>
thanks
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 260 seconds]
|jemc| has quit [Ping timeout: 260 seconds]
bb010g has quit [Quit: Connection closed for inactivity]
jadewang has joined #sandstorm
jadewang has quit [Ping timeout: 260 seconds]
|jemc| has joined #sandstorm
notevil has joined #sandstorm
larjona has joined #sandstorm
funwhilelost has joined #sandstorm
jadewang has joined #sandstorm
chris_severs has quit [Ping timeout: 264 seconds]
chris_severs has joined #sandstorm
kecolus has joined #sandstorm
rustyrazorblade has quit [Quit: rustyrazorblade]
funwhilelost has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<|jemc|>
what is the theoretical maximum size of a capnp segment?
<|jemc|>
I know that there will likely be some practical restrictions based on the byte width of the numbers that define offsets and such, but is there a formally defined limit to segment size?
<|jemc|>
also thinking about how this plays into the notion that the "canonical" form of a message is allocated to only one segment - does this impose restrictions on how big (or deep) a message can be?
<zarvox>
Looks like Firefox Accounts uses Hawk request signatures (Authorization: Hawk <comma-separated list of key-value pairs for id, ts, nonce, hash, ext, mac>)
<zarvox>
so unless we teach the proxy about Hawk ( https://github.com/hueniverse/hawk ) I don't think we can run a Firefox Accounts server instance
<XgF>
maurer: You run into pointer-encoidng issues when a list reaches that size anyway
<maurer>
XgF: that is what I was suggesting would be awkward...
<maurer>
XgF: not the java impl limits, but the pointer limits
<XgF>
(but the only practical csae which occurs to me as a problem is storing files in capnp)
<asheesh>
zarvox: I think teaching the proxy about Authorization: Hawk seems within reason.
<asheesh>
Conveniently it's not something browsers know how to do for web pages!
<maurer>
XgF: It's probably a non-issue for all the uses I'm aware of today, but for example consider what you'd get if you decided to use capnp as a container format for video
<maurer>
XgF: it'd be well suited to it other than that restriction, due to incremental decode
<XgF>
maurer: Still fine. Video is packetized
<maurer>
XgF: packets are probably structs in a big list
<XgF>
no, packets are variable length
<XgF>
(so if they are structs in a list, the structs will contain pointers to their actual data)
<maurer>
OK, so you suspect that'd still fit?
<XgF>
yes
<XgF>
Like, you'd be packing the per-frame data as e.g. bytes one frame at a time
<zarvox>
asheesh: I'm not sure that's possible given the protocol
<zarvox>
it's HMAC'd end-to-end
<zarvox>
at the very least we'd wind up storing the secrets for the HMAC'd requests
funwhilelost has joined #sandstorm
<asheesh>
Fascinating, zarvox.
<zarvox>
It's also not clear to me how we'd bridge that to Firefox, whose sync code we probably can't reasonably modify and which expects flows for account creation and so on.
<asheesh>
I imagine there's some way to do something self-hosted something.
<zarvox>
This is already in the realm of "edit your about:config" though
<zarvox>
There is, I just don't know if it's compatible with Sandstorm's accounts system and permissions and protocols.
<zarvox>
there are two pieces: the sync server relies on some auth infrastructure provided by fxa (Firefox Accounts)
<asheesh>
Ah hah, "Enter “about:config” in the URL bar, search for items containing “fxaccounts”, and edit them to use your self-hosted URLs"
<maurer>
I suspect the best way to make something like a sync grain "producty" would be to make the sync grain autogenerate a ff extension that edits about:config correctly for you
<maurer>
so that you could punch a button on the sync grain, and have your ff swap to it
<zarvox>
But then the extension would be unsigned, which would make it only usable if you enable testing extensions, I think?
<maurer>
Oh, I suppose so :(
<maurer>
I forgot that I have those on
<maurer>
Dunno what would be best then
erikmaarten has joined #sandstorm
<asheesh>
"Obviously" it should submit to the Fx app market thing an extension called "Asheesh's Sync Settings"
<asheesh>
"Obviously" it should submit to the Fx app market thing an extension called "{{profile.name}}'s sync Settings" rather
<maurer>
asheesh: No, you had it right the first time
* asheesh
chuckles.
<maurer>
all users should install a differently built version of "Asheesh's Sync Settings"
<maurer>
:)
<asheesh>
FWIW we could make an addon called "Change Sync Settings"
<asheesh>
we/you/etc could make that addon.
<asheesh>
And it could support a special URI scheme that lets people choose the sync settings.
<maurer>
Yeah, that'd make sense
<asheesh>
rather, lets people specify the sync settings as URI parameters so the Sandstorm app can push the info out.
<maurer>
you'd need to be careful with how it worked though to make sure it couldn't cause people to swap sync servers when links got hit
<maurer>
without their knowledge
<erikmaarten>
hello everyone!
<zarvox>
hi erikmaarten!
<asheesh>
Likewise erikmaarten
<asheesh>
Oh snap Thanksgiving ended but I didn't finish giving you advice.
<erikmaarten>
asheesh: I've gotten Polytunes to work decently in two-player, think I just need one more session to polish it a bit and get it to work in singe mode
<erikmaarten>
:)
<asheesh>
Oh, awesome!
<erikmaarten>
I still haven't figured out that problem though
<erikmaarten>
the Sandstorm env variable thing I told you about
<asheesh>
Yeah!
<asheesh>
Let's see.
<asheesh>
Maybe I can take a sec to vagrant-spk dev what you have.
<erikmaarten>
zarvox: ah, yeah, I think I followed your example and did the same with the board game app I made
* asheesh
wanders off for a few minutes and hopes you two can figure it out.
<zarvox>
if Meteor saves the settings from the environment into Mongo on first-run, then you're in good shape. if not, maybe you can modify your app to explicitly save the settings from the environment to the DB on startup
<erikmaarten>
but I think that's not the issue at the moment. When starting up the app, regardless of which command used, there will always be the env var from the two-player command
<erikmaarten>
When starting a new game, that is
<zarvox>
Oh. That's weird, and sounds like a bug.
<erikmaarten>
Or at least as far as I can see, that's what happens
<erikmaarten>
there's definitely an inkling of a possibility that I made a mistake somewhere too :D
mnutt__ has quit [Ping timeout: 260 seconds]
<asheesh>
I must AFK again a bit, but will be back in half an hour or so.
kecolus has quit [Quit: Leaving]
mnutt__ has joined #sandstorm
<erikmaarten>
I'm going to bed, will continue tomorrow or during the weekend. I guess it could possibly be a bug...
KCinJP has joined #sandstorm
erikmaarten has quit [Quit: Leaving]
notevil has quit [Quit: import antigravity]
jadewang has quit [Remote host closed the connection]
jadewang has joined #sandstorm
natea has quit [Quit: natea]
<kentonv>
dwrensha: I just noticed that github now reports whether email addresses are verified!
<kentonv>
dwrensha: in user.services.github. This seems new. Can we use it?
<dwrensha>
nice!
<kentonv>
it's in user.services.github.emails (plural), which I think only now shows up because you(?) changed it to request the user's private emails.
<dwrensha>
should be straightforward
<dwrensha>
where is that request?
<dwrensha>
I don't think I changed any such thing
mnutt__ has quit [Ping timeout: 245 seconds]
<kentonv>
maybe Meteor changed it, then
<kentonv>
but if you log out and log back in on a recent dev server you'll see it request a new permission: "see your private email addresses"
<kentonv>
(or something like that)
<dwrensha>
oh, I do think I remember wondering why I got a screen like that
<dwrensha>
recently
<kentonv>
anyway, we now get a *list* of emails from github, some of which may be verified and some not
<dwrensha>
so Meteor is automatically populating it, even for users that have existed for a while?
<kentonv>
it'll only be populated the next time they log in
|jemc| has quit [Ping timeout: 246 seconds]
<dwrensha>
oh, I wonder if the username and such are also updated then
<dwrensha>
I guess they probably are
<kentonv>
probably, yes
<kentonv>
github also reports which email was primary
<kentonv>
can our code handle one identity providing multiple verified emails?
<dwrensha>
SandstormDb.getVerifiedEmail() -> SandstormDb.getVerifiedEmails() would not be too hard
<dwrensha>
since the latest migration we don't actually store anything related to the verified email in identities
<dwrensha>
we just rely on user.services
<kentonv>
cool
<dwrensha>
we would want a visual design for displaying multiple verified emails in the profile editor
<kentonv>
it's ironic that the verified email problem, which is one reason we pushed on multiple identities, solved itself just as we're about to launch multiple identities.
<asheesh>
kentonv: literal lol
<asheesh>
borderline rofl
<jadewang>
actual lol
<dwrensha>
I think "ability to migrate away from identity providers you no longer fully trust" and "discoverability of identities that you know through particular identity providers" have always been at least of strong motivations for multi-identity.
<dwrensha>
I'm working on using user.services.github.emails right now...
<zarvox>
so then the next question becomes: should you be able to tag someone via their github-verified email, if they haven't made it an email identity?
<zarvox>
I think the answer is "no", because it leaks your "private" GitHub emails to your contacts
<dwrensha>
"tag someone"?
<zarvox>
@-mention
<zarvox>
the contact-autocomplete thing
<zarvox>
I guess verified email addresses have always been private, except for email address identities, so no change there.
jadewang has quit [Remote host closed the connection]
simonv3 has quit [Quit: Connection closed for inactivity]