#dev 2022-06-23

2022-06-23 UTC
jacky, geoffo_ and angelo joined the channel
#
jacky
angelo: ^
#
GWG
My question re yours is every like you sent me is parsed as /about and that's what I link to
#
GWG
Is that what you expect
#
jacky
what do you mean? like the URL of the source is my h-card?
#
jacky
or that of my author in my like?
#
GWG
The author property, and Webmentions implementation links to the author url on likes, not the like
#
angelo
i'm getting the h-card on your homepage jacky by looking for a repr h-card. the additional logic i'm thinking about implementing is: if the url/uid is something /other/ than the current page then go fetch that and merge the two
#
jacky
angelo: that'd work
#
angelo
although in your case both cards appear to be the same
#
angelo
and what i want is your rel-mes
#
jacky
versus
#
jacky
GWG: I'll check my implementation
#
GWG
jacky: Asking what your intent is
#
jacky
I don't think I follow tbh
#
jacky
do you have a URL from me to demostrate/show what you're referring to?
#
GWG
You sent me a Webmention on this
angelo_, AramZS, cjw6k, mro, gRegor, gRegorLove_, GWG, [Murray], tetov-irc, [tantek], [tonz], geoffo_ and jacky joined the channel
#
jacky
GWG: gotcha and are you saying that the 'source' of the Webmention was https://jacky.wtf/about
#
jacky
if so, that's something I can add a test for
#
jacky
it _might_ be because my approach pulls all of the links out of the associated entry and my h-card URL qualifies
geoffo_ and jacky joined the channel; jacky left the channel
#
GWG
jacky: No, the author property...not the source. I display the author property on likes and mentions..only show the source url on replies
#
jacky
okay so wait what was the issue? I'm confused even more
#
jacky
is it the sparseness of the field?
#
GWG
jacky: I was asking you if your intent was for /about to be what people displayed
#
jacky
Oh yeah, it is
#
jacky
like that link + the photo at that page via the provided h-card
#
jacky
I figured that's how it _should_ work (unless I'm misinterpreting now how to do rep h-cards)
jacky joined the channel
#
omz13
jacky I ran my profile discovery code (https://toolbox.imoxia.com/#profiledisco) against jacky.wtf and it determines your representative h-card is at your /about
#
omz13
but it picked up a few nits... the main one being how you specify your birthday... it has a timezone component?!
#
sknebel
jacky: line 26 on /about has a stray </head>
#
jacky
ah thanks
#
[tantek]
jacky, re: AP tests, btw people do ask about this still and it's :cricket: :cricket: https://twitter.com/nileshtrivedi/status/1539863013225816065
#
@nileshtrivedi
@sl007 @bluesky @t Interesting. BTW, is there a ActivityPub conformance test suite somewhere? Would love to see how Mastodon actually scores on that. 😄
(twitter.com/_/status/1539863013225816065)
#
omz13
and the bday "-07-10T12:00:00-05:00" should be "--07-10" (if you want to drop the year)
#
[tantek]
agreed re: yearless dates. the TZ offset should drop the ":" for better readability/disambiguating time vs tz
#
[tantek]
there's a mf2/vcp parsing output c14n proposal on that also, see #microformats
#
sknebel
not vcp in this case
#
jacky
right but that's maintained by the AP community, no?
#
jacky
omz13: ah, excellent ty!
#
[tantek]
jacky, yeah, I'm not sure if "maintained" is the right verb for something that's offline and not planned on being resurrected in that form
#
[tantek]
sknebel, hmm, interesting point. perhaps TZ c14n needs to be higher level than vcp in that case
#
[tantek]
interesting that the iOS edited message fallback will be very similar to how discord users's edited messages show up here in IRC/Slack: https://www.theverge.com/2022/6/23/23179820/ios-16-imessage-edit-messages-incompatible-previous-versions-of-ios
gRegor and [Scott_Jack] joined the channel
#
gRegor
[tantek], yeah, there's an mf2p issue to move those vcp rules to apply to all dt-* https://github.com/microformats/microformats2-parsing/issues/12
#
Loqi
[tantek] #12 should dt-* parsing do date and time parsing for all values?
#
[snarfed]
interesting approach to the database antipattern, looks like a more powerful version of the flat file storage approach some of us use: https://www.gnu.org/software/recutils/
#
[tantek]
has anyone built a Faraday cage (which presumably would mitigate some of these effects, at least for long term disconnected storage like backups) at home? preserving your "long indieweb" as it were: https://www.wired.com/story/sun-storm-end-civilization/
#
[tantek]
^ how much of existing servers, any magentic based storage, maybe even SSDs would such an event wipe out today?
gxt, angelo_ and jacky joined the channel
#
superkuh
Very few.
#
superkuh
The timescale of geomagnetic events from an interplanetary coronal mass ejection is very slow. It does not induce currents effectively in small conductors. Only very long ones. The danger is to the grid backbone transformers (which take months to replace in the best of times).
#
superkuh
The only damage to computers would be from the power voltage spikes and power outage. Data-centers are equipped for that.
#
jacky
> The problem with YAML is that it was designed as a “data serialization language” and thus to map the data constructs usually found in programming languages. That makes it too complex for the simple task of storing plain lists of items.
#
jacky
feels wrong?
#
jacky
like I do this all of the time with ansible
#
jacky
unless they mean a lack of typing or column integrity..
#
jacky
(from the thing [snarfed] posted earlier)
#
AramZ-S[m]
I do think regular backup to disconectable-storage is wise, but there are still some minor risks even with powered down storage from magnetic events. I'm not sure how much but still. I know there are a bunch of for-sale faraday cages some of which can be pretty good
#
AramZ-S[m]
There's an interesting review I found in my archives over here - https://www.mattblaze.org/blog/faraday
jacky joined the channel
#
AramZ-S[m]
And I've not tested these, but I liked reading through this article - https://www.pcmag.com/how-to/how-to-make-your-own-faraday-cage
#
AramZ-S[m]
USB Keys have enough storage that it would be easy to think about backing up a site, or even a whole server stack, to a USB Key or a RPi and dropping it into a phone-sized faraday packet and then dropping that into an offsite storage unit. This is actually exactly what I considered doing a while back. I buy a portable HDD, back everything on my systems and sites up to it, and then drop it in a storage unit about once a yea
#
[tantek]
what is backup
#
Loqi
A backup of an indieweb site should have everything necessary to redeploy it on a new webhost https://indieweb.org/backup
#
[tantek]
AramZ-S[m]++ those are great tips & practices! definitely worth adding to at least a Brainstorming section in ^
#
[tantek]
That PCMag article is great! Love this: "a Faraday tent under which you can park a car and probably your entire family."
#
[tantek]
I'm thinking things like lining the inside of a closet (and closet door) with a faraday cage mesh, so anything you toss inside is by default better protected, and quite well protected when the door is closed.
#
[tantek]
i.e. making such protection a default (built-in feature) for such an ordinary "space" rather than making the Faraday cage its own special thing
#
AramZ-S[m]
Yeah, but I feel like if there's really a big issue with problems large enough to need a faraday cage, you're going to need it to be off-site as well. I think you'd really need something inside the closet, not the whole thing, because for a really effective faraday cage it needs to be a sealed up as possible. Maybe a big box you can open and close easily.
#
[tantek]
I mean sure, defense in depth and such
#
[tantek]
maybe a box inside the closet for your USB keys you want protected from solar radiation (or EMP weaponry?)
[chrisaldrich] joined the channel
#
AramZ-S[m]
Yeah, that's a cool idea. I already have a fireproof bag for that purpose, so adding an additional layer of protection around it would be useful.
jacky, tetov-irc and gRegorLove_ joined the channel