#fdroid-dev

/

      • _hc has quit
      • _hc joined the channel
      • _hc has quit
      • _hc joined the channel
      • zwliew joined the channel
      • mvp76 joined the channel
      • _hc1 joined the channel
      • _hc1 is now known as _hc
      • _hc
        hi cdesai uniq[m] pserwylo krt[m] seems that Tovok7 can't get onto freenode
      • uniq[m]
        hey
      • cdesai
        hello
      • _hc
        thezero
      • thezero
        hello
      • Tovok7 joined the channel
      • Tovok7
        Now, I got a connection, strange thing.
      • _hc
        well hello Tovok7, they let types like you in now ;-)
      • Tovok7
        hahah :)
      • _hc
        from the "wrong" part of the internet
      • mimi89999_0: nailyk est31 larjona CiaranG: ping, in case you're interested in following or joining the meeting
      • ok, I'm a little slow today, does anyone want to start?
      • est31
        hi
      • Tovok7
        I can go
      • There is not so much to report this week for me, because I was traveling to Cryptorave.
      • Just did some minor changes like making repomaker work with MariaDb/MySql, improving the form styling.
      • A bigger chunk of work was sanitizing the HTML of (localized) app descriptions and let the user of repomaker edit it with TinyMCE.
      • At the moment, I am working on rm#29.
      • NOTICE: [repomaker] #29: Support Feature Graphics - https://gitlab.com/fdroid/repomaker/issues/29
      • In other news, we are close to getting the first UX designs ready which was slower than expected and start to implement those.
      • EOF
      • _hc
        Tovok7: tell us more about Cryptorave :)
      • Tovok7
        It was a nice small non-commercial crypto/internet security event organized by volunteers. It was my second year and I ususally like it, but you can't compare it with European conferences of this type like CCC.
      • Things like that are just blossoming in Brazil.
      • _hc
        CCC seems way too big
      • how big was cryptorave?
      • Tovok7: what kind of sanitizing did you do of the HTML?
      • Tovok7
        hard to tell, because there were no tickets. They say they had 2500 people registering, but it was way less people at any given moment in time. Maybe a coup[le of hundred simultaneoulsy.
      • _hc
        good size then
      • lots of people, but not too anonymous
      • Tovok7
        I do whitelist sanitzing with bleach: https://github.com/mozilla/bleach
      • So everything not in the whitelist gets removed from the descriptions
      • nailyk
        _hc: sorry just came. What is the purpose of the meeting please?
      • Tovok7
        Who wants to know more about cryptorave: https://cryptorave.org/
      • _hc
        nailyk: its the weekly fdroid dev meeting :)
      • I thought you'd been part of it before, that's why I asked
      • larjona lurks
      • hi
      • Tovok7: seems like we might need something like bleach if we support HTML in the descriptions
      • nailyk
        thanks. Will read silently. Am really interrested on it :)
      • thezero
        Tovok7, is there an english page? xD
      • Tovok7
        _hc: you mean in fdroidserver?
      • _hc
        but the limited HTML support of Android Html.fromHtml() might be enough
      • yeah
      • thezero: portuguese isn't close enough to Italian? I can usually fake my way through Dutch, knowing German and English
      • Tovok7
        thezero: unfortunately not, but I ran into surprisingly many non-brazilians there
      • there was a tails developer from france for example
      • _hc
        which one?
      • intrigeri?
      • or perhaps he didn't want it known...
      • they live on the internet :-D
      • about HTML, the question is: should we support HTML in the descriptions, or Markdown? neither? both? I propose we have a short realtime voice session on https://meet.jit.si/fdroid to figure this out. Who is interested?
      • nailyk
        I am. But not really contributing to frdoid.
      • _hc
        guess people are pretty quiet/slow/multi-tasking today
      • larjona
        I can't. And I have no opinion... I would say the same that google play does, if we can know...
      • cdesai
        I'm around but not really sure if I can add / say much about the description discussion.
      • larjona: I was just trying to figure out what play store supports
      • _hc
        I'm guessing mvdan krt[m] Tovok7 pserwylo and me would have the strongest opinions on the matter
      • cdesai: it seems to support basic HTML
      • thezero
        I think it's better to just support markdown, isn't it enough?
      • Tovok7
        I'd take the meeting if it doesn't happen right now: My take on it is to support a well defined subset of HTML i the repo index and then whatever tools fill the findex can use other languages as well if they like, as long as they map to that subset of HTML. So you could have markdown in the metadata files and fdroidserver transforms that to HTML when putting the description into the index.
      • _hc
        the index is also used by the website, and it'll do much better with Markdown than HTML
      • cdesai
        html might be easier to support on android, but markdown maybe easier to constraint given that it doesn't have as big of a surface as html.
      • est31
        markdon is afaik a superset of html
      • _hc
        I'm thinking we should deliver raw description text, and each consumer chooses how to render it.
      • Tovok7
        in what times are we living, now websites do better with markdown than HTML alrady ;)
      • _hc
        the raw form could be plain text, plain text with allowed HTML tags, or Markdown (which allows some HTML)
      • Tovok7
        how to consumers recogize the format then?
      • *do
      • _hc
        I mean we choose just one of those as the standard
      • then there is only one standard format
      • thezero
        est31, depends if you permit HTML along with Markdown syntax
      • _hc
        from the markdown manual: "For any markup that is not covered by Markdown’s syntax, you simply use HTML itself. There’s no need to preface it or delimit it to indicate that you’re switching from Markdown to HTML; you just use the tags."
      • the only tricky thing is: "Note that Markdown formatting syntax is not processed within block-level HTML tags. E.g., you can’t use Markdown-style emphasis inside an HTML block."
      • thezero
        exactly
      • uniq[m]
        I'd rather see markdown translated to html serverside, because it saves clients the trouble of including a markdown lib
      • _hc
        that's not very hard to do these days
      • uniq[m]
        as for building search indexes, stripping out html tags should be a fairly easy regex.
      • Coffee[m]
        If you want to avoid Ambiguity Hell, I'd stick with a simple html subset.
      • _hc
        and translating server-side, that means ending up with hard problems with formatting on the website vs the android app
      • uniq[m]
        i don't see how markdown could resolve any issues of this kind
      • Tovok7
        me neither, especially if it also includes arbitrary HTML
      • _hc
        jekyll adds styles to the markdown
      • in a different way than it happens on Android
      • if its converted to html server-side, the HTML is going to have a point of view in how its marked up for formating
      • but maybe the subset of HTML that we use would be small and simple enough that there never needs to be id="" or class=""
      • only <p>, <li>, <b>, etc
      • Tovok7
        We are just talking about < p > < a > < strong > and < i > here, or is there more?
      • _hc
        lists
      • Tovok7
        arghhhh my client is eating html tags
      • _hc
        lol
      • Tovok7
        yes, I'd say no id or class allowed
      • uniq[m]
        ^ - ^
      • Tovok7
        these ID and classes are unlikely to have the meaning on other systems where that HTML is processed
      • make the subset as basic as possible, maybe only allow href if you want to support links
      • _hc
        it looks like there is good/easy support for Markdown in jekyll, django, and Android, and the current model of fdroidserver's index is a pure conduit for data, not
      • nailyk
        am not sure of the purpose but I contribute in some website github sometimes. Thoses using jekyll had an advantage, it is really easy to add/change pages. However it is an heavy process to test before pushing until you get your own instance.
      • _hc
        as far as I can tell, Google Play also just delivers the raw description to the website and Play app
      • nailyk: if you push to your own fork's master branch, you get an automatic, complete render. Here's mine: https://eighthave.gitlab.io/fdroid-website/
      • how about we let this discussion stew in our minds a bit more? Here is an issue to track this: #1000
      • NOTICE: [client] #1000: support displaying basic HTML in app descriptions… - https://gitlab.com/fdroid/fdroidclient/issues/1000
      • thezero
        You can strip HTML from the description and only use Markdown if you needs only basic tags
      • _hc
        I'm assuming silence means next topic?
      • so I go with my status:
      • I've been reviewing the Bazaar2 proposal to make sure that we are delivering what we promised in the contract.
      • As we get closer to the end of the funding, I also have to do more accounting work :-D :-|
      • The localization stuff is getting pretty polished. I'd like to set up 5-10 apps that are translated via fdroiddata.
      • for most apps, we should help the developers get setup with fastlane or Triple-T:
      • But there are a handful of popular apps that are unlikely to use these methods, e.g. Firefox and derivatives, VLC, Nextcloud, etc.
      • For these, I propose that we maintain translations in fdroiddata.
      • Or maybe we maintain just the translated name/summary for like 50-100 apps
      • this can be nicely integrated with Weblate
      • I'll like feedback on this idea, especially from people who've worked a lot on fdroiddata, like mvdan, krt[m], est31, etc
      • Tovok7
        If you have any Bazaar2 proposal deliverables for repomaker, please send them over, to make sure we actually deliver them.
      • _hc
        some general ones are related:
      • * FDroidappstoreandwebsitefullylocalized
      • * available80%oftestgroupratestheappstoreexperienceasgoodandhelpful
      • * neworganizationsrunningtheirownapp/mediarepositories(Target:3)
      • * FDroiddownloads(Target:100,000)
      • * appswithallrelatedcontentfullytranslatedintoSpanish,Farsi,andTibetan(Target:10)
      • * 95%ofalltextcontentabletobetranslated
      • sorry for the weird formatting
      • EOF
      • whoever is ready to report, go right ahead
      • uniq[m]
        This week I've finished support for rewriting metadata to yaml.
      • Unfortunaly parsing yml metadata is not ready yes, but I'm working on it.
      • *yet
      • Any feedback about how yaml metadata should look like would be appreciated.
      • _hc
        what's left to do on parsing .yml?
      • uniq[m]
        build infos don't get parsed correctly
      • _hc
        I'm excited about switching to .yml since it'll mean the data key names will be synced across the whole project (only different being camelCase start letter, e.g. WebSite vs webSite)
      • plus YAML will be easier to edit, since lots of editors support it
      • syntax highlighting, etc.
      • uniq[m]: how about your trip?
      • uniq[m]
        I've also participated on a reproducible builds hackathon, kindly hosted by debian.
      • So I've tried get some apps building repoducibly and I managed to get 2 to do so. :)
      • (Thanks _hc for the pointing out which ones would be interesting)
      • cdesai
        maybe somewhat relevant question - if an app can be reproducilbly built, what signature would it use in fdroid?
      • _hc
        well, they are also ones that the funder wants to see
      • cdesai
        I was talking with the LineageOS guys about having F-Droid / putting up some of their apps on F-Droid, and one of the concerns they had was signatures.
      • _hc
        uniq[m] and i Just worked out the plan to handle that