XSF Discussion - 2017-03-24

  1. Guus


  2. jonasw

    we should make a nice graph of at which time of day people make their renewal commits :)

  3. Ge0rG

    jonasw: yeah, it might be leaking private information.

  4. Guus

    jonasw: did you account for people setting the renewal date in the future? :)

  5. jonasw

    Guus, not in the linting, but good idea

  6. edhelas

    hi guys, once again we are working on a new XEP that will replace Bookmarks in the futur, https://lite5.framapad.org/p/OlHwr5GBlY

  7. jonasw

    Ge0rG, *shrug*, so does making a commit in your library

  8. jonasw

    Guus: https://github.com/xsf/xmpp.org/pull/300

  9. edhelas

    also https://github.com/xsf/xeps/pull/458 :)

  10. Guus

    edhelas: I weild no power in that repo. You'll need SamWhited or friends for that, I think.

  11. edhelas

    Guus, no problem ;) it was just to drop the link here

  12. nyco

    Guus, yes?

  13. Guus

    nyco: l wanted to discuss with you if I should do something that I now already have done :)

  14. Guus

    (see github)

  15. Guus

    we should get your blogpost published!

  16. jonasw

    it would be great if the website showed the commit used to build it somewhere, even if only somewhere hidden in the source.

  17. Guus

    jonasw: make it so.

  18. jonasw

    that’d save us the "has it rebuilt yet or did something obscure fail?" guesswork

  19. nyco

    Guus, I agree, but we've just published one... this afternoon maybe ?

  20. nyco

    Guus, category Events, fixed

  21. Guus

    nyco - there will be a unknown delay between the merge being done, and the change being visible on the website.

  22. nyco

    like minutes?

  23. Guus

    unsure. I think there is a process that checks a couple of times per day if the website needs to be updated.

  24. Guus

    so, possibly, hours.

  25. jonasw

    I think so, too

  26. Guus

    jonasw, I think you're the resident expert: is there value in upgrading pelican from 3.3 to whatever is the latest version?

  27. jonasw

    I don’t know, actually

  28. jonasw

    I have started using pelican only a few weeks ago, either ;-)

  29. jonasw

    (but then spent a few intense days to port two websites to it)

  30. nyco

    ok, then let's push when we think it's good, and it's updated when it's updated... ;-)

  31. jonasw

    Guus: in case of doubt, don’t change a running system

  32. Guus

    which makes you the expert. :)

  33. jonasw

    I don’t even know which version is recent…

  34. Guus

    jonasw: agreed. I was actually running locally with 3.7 (which was wat I got with a default install), which gave me errors.

  35. jonasw

    3.7 according to my debian…

  36. jonasw


  37. jonasw

    it builds the website fine here

  38. jonasw


  39. Guus

    nyco: it looks good to me now. Do you want another screenshot?

  40. nyco

    yes please, if it takes less than 30 seconds

  41. Guus

    jonasw: for me, the menu was kaput. Also, 3.7 was more forgiving when I neglected to specify the number of items in a menu correctly.

  42. jonasw

    Guus: you need to run make devserver or something though instead of simply running pelican for reasons unknown to me

  43. jonasw

    oh right

  44. jonasw


  45. jonasw

    that explains why I had a hard time finding the software lists when testing :D

  46. jonasw

    well, that’s one good reason to wait with upgrading

  47. jonasw

    or at least plan it properly

  48. Guus

    nyco: 32. I counted. Nost of that was trying to resize the browser window.

  49. nyco

    toooooo long!!!

  50. Guus

    jonasw: agreed - but it does suggest that there are important changes, which made me wonder.

  51. jonasw

    Guus: ctrl+shift+m on firefox (re resizing)

  52. Guus

    jonasw: https://github.com/xsf/xmpp.org/issues/247

  53. nyco

    yes, cool, I see no glitch... thx Guus! ;-)

  54. nyco

    also, on our blog, we should insert visuals

  55. nyco

    can I do that, or it's too late?

  56. Guus

    go for it

  57. nyco

    I upload a picture to the repo? or use an online pic?

  58. Guus

    lol, I added the screenshot to the wrong PR :P

  59. jonasw

    nyco: put it in content/images/, maybe in a new blog subdirectroy

  60. Guus

    nyco: I have no clue.

  61. jonasw

    then the only question is how to reference it properly, this varies with pelican versions unfortunately

  62. nyco

    ok, we'll try on another blog post

  63. jonasw

    what *should* work is this: |filename|/images/blog/thenameofyourimage.foo

  64. nyco

    sorry no time for today

  65. nyco


  66. jonasw

    (filename is literally filename, see here: http://docs.getpelican.com/en/stable/content.html#deprecated-internal-link-syntax )

  67. Guus

    nyco, are or are you not going to change anything?

  68. nyco

    Guus, not this time

  69. Guus

    Then I"ll merge now.

  70. Guus


  71. Guus

    any concerns about https://github.com/xsf/xmpp.org/pull/294 ?

  72. jonasw


  73. Guus

    jonasw: what concerns?

  74. jonasw

    making a review right now, at least I found a typo

  75. Guus

    nyco: can we remove https://wiki.xmpp.org/web/Meetups and use the website instead?

  76. Guus

    duplicate accounting will only result in problems

  77. jonasw

    Guus: review done

  78. nyco

    good catch!

  79. nyco

    I'm on it

  80. Guus

    typo fixed

  81. Guus

    link added

  82. Guus

    I'm tempted to make futher improvements in future PRs, if needed

  83. jonasw

    no objections

  84. Guus

    could you do one last sanity check?

  85. jonasw

    just thought the typo needs to be fixed before, the others are merely MAY suggestions

  86. Guus

    absolutely, tx

  87. jonasw

    Guus: works for me

  88. jonasw


  89. Guus


  90. Guus

    I'm gonna merge my own updates too

  91. Guus


  92. jonasw

    do it

  93. Guus

    done it

  94. jonasw

    looks better that way than with three open renewal PRs pending

  95. jonasw

    that may evoke bad feeling for some who waited months to have their software listed ;-)

  96. jonasw


  97. jonasw

    (or servers)

  98. jonasw

    can we close website PR # 193?

  99. jonasw


  100. jonasw

    the changes it carries are obsolte anyways

  101. Guus

    great minds

  102. Guus

    I just closed it

  103. jonasw


  104. jonasw

    goffi makes a good point here: <https://mail.jabber.org/pipermail/jdev/2017-March/090299.html>

  105. jonasw

    do we want that? If so, on a separate page or below the servers listing?

  106. jonasw

    Finding good components is not trivial to be sure

  107. Ge0rG

    my gut feeling is to merge it with the servers lisst

  108. jonasw

    and also it highlights a core feature of XMPP, being able to interface with other protocols by embedding them

  109. Ge0rG

    my gut feeling is to merge it with the servers list

  110. jonasw

    I wouldn’t merge it, because they’re different things, but putting them on the same page might work

  111. Ge0rG

    good to me too

  112. Ge0rG

    same page, different table

  113. jonasw


  114. Ge0rG

    also we really should have that description field.

  115. jonasw

    I just now realised you’re not Guus. I got tricked despite colours. (I blame it on the fact that the pidgin colours changed radically when I changed the background color by only a few bits from 99% white to 98% white or something)

  116. jonasw

    yes, for components it’s definitely needed, otherwise you don’t know what it is.

  117. jonasw


  118. Guus

    Stop impersonating me Ge0rG!

  119. Guus

    There's something to adding server components to a public list. I don't have a preference how that's done.

  120. Ge0rG

    Guus: I'd change my nickname, but that's impossible because of MSN restrictions.

  121. intosi

    Ge0rG: switch to AIM, and you're done.

  122. jonasw

    *switch to AIM to reach your goals

  123. Ge0rG

    but my aim isn't very good

  124. Guus

    I think this concludes my round of fixes for today

  125. Guus

    I'd be greatful if someone *eyes intosi* could help sort out some of the issues that need access to the file system of the webserver

  126. Guus

    (where did the XEP archives go? Can we add various redirects? Delete the old ical files + related infra, etc)

  127. intosi

    Guus: tickets?

  128. intosi

    Or a ticket, at least ;)

  129. Guus


  130. Guus


  131. Guus

    possibly https://github.com/xsf/xmpp.org/issues/242

  132. Guus


  133. Guus

    https://github.com/xsf/xmpp.org/issues/224 (which is somewhat duplicated by another one)

  134. Guus

    also, I don't know what this means: https://github.com/xsf/xmpp.org/issues/152

  135. Guus

    ... and that's just from a quick glance over the issue list... :)

  136. jonasw

    of 152, I think most don’t apply to a static website without confidential content or interaction.

  137. intosi


  138. intosi


  139. jonasw

    you asked for it! ;-)

  140. Guus

    I'm trying to get that fixed intosi, but there's only so much I can do myself

  141. Guus

    (which i mostly did, afaik)

  142. Guus

    what do you guys think about https://github.com/xsf/xmpp.org/issues/212 ? We should either commit to it, or close the issue.

  143. Ge0rG

    IMO, closing 152 as WONTFIX would be OK

  144. intosi

    I fail to see what 279 tries to cover that 224 doesn't.

  145. Guus

    please comment on the ticket, which makes it less awkward for me to close it.

  146. Ge0rG

    Guus: +1 to WONTFIXing 212

  147. Guus

    intosi: yeah, probably duplicates.

  148. Guus

    also, https://github.com/xsf/xmpp.org/issues/275 - no one but me really likes that, right?

  149. Guus

    in which case it should be closed

  150. Ge0rG

    meh. you can't click on blog post titles on https://xmpp.org/blog.html to read the full version.

  151. intosi

    275 is not something I will fix for you ;) It's content anyway, so you can fix it, or not fix it, without help ;)

  152. Guus

    yes - apart from the annoying bit that I don't know any of those people :D

  153. Ge0rG

    I think the issue with 275 is to find the right text for the individual persons listed

  154. jonasw

    Ge0rG: right, that annoyed me too

  155. jonasw

    it’s easy-ish to fix though

  156. Guus

    jonasw: please do

  157. Ge0rG

    jonasw: +1

  158. jonasw

    you two are guilty for the dirty kitchen!

  159. jonasw

    I take it back. the CSS is SASS

  160. jonasw

    I cannot into sass

  161. Guus

    jonasw: please note that I am here just slightly longer than you :)

  162. Guus

    but I like cooking...

  163. Ge0rG

    jonasw: how is that a CSS issue? Isn't it rather about HTML <a>?

  164. jonasw

    yes, but it changes the colour of the header then

  165. Ge0rG

    jonasw> of 152, I think most don’t apply to a static website without confidential content or interaction. can you add that to the ticket?

  166. jonasw


  167. Guus

    intosi, out of curiosity: what is the update frequency of the website (how often does it check for changes in github), and why?

  168. jonasw

    Guus, Ge0rG: https://github.com/xsf/xmpp.org/pull/301

  169. Guus


  170. jonasw

    on my way

  171. jonasw


  172. jonasw

    github doesn’t like my screenshots

  173. jonasw

    "That file is hidden"

  174. jonasw

    "Something went really wrong"

  175. Guus


  176. Guus

    I just use 'select file' in the comment box

  177. jonasw


  178. jonasw

    that’s what I do

  179. Guus

    file permission problem?

  180. jonasw

    now it works

  181. jonasw

    screeshot attached

  182. Guus

    perhaps put quotes around the title in the link?

  183. jonasw

    I also thought about prefixing it with "about"

  184. Guus

    Continue reading 'New XMPP Software Listing Rules'

  185. jonasw

    it’s reaaallly hard to edit those files while the devserver thing writes over your vim terminal btw :D

  186. Guus

    I have this handy Vagrant image...

  187. Guus

    (or you just could use two terminals)

  188. Guus

    yeah, that's better

  189. jonasw

    new commit+screenshot attached

  190. jonasw


  191. jonasw

    ping Ge0rG

  192. Guus

    I'm not bothered by the missing green

  193. Guus

    if anything, it makes it more coherent with the rest of the site

  194. jonasw

    I liked the touch of colour, and I’ll probably bring it back once I figure out how to use sass

  195. jonasw

    it makes it less coherent on how the blog article looks after you followed that link

  196. Guus

    ah, right.

  197. jonasw

    last time I tried to use sass I was baffled that it takes 30 seconds to process a few files

  198. jonasw


  199. Guus

    yeah, I frequenly work with people that gave Maven flak for downloading a lot of stuff and taking a lot of time. I now get to look at them funny every time they make me do npm install.

  200. jonasw


  201. Guus

    The PR looks good to me. I'll leave it sit for a bit for others to be able to comment

  202. jonasw

    looking forward to do apt remove npm grunt; apt autoremove after this

  203. jonasw

    grave bugs of npm (→ 1.4.21+ds-2) <Outstanding> b1 - #857986 - npm: package is 3 years old (consider removal?) Summary: npm(1 bug)

  204. jonasw

    I cannot into sass, now it’s not finding some included files which also aren’t in the repository.

  205. Guus

    intosi, anything I can help with?

  206. intosi

    Just triaged right now. Where I made remarks, consider them and you (or others) provide an answer, and I'll make it so when it's more convenient, and doesn't interfere with my actual job ;)

  207. jonasw

    ah, right, intosi, can you give some details on how the website is built on the server? I’d like to know (a) if it is rebuild completely (i.e. with a fresh output directory) every time and (b) if so, does it do that unconditionally any N hours or does it check if a "change" happened and (c) if the latter, how does it detect a "change"?

  208. intosi

    Peace, out.

  209. jonasw

    ah, darn, too slow

  210. Guus

    thanks for your help! I'm off to do some work that actually pays bills too :)

  211. intosi


  212. Guus

    I've provided feedback where needed, I think. As far as I'm concerned: just do it. (c)(r)(tm).

  213. Guus

    things here are left in limbo way too long for my comfort

  214. Guus

    an improvement that is later improved upon further is better than having no improvement at all.

  215. Guus

    bear, can this be closed: https://github.com/xsf/xmpp.org/issues/109 ?

  216. Guus

    hmm: I can't add an "item for discussion" on the Trello board for the Board. Could someone add https://github.com/xsf/xmpp.org/issues/175 there? It's a relative simple clean-up.

  217. Guus

    dwd / ralphm / samWhited ?

  218. Ge0rG

    jonasw: +1 to the blog changes, even with blue instead of green

  219. nyco

    OH LOOK... on Hacker News: the URL may contain news.ycombinator.com and then /newest please upvote the meetup entry quickly (before it goes away) if you want/can and then comment, still if you want

  220. intosi

    nyco: an actual url to the HN article in question might be a few orders of magnitude more efficient...

  221. intosi

    https://news.ycombinator.com/item?id=13948527 for instance, that would've helped.

  222. Zash

    Wasn't the theory that they count votes less if you go there directly?

  223. nyco


  224. nyco

    it has the actual opposite effect!

  225. Ge0rG

    Less is more. War is peace.

  226. nyco

    efficient for what? so that you click?

  227. Ge0rG

    we are already identified by HN as being part of the "xmpp trolls" cluster.

  228. Zash

    Ignorance is bliss

  229. Guus

    I am shocked.

  230. jonasw

    you broke the pattern

  231. nyco

    clicking on the URL in a short amount of time by a few people actually decreases the scoring thus it will be less visible, and visible for a shorter period of time

  232. Ge0rG

    nyco: by any people or just by logged-in HN members?

  233. Guus

    so, here's a tought: perhaps we shouldn't try to affect the ranking at all, and let the actual algorithm do its thing? :)

  234. Ge0rG

    Guus: no, we need a blackhat SEO to make XMPP more prominent.

  235. nyco

    Ge0rG, both

  236. nyco

    Guus, exactly, so say to people how they can reach it, do not push the actual URL

  237. jonasw

    nyco: so by placing a ycombinator-link to, for example, slashdot, I can kill it?

  238. jonasw

    that sounds like fun

  239. nyco


  240. jonasw

    how broken is that?

  241. nyco

    Reddit has that as well, and so forth with Product Hunt, etc.

  242. jonasw

    and that are probably the same people who complain about facebooks filtering

  243. nyco

    no, no, nooo, it is NOT broken, it is actually FIXING the click shemas!

  244. jonasw

    no, I don’t get why clicking a link without interacting with the voting would have any effect

  245. jonasw

    either positive or negative

  246. jonasw

    that’s just broken

  247. nyco

    an enterprise asks all the employees to click "+" in a URL... that's a common practice, that must be fought

  248. jonasw

    that’s a different story, I agree

  249. jonasw

    but if I don’t vote, I think my click on the link should not have any influence

  250. Guus

    awesome! and you just told us to go click somewhere :D

  251. jonasw

    +1 Guus

  252. nyco

    Guus, not "told", asked politely... and I have not given the direct URL, that's the point

  253. Ge0rG

    "go click yourself!"

  254. nyco

    jonasw, -1

  255. jonasw

    nyco: only to trick the system

  256. Guus

    potatoes potatoes :D

  257. jonasw

    you would’ve given the URL if it wouldn’t affect our score, right?

  258. jonasw

    Ge0rG: click the button?

  259. nyco

    jonasw, no, that's a "if you want/can"

  260. nyco

    jonasw, that was not a "do that"

  261. jonasw

    I don’t feel that makes a huge difference

  262. jonasw

    but then again, I’m allergic to viral marketing

  263. nyco

    that is a little polite push, not a massive fraud

  264. Guus

    nyco: we're not discussing the manner in which the question is asked, but the fact that it's an artificial attempt to influence a score.

  265. nyco

    jonasw, it is not about what you feel, it is about the algo behind

  266. jonasw

    "Die Grenzen sind fließend.", as we say in germany

  267. nyco

    jonasw, which means?

  268. jonasw

    "the line is blurry", probably

  269. Guus

    fluent borders

  270. Guus


  271. Guus

    ah, that makes more sense :)

  272. nyco

    wow guys... we cannot win...

  273. Guus

    I don't consider myself to be in a competition

  274. Guus

    <-- most aweful marketing person *ever*

  275. nyco

    sure, but the other guys will win

  276. jonasw

    oh, I am.

  277. nyco

    Guus, that's the point

  278. nyco

    we massively lack marketing at the XSF

  279. jonasw

    nyco: I agree that some marketing is important. I disagree with manipulating scores.

  280. nyco

    so don't slow it down when we attempt to fix a little the awfulll situation that we are in

  281. jonasw

    I’d like to know what kind of marketing daniel did to make Conversations fly actually.

  282. nyco

    a little openness guys...

  283. jonasw

    oh, waqas, you’re around. do you have a minute to take a look at XEP-0390? https://xmpp.org/extensions/xep-0390.html

  284. nyco

    Conversations does not fly, it has 10 to 50k downloads on Play Store

  285. Guus

    yeah, again - i'm not in an aweful situation. I get to do cool stuff - I could not care less if it's used by 10 or 10,000 people.

  286. nyco

    also he gets out of the building quite often: Twitter, blog posts, events/slides

  287. jonasw

    nyco: that’s pretty flying

  288. Guus

    but hey, I won't stop anyone from promoting XMPP either :)

  289. nyco

    jonasw, WhatsApp and FB Messenger are above a billion, a billion

  290. jonasw

    I know

  291. jonasw

    but the network effect isn’t linear either

  292. Guus

    I think that Daniel primarily build a pretty good client.

  293. nyco

    at 50k you don't get any network effect

  294. Guus

    which made him stand out.

  295. jonasw

    promotion won’t fix anything unless we can provide people with software which doesn’t suck though

  296. nyco

    5 years ago you could reach critical mass at 1 million, now it is rather several millions

  297. nyco

    jonasw, That. Let's do that now.

  298. jonasw

    I’m on it!

  299. Ge0rG

    I have a paid amateur radio app on Play Store that's also in the 10k-50k bin. It's the #1 app for its use case though, and people know it.

  300. jonasw

    I just ditched a class to be able to do some polishing on my client before I hand it in to Prototype Fund. which will also get publicity if it’s accepted.

  301. jonasw

    Ge0rG: you also have a dice app I found out. I actually have that installed :D

  302. Ge0rG

    jonasw: did you have it installed before you knew me?

  303. jonasw


  304. jonasw


  305. Ge0rG

    yay! I'm famous!

  306. Ge0rG

    jonasw: I only wrote that app because there was a bunch of dice apps before, and they all sucked.

  307. jonasw

    that’s the best reason for writing an app

  308. Guus

    perhaps we should do that for XMPP clients.

  309. jonasw

    I’m on it.

  310. Ge0rG

    I usually need a dice app because I forgot my dice, and not because I want to shake my phone like crazy.

  311. jonasw

    when I’m not busy with website PRs :-)

  312. Guus

    (obligatory xkcd reference in 3, 2, 1 ...)

  313. jonasw

    Ge0rG: seriously?

  314. Ge0rG

    Guus: this is also why I got involved with yaxim... back in 2009

  315. Ge0rG

    jonasw: yeah.

  316. jonasw

    oh-kay, so do those people also write SMS by moving their phone like a pen?

  317. Guus

    presumably, yes. If you can write them the app to do that.

  318. Ge0rG

    I'm sure there is a market for that.

  319. jonasw


  320. Ge0rG

    Afer all, ICQ got famous for using a typewriter click sound on each key press

  321. jonasw

    there are terminal apps which do the same

  322. waqas

    jonasw: I'm skimming through XEP-0390, it seems sensible at first glance

  323. jonasw

    thanks for taking a look!

  324. waqas

    jonasw: I'll later try to implement this in simple code, which would help think through any edge cases

  325. jonasw

    that’d be awesome

  326. Guus

    "if I didn't find anything I don't think that would be enough to count as evidence that problems don't exist." <-- I think Kev was just dethroned as the king of double negatives :)

  327. jonasw

    I fail to parse that

  328. Zash

    I haven't had enough coffee yet to even try

  329. intosi

    Don't see anything wrong with that sentence.

  330. jonasw

    my parsing suggests to add a ", but" behind "anything" to make it readable

  331. Holger

    I don't think I would say that I don't see anything wrong with that sentence, but that doesn't mean any problems exist in it.

  332. Zash

    "I have only proved the sentence correct, I have not tested it"

  333. Holger

    Does it compile?

  334. Zash


  335. nyco

    hey, do we have stats/traffic numbers on the xmpp.org website?

  336. Guus

    I'm not seeing an obvious tracker in the source of the pages (like google analytics). Unsure if we generate statistics from the access logs.

  337. Zash

    Nothing says "Privacy-focused" like Google Analytics

  338. Guus

    Nothing says "Privacy-focused".

  339. Guus

    sorry. You did. :)

  340. Zash

    and https://xmpp.org/

  341. Guus

    I sit corrected.

  342. jonasw

    Zash: what about embedding ajax.googleapis.com? :-)

  343. jonasw

    Zash: is https://prosody.im/chat/ that chat.tar.gz thing?

  344. Zash


  345. jonasw

    Zash: any comments on https://github.com/xsf/xmpp.org/issues/101 ?

  346. Zash

    I'm still sad that Candy isn't as maintained as it should be

  347. SamWhited

    I tried the Candy demo again recently, it just said "Connecting" forever so I gave up

  348. SamWhited

    (where "recently" == "just now")

  349. Guus

    We run Candy at www.igniterealtime.org/support/group_chat.jsp

  350. Zash

    SamWhited: Seen mod_candy?

  351. Guus

    It has quircks, but it's usable.

  352. SamWhited

    nicknames seem to be fixed anyways

  353. Zash

    Has the issue been fixed where if I join as both "Zash" and "zash", one gets ignored?

  354. SamWhited

    "username already in use"

  355. Ge0rG

    there is an SPA version of Converse.js at https://github.com/jcbrand/inverse.js which is even-more-beta

  356. Guus

    unsure - there are many oddities. One of my biggest gripes with Candy is that it dies (without letting the user know) on any uncaught error.

  357. Guus

    which gives people the impression that no-one is responding.

  358. waqas

    My main gripe with Candy were the XHTML-IM security issues…

  359. Guus

    waqas: I disabled XHTML-IM in our instance - it introduces instability

  360. waqas


  361. Guus

    so those might still be in there

  362. SamWhited

    they do appear to be escaping things properly and not sticking stuff in attributes now at least; do they support XHTML-IM? Regardless of how good things are otherwise, as far as I'm concerned that is a security risk in and of itself.

  363. Guus

    they support it, but you can disable the support

  364. Guus

    and/or not enable it - I'm not sure anymore, it was a long time ago

  365. Ge0rG

    What's the use case we are talking about? single-page anonymous access to a MUC?

  366. SamWhited


  367. Ge0rG

    then maybe prosody's chat is the way to go, with some fancy CSS around it

  368. Ge0rG

    jonasw: buggy has a question mark behind it, mainly because it's just a tarball on a web page that hasn't been touched in years

  369. waqas

    Mainly because nothing has really changed in terms of requirements

  370. SamWhited

    Candy seems okay now (at least, the way I think I got it last time has been fixed), but I have gotten it to execute arbitrary JS in the past. Prosody's web chat is literally the only one that I don't think I've been able to inject or otherwise break too badly.

  371. SamWhited

    only one out of the ones I've tried, obviously.

  372. waqas

    SamWhited: Same. And I'd tried many, including ones without xhtml-im and desktop clients...

  373. SamWhited


  374. Zash

    I'm not sure the tarball reflects what is actually used either

  375. SamWhited

    It's fairly easy to style too; eg. https://muc.mellium.im/ for a project that doesn't actually exist that I will launch one day.

  376. Guus

    for the record: i think you guys are overly cautious. I'd rather go with a well maintained project than a dormant one.

  377. Guus

    but, if we can pick any one client, i'm happy to go with that.

  378. SamWhited

    I'd rather go with simple and low footprint for attack over anything else; but maintained would be nice too.

  379. Ge0rG

    if the prosody devs would make it a real project, with an issue tracker etc, it would instantly lift off

  380. Guus

    I think that any client has a decent amount of complexity, and vectors of attack - perhaps not that one that we test for now, but eventually, something will be found.

  381. Ge0rG

    as it is, one could get the impression it is an unmaintained student project that was dropped on a web page a decade ago

  382. SamWhited

    Ge0rG: Who would maintain it in that case?

  383. waqas

    I'm curious how much demand there is

  384. Zash

    Ge0rG: Someone did that, then added tons of broken features and then forgot all about it.

  385. waqas

    Yeah, I think the main concern would be scope creep. It was designed to be one simple thing: provide users with chatrooms anonymously. Stuff like admin-tools, etc were out of scope.

  386. Ge0rG

    I'm okay with the feature scope, but I need to think of it like an admin: if I install it onto my machine, I need some way to update it, monitor for CVEs etc.

  387. Ge0rG

    As it is now, it doesn't even have a name

  388. Zash


  389. Zash

    Have you noticed how our issue tracker is called issue tracker?

  390. Ge0rG

    Zash: "chat.tar.gz" rather?

  391. SamWhited

    "chat tarball" — good name, I like it!

  392. Ge0rG

    jonasw> Zash: is https://prosody.im/chat/ that chat.tar.gz thing?

  393. Ge0rG

    let's call it "that chat.tar.gz thing"

  394. Zash


  395. jonasw


  396. jonasw


  397. Guus

    Chatty McChatface, please.

  398. SamWhited

    The Client Formerly Known as chat.tar.gz

  399. waqas

    chat tarball… chatball.. magic chatball! It's perfect!

  400. Ge0rG

    Guus: +1

  401. Guus

    so... after a year and a half, let's pick a webchat client.

  402. Guus

    who can create a PR that introduced Chatty McChatface?

  403. Guus


  404. jonasw

    *That Chat.tar.gz Thing

  405. Zash

    Watch out, it's that chat.tar.gz thing!

  406. Guus

    ... anyone?

  407. MattJ

    Ge0rG, in your post about Chatball (which I hadn't seen before) maybe mention mod_muc_ban_ip in the Prosody config part as a side note

  408. Ge0rG

    MattJ: I'm not sure what I should say about mod_muc_ban_ip in there

  409. MattJ

    Guus, I guess I could

  410. Guus

    Thanks mattJ!

  411. Guus

    for reference, we've been tracking the effort in https://github.com/xsf/xmpp.org/issues/101

  412. MattJ

    BOSH endpoint?

  413. Zash


  414. MattJ

    I don't think there's currently one on xmpp.org, unless I'm mistaken

  415. MattJ

    I possibly still have SSH access there so I could set it up

  416. Zash

    MattJ: Can I convince you to get rid of mod_muc_log_http?

  417. MattJ

    and replace with?

  418. Zash

    mod_http_muc_log :)

  419. Guus

    MattJ, could you arrange that with the powers-that-be in the infra team (Kev, Intosi, Tobias, others?)

  420. MattJ

    Unless I was quietly kicked out, I think I'm still a (dormant) member of iteam :)

  421. MattJ

    Haven't worked on it for at least a year though, I reckon

  422. Guus

    Could you arrange that with yourself then? :)

  423. MattJ

    I'll do whatever is necessary to make the web chat work, and I won't stop till I'm done!

  424. MattJ

    But first, meetings...

  425. intosi

    Close shave.

  426. SamWhited

    that reminds me, I need to go to the barber.

  427. jonasw

    any opinions on that one? https://github.com/xsf/xmpp.org/pull/302 I didn’t even know that page existed…

  428. Ge0rG

    am I the only one who can't click on the menu items on xmpp.org?

  429. jonasw

    worksforme, but I have 1st party JS enabled

  430. SamWhited

    you're not the only one; anytime menus are made with javascript bad things happen.

  431. jonasw

    it also works without though

  432. Flow

    jonasw: hmm 100k users

  433. Zash

    Doesn't jabber.org have on the order of 1M users?

  434. Ge0rG

    I do have JS enabled, the menups properly pop up and show the link, but clicking only makes the menu collapse, on Chromium 57

  435. Guus

    jonasw: I was somewhat unsure of that one too. I can't find any references to it on wikipedia (english nor italian)

  436. jonasw

    Ge0rG: all items or only specific ones?

  437. Guus

    not sure if that proves anything.

  438. Ge0rG

    jonasw: all items. I can click on the menu itself, and then click through from that category page

  439. SamWhited

    Maybe the number of users column should go away since it's a thing that will change and probably hard to verify anyways?

  440. Zash

    Maybe the page should die in a fire before it becomes used for marketing

  441. Guus

    SamWhited, sure.

  442. Guus

    Zash, sure.

  443. Flow

    what Zash said

  444. Zash

    Like how XEP support pages usually end up

  445. Guus

    Guys - make it so.

  446. jonasw

    remember to check for links to that page before burninating it

  447. jonasw

    I’m not available for any website patches anymore today, EBUSY

  448. Guus

    (perhaps dropping the table instead of the entire page would be a good compromise)

  449. Zash

    chatme is on the xmpp.net list already, so duplicating it on that page seems weird to me

  450. Ge0rG

    Maybe that page should only list XMPP-based services that are not XMPP servers?

  451. Ge0rG

    but I'm fine with killing it as well

  452. jonasw

    it advertises WhatsApp!. kill it with fiiiiireeeeee. (well, okay, it may make people think "uh, whatsapp uses XMPP, maybe it isn’t *that* bad)

  453. Zash

    A clearly defined purpose / scope would be good

  454. Ge0rG

    jonasw: "it isn’t *that* bad" - WhatsApp or XMPP? :P

  455. jonasw

    Ge0rG: your choice

  456. Zash

    But HN says that WA is no longer XMPP

  457. Ge0rG

    jonasw: I take "neither".

  458. dwd

    Zash, It hasn't been in ages, though. Question is whether it's still fundamentally based on XMPP or not.

  459. Zash

    As in, "they switched to Signal Protocol", which doesn't even make sense

  460. Zash

    I haven't seen any credible proof that it's not still "FunXMPP"

  461. jonasw

    I don’t care either.

  462. fippo

    https://twitter.com/backlon/status/845327874928238593 -- did i miss the obituary notice?

  463. Zash

    fippo: No, you've forgotten about it because it was soooo long ago

  464. Zash

    https://q.zash.se/86ed5344ab0a.txt .. seems Facebooks XMPP isn't entirely killed yet either

  465. fippo

    https://twitter.com/markb/status/845328381117956097 ... hrm... brb, I need to ask someone about this...

  466. lovetox

    i need help, i tried to read RFC6121

  467. lovetox

    about IQ stanzas

  468. lovetox

    when using in band bytestream

  469. lovetox

    can it make sense to adress the stanzas to a barejid?

  470. lovetox

    where does the server route such an iq stanza

  471. lovetox

    im terrible with the whole filetransfer thing

  472. fippo

    that doesn't make sense. you need to figure out the full jid of someone before sending them a file

  473. lovetox

    thats what i thought

  474. lovetox

    the xep also has only examples with full jid, but it is never mentioned explicitly that stanzas have to be adressed to full jids

  475. lovetox

    maybe because its obvious

  476. Zash


  477. Zash

    > For an IQ stanza, the server itself MUST reply on behalf of the user with either an IQ result or an IQ error, and MUST NOT deliver the IQ stanza to any of the user's available resources.

  478. lovetox

    so in an ibb there can never be a iq involved that is not adressed to a fulljid

  479. Zash

    In theory you could send the data in messages, but don't do that.

  480. lovetox

    no gajim does not do that

  481. Zash


  482. Zash


  483. lovetox

    thanks zash, but we thats not an issue, gajim does not use messages for ibb

  484. Zash

    I don't understand

  485. Zash

    I'm saying that the XEP says you can if you really want.

  486. lovetox

    my question was about IQ stanzas and ibb, you posting of the rfc was the help i needed

  487. arc

    official announcement is out, gtalk will be officially killed off this summer

  488. moparisthebest

    arc: it's more akin to a merciful finishing off isn't it? It's better than staying running as broken as it is and giving xmpp a bad name

  489. MattJ

    Most people think it's already turned off