00:10:15 -!- arseniiv has quit (Ping timeout: 260 seconds). 00:22:49 -!- normsaa has joined. 00:35:55 -!- src has quit (Ping timeout: 260 seconds). 01:00:18 Is Tom7 interested in Free Hero Mesh now that bizarro is implemented? Someone told me that it is, but I don't really know. (Maybe, should also be implemented changing execution order of objects (like, I think, Tom7 did), but that might be difficult to know how to do it properly) 01:03:03 what is bizarro 01:03:54 The bizarro world is another grid of objects which is normally invisible and intangible, but is still affected by most turn-based actions 01:07:35 dunno, you'll have to ask him. he has an email. 01:07:47 and a blog with comment forms. 01:07:59 OK. I looked and I found two different email addresses. 01:10:32 But, I will try. 02:04:35 in git, how do you list the refs of a remote repo? 02:06:07 b_jonas: does "git branch -v --remote" do what you want? 02:17:31 ais523: no, that only shows the local tags created that follow the remote tags. but (git ls-remote) does work. 02:18:58 ais523: the context is that I was trying to pull from a git repo that renamed its master branch to main, and didn't leave master even as a synonym, but I still have a local remotes/origin/main reference. so I wanted git to list remote refs to make sure that indeed, master doesn't exist on the remote. 02:19:11 *sigh* 02:23:12 the "master" nonsense started out sanely, with only the fancy web interface on github localizing "master" to "main", and using "main" as the branch name in repos if you create one with their fancy GUI tool. they don't need to touch actual branch names, because that would break compatibility, and the kind of people who don't understand that won't see the real branch names because they don't type (git 02:23:18 lsref) or anything in the command-line, they just look at the github frontpage or something. 02:23:36 but no, at least one repository actually has the existing branch name deleted now. how did we get to this? 02:24:54 immibis: ok, you were right 02:24:56 um 02:24:58 int-e: you were right 02:25:20 and it sounds like it's from one of those puzzle/contest sites, not from coursework homework 02:25:36 fwiw, I think it makes the most sense to name the primary branch after the project 02:26:37 I don't care what you name the branch, I only care when it's renamed in an existing public repo with no synonym left 02:30:01 (I also found another repo that did a non-fastforward change to their master 02:32:36 What I read is that git doesn't care about the branch name (except that it creates the "master" branch by default). But, you might want to use other branch names in any system, one reason being due to converting it, e.g. fossil<->git, since fossil uses name "trunk" by default, but again it doesn't care the branch name you can use other names. 02:34:00 (Fossil also has propagating tags too, so you can reference a range of commits using a name of any tag and you can have more than one. The branch name is also a tag, so that can also be used.) 02:35:33 what I don't know is how git knows which branch is the default? when I clone a repo, I don't explicitly tell it to check out the master branch, it decides that on its own. 02:35:41 is there something in the config that tells that? 02:37:41 Does anything but Subversion use that model where you just have one big tree, and it's just a convention that you create top-level directories "trunk", "tags" and "branches" under the root, and make copies of all your code in them? 02:38:07 hmm... I think git checks out the HEAD of the remote 02:39:00 fizzie: no idea, ask someone who uses mercurial, they have like three or four different branch mechanisms so perhaps one of them is that 02:39:11 :-/ 02:39:19 git clone "-- creates and checks out an initial branch that is forked from the cloned repository’s currently active branch." 02:40:10 I guess a bare repository has an active branch as well? 02:40:49 fizzie: apparently 02:40:55 And we use Mercurial at work, though weirdly. 02:42:02 fizzie: aren't you using git at work? 02:42:54 No. Well, except for a lot of open-source things of course. 02:43:22 Chromium and Android and such. 02:43:59 well yeah, everyone has a lot of version control systems installed on their machine these days just to be able to check out third party repositories 02:44:37 that's why I have cvs, mercurial, darcs, fossil, bazaar installed 02:44:58 (svn and git I have used for actual work or hobby) 02:48:00 I think we moved from Subversion to git for the "big" (not that big) shared ASR codebase at the university when I was there doing research. 02:49:36 Hey, that fox is back. 02:50:14 Well, okay, it might be an entirely different fox, I think there's many of them. 02:51:37 Hadn't seen any for a while, but before that I frequently saw one of two at night out there crossing the (pretty quiet) side street or just walking on the sidewalk. 02:52:37 -!- earendel has joined. 03:05:32 new connection would also mean new secure socket layer handshake roundtrips overhead..no? this sucks hard on gprs connections, when scripts demand preflight check for modifications before it serves from cache. a page like irccloud easily make 30-40 https requests. it is a pain. not sure what keepAlive even does. probably additional overhead ping pongs. it in some ways stays open. via stream end 03:12:11 a better protocol would put all requests in a single one. and fetch the whole package deflated, encypted.. vice versa.. and then evaluated "in time". and the user/webdeveloper would only notice when inspecting the network tab. good software never jumps in your way demanding your attention. 03:13:41 Sure, computers suck. 03:14:04 earendel: what is the context for this? and yes, we've been doing multiple HTTP requests per TCP connection long ago, regardless of SSL or not 03:14:18 yes, but not pleasently 03:14:47 well yes, HTTP 2 is trying to improve on that a bit I think 03:15:07 in that it allows you to abort a request without breaking the TCP stream 03:15:21 but the basics work well enough, and routinely used for downloading several small files 03:15:26 [ardon.. a few thoughts. context is my mobile connecion here :/ 03:16:07 HTTP/3 uses UDP and has a 0-RTT encrypted handshake to known servers. Or at least QUIC did. 03:16:34 I mean I have an old custom HTTP downloader program (I really should make a new one soon because it sucks), it has a delay mode where it downloads things slowly, and in that delay more I batch multiple requests together and longer delays between just so that I can take advantage of HTTP keepalive connections 03:16:59 and indeed I've done a debug output to verify that that works, it does not start a new TCP connection within the batches usually 03:17:09 (And TCP keepalive is pretty much just redundant 0-size frames + their acks.) 03:17:27 fizzie: this is a different keepalive 03:17:30 not TCP keepalive 03:18:16 Oh, I got confused, because that was talked about earlier. 03:18:27 i am afraid all of this is crappy om plicated and just doesn't work. or solve the problem it was invented for. obviously. so http is ok. one as well use bundlers. and just.. test a webapp in low bandwith. 03:18:43 right keepalive we were. 03:18:59 Yeah, the HTTP "keepalive" is almost a different meaning of the word. 03:19:14 it's just some stuff in the request and response headers where the server and client negotiate that you can reuse the TCP connection for another request, for which both parties have to make sure that if a request/response has a body, it uses either Content-Length or chunked encoding so the other party can tell where the body ends 03:19:45 but earendel might have asked about something else and I jumped to a conclusion 03:19:46 it has crossed my mind recently that many web pages have heavy JavaScript for doing things like replacing small portions of the page so that there's no need to redownload the whole thing 03:19:56 I don't know the context that earendel was trying to ask about 03:19:59 i was a bit off as i said. 1st sgt logbook entry. i understand we talked about tcp. 03:19:59 when in fact, you could just do it all as plain HTML and it'd be faster due to the smaller download sizes 03:20:41 there are complicated javascript-based web forums which have things like fake scrollbars because the real one wouldn't work properly, because the page isn't loaded all at once 03:20:51 yup. 03:20:55 but, if it weren't for images (which you may be able to use placeholders for), you probably *could* just load the page all at once 03:21:07 and it might well be lighter 03:21:28 ais523: and websites that put on a blur filter and an overlay popup that says you have to buy a payed subscription to view the article, on client side 03:21:52 or similarly a blur filter and an overlay popup saying that you have to claim to be at least 18 years old to view the content 03:22:00 usually also disabling the scrolling behind it 03:22:12 b_jonas: I've discovered that surprisingly many of those, the filter, overlay, etc. just don't load if you have JavaScript off 03:22:34 ais523: nah, on some sites nothing loads if you have Javascript off 03:22:37 my "default" for browsing the web is JS-off (I have ways to turn it on if I need to), and it's actually much more usable that way, surprisingly 03:22:47 yes. images can, and were put into sprite bitmaps to keep conn low. 03:23:00 but images are not used much as back then. 03:23:15 ais523: but sure that happens too 03:23:37 usully u can do a lot with that minimal symbols that are fashinable and minimal (which i like too) 03:24:10 it helps indeed sometimes. 03:24:30 ais523: there's at least one site where I enable javascript just to download the content, AND have a client-side CSS rule to remove the overlay and reenable scrolling 03:24:41 btw. gitter is lynx compatible. 03:25:39 ais523: oh, and you need to accept at least some cookies for it to load too 03:25:53 i think most websites are mining coins when i use their interface in the background. and keeps machine learning my type style. and metriics more terrifying. :> 03:30:05 earendel: probably not very fast, because I'd hear the fan spinning 03:31:11 or contribute to SETI 03:31:37 which i would cancel severly. 03:31:40 ais523: I agree; it does usually work better. 03:32:06 Sometimes, even though nothing is displayed if JS is disabled, you can disable CSS as well and then it will be displayed. However, sometimes this does not work. 03:33:15 (Sometimes by viewing the source you can still find a link to the data, though.) 03:33:36 I discovered recently that Firefox has a built-in way to disable CSS on the page you're viewing 03:33:53 alt-v, y, n 03:34:04 this also lets you change between stylesheets, if the site defined more than one 03:34:07 about:config? 03:34:17 no, in the menus for one page, rather than globally 03:34:27 it's probably got a global CSS disable too, but I doubt it's a good idea to use it 03:35:59 I use that command a lot actually 03:37:15 ais523: yes, that's very old. the alternate stylesheets too, though mostly your readers won't guess they exist unless you include a built-in interface, whether client-side script to change which stylesheets are active 03:37:45 that matrix addon was quite powerful. usually its a bit too much work, but it could pay on 56KBit ~ 03:37:50 b_jonas: still, it means that you could combine the JS stylesheet change, with the browser stylesheet change 03:37:55 and have a page that works for non-JS viewers 03:37:59 also I think different browsers implement the alternate stylesheets a bit differently so you have to be careful, to tell if they're disabled when loading the page, some look at one attribute, some look at another 03:38:31 one attribute tells if it's a normal or alternate stylesheet; the other gives the name of the alternat stylesheet that the menu or scripts can reference to enable/disable it 03:38:59 aaand. markdown would be totally sufficient for informational pages. articles. 03:39:44 I wouldn't like Markdown specifically for that because it has a few annoying corner cases 03:39:48 like in the escaping rules 03:39:57 Another alternative is the simpler gemini format 03:39:58 but, the sort of page that can be generated from a markdown-ish input, yes 03:40:04 or even POD 03:40:46 (Although gemini doesn't have inline formatting; there is no emphasis, strong, fixpitch (except for blocks), or inline links (links on lines by themself are possible).) 03:40:56 there can still be app vr world integrated within toaster and fridge. also rss has to somehow be spun together. then we can drop social media tech monopolies if we want. (we could keep the, if they behave nicely :) 03:41:06 I actually really like POD's feature set, except for the non-semantic I<> and B<> 03:41:31 ais523: yes. but the menu only allows the reader to enable one alt stylesheet and disable all others, or enable no alt stylesheet. if you want multiple orthogonal axis by stylesheet, such as colors and font sizes, you need more than the menu. 03:41:48 I also don't like markdown. it's harder to use than HTML. 03:41:53 well, browsers have a built-in for variable font sizes nowadays too, but yes 03:42:11 I was working on my own markdown-alike, but never finished, there were some things I couldn't decide how they should work 03:42:15 some p2p indexing distributed queryable node mastermind 03:42:21 you'd THINK it's hard to make a markup language that's worse than HTML with all its historical baggage, yet a lot of wiki or forum sites manage exactly that 03:42:25 I /did/ decide that
was worthy of having a tag-equivalent, though 03:42:38 and find an escaping syntax I was happy with 03:42:39 so. i hope u have noted this. :p 03:42:40 it'd be quite marvelous if I didn't want to post on some of those sites 03:42:46 I think plain HTML without CSS or scripts, and with some commands omitted and also some more things such as footnotes, can be good, perhaps. 03:43:04 I think there are three goals for a markdown-alike, in some amount of competition: a) be readable as plaintext, b) be easy to write, c) round-trip well with HTML-alikes 03:43:13 I was primarily focusing on a) and c), but other combinations also make sense 03:43:25 Markdown is moderately bad in all three areas 03:43:45 ais523: yeah, if I make a custom markup language, it will be HTML plus a lot of custom pseudo-tags that are rewritten on the server 03:43:50 hihi 03:43:52 MediaWiki markup is good at c) and moderately good at b), less good at a) 03:43:59 I also had idea of markdown-alike too, with goals a) and b), and without the ability to embed HTML. Also some other differences, such as <...@...> for message IDs, and using ^H and ^W to represent strike-out 03:45:23 But regardless, I should think that a) is important 03:47:13 When writing a web page, ensure working without scripts if possible. Include a proper