| --- Log opened Tue Nov 17 00:00:25 2020 |
| 00:22 | | FLHerne [flh@Nightstar-6tv.748.10.86.IP] has quit [Ping timeout: 121 seconds] |
| 01:12 | | Kindamoody is now known as Kindamoody[zZz] |
| 01:40 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 02:02 | | catalyst [catalyst@Nightstar-4dpc87.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 02:04 | | catalyst [catalyst@Nightstar-4dpc87.dab.02.net] has joined #code |
| 02:09 | | Degi [Degi@Nightstar-9sgdpn.dyn.telefonica.de] has quit [Ping timeout: 121 seconds] |
| 02:18 | | Degi [Degi@Nightstar-asjrr1.pool.telefonica.de] has joined #code |
| 02:29 | | * Reiv hrns |
| 02:29 | <@Reiv> | Hey guys, could I get a hand with a regex? |
| 02:30 | <@Reiv> | I do not understand regex at all but I have a bit of code I need to clean up and it is annoying me. |
| 02:30 | < Mahal> | what is the problem? |
| 02:30 | < Mahal> | I am only mediocre but trying to learn |
| 02:31 | <@Reiv> | I have a couple dozen .asp files with snippets of dynamic SQL that I will be manually cleansing, taking query dumps from, and then shoving uncerimoniously into a data warehouse that they need never be touched again, because the old DB is dead but we need the data for historic purposes. |
| 02:31 | <@Reiv> | "select fnh_descr_33, fnt_text "&_ |
| 02:31 | <@Reiv> | "from fa_note_hdr a left outer join fa_note_text b on a.fnh_text_id = b.fnt_text_id "&_ |
| 02:31 | <@Reiv> | "where fnh_asset_id = "&v_fa_id&" "&_ |
| 02:31 | <@Reiv> | "order by fnt_text_id, fnt_seq " |
| 02:31 | <@Reiv> | I need to match, so I may delete, the strings "&_[newline][all the tabs, number varies]" |
| 02:32 | <@Reiv> | I can handle the leading and trailing quotes, and the where clauses etc will be what I'm manually cleaning up. |
| 02:32 | <@Reiv> | But normalising these things into a single line is going to be an annoying chunk of busywork |
| 02:36 | <&ToxicFrog> | Some regex engines have a "match across newline" option |
| 02:36 | <&ToxicFrog> | That will make your life significantly easier here |
| 02:37 | <&ToxicFrog> | I would probably actually do this in Lua, which matches across newlines by default: |
| 02:39 | <&ToxicFrog> | print((io.read('*a'):gsub('"&_\n\t*"', ''))) |
| 02:39 | <&ToxicFrog> | You can also do it in sed, but it'll be...hard to read |
| 02:40 | <&ToxicFrog> | The actual regex you want, in engines that support multiline matching, is quite straightforward: |
| 02:40 | <&ToxicFrog> | "&_\n\t*" |
| 02:41 | <&ToxicFrog> | (assuming it uses LF newlines or is opened in text mode on a windows system; if it uses CRLF you may need \r\n instead of \n depending on your environment) |
| 02:41 | | * ToxicFrog fiddles with sed |
| 02:50 | <&ToxicFrog> | Reiv: ok, this will fix all of them in place (make sure you have everything in version control just in case it explodes): |
| 02:50 | <&ToxicFrog> | sed -i -E ':start /&_$/ { N; b start; }; /"&_/ s,"&_\n[\t ]*",,g;' *.asp |
| 02:51 | <&McMartin> | This has been making the rounds https://twitter.com/pat_wilson/status/1328240487279411205 |
| 02:52 | <&ToxicFrog> | That's just snake_case with the "types start with uppercase letters" convention, presumably |
| 02:53 | <&McMartin> | OpenGL C APIs are usually glCamelCase |
| 02:53 | <&McMartin> | I think the underscore is unique to GLSL special variables, which also stop existing as of OpenGL 3.x |
| 02:56 | <@Reiv> | Oh, thank you TF! |
| 02:57 | <&ToxicFrog> | I can break down the sed script if you want |
| 02:57 | <&ToxicFrog> | (possibly tomorrow when I'm awake) |
| 02:57 | <@Reiv> | I tried your fancy one but it didn't work, but the simple line worked great. |
| 02:58 | <@Reiv> | The simple one made plenty of sense though, once I comprehended what I was looking at, so much appreciated :) |
| 02:58 | <&ToxicFrog> | ...which one is the fancy one? |
| 02:58 | <@Reiv> | The 'fix them all in place' |
| 02:58 | <@Reiv> | I instead used the Find/Replace feature of my text editor with the regex, and it worked fine tho |
| 02:58 | <&ToxicFrog> | Aah |
| 02:58 | <@Reiv> | So I call that an unmitigated win |
| 02:58 | <&ToxicFrog> | Yeah, the "fix all in place" is a command line that invokes `sed` |
| 02:58 | <@Reiv> | (And now that I see it, it makes sense enough) |
| 02:58 | <@Reiv> | Gotcha |
| 02:59 | <&ToxicFrog> | it's not for use inside an editor, it's a tool for making automated bulk edits to text files |
| 02:59 | <@Reiv> | Nonetheless, thank you very much :) |
| 02:59 | <&ToxicFrog> | (everything from ':start to ,,g;' there is the sed program, which includes the regex but also instructions to sed itself) |
| 02:59 | <&ToxicFrog> | yw! |
| 03:11 | <@Reiv> | Hm |
| 03:11 | <@Reiv> | Well, if I had sed |
| 03:12 | <@Reiv> | I would then definitely use it to find that regex and apply it to files and extract the final result anywhere there was a SELECT*FROM* clear through to the end of that set of &_s, and spit them out in a csv that included file name in an adjacent column... |
| 03:12 | <&ToxicFrog> | If you have linux, wsl, cygwin, or maaaybe OSX you probably have sed |
| 03:13 | <@Reiv> | ... but since I don't, and that's Real Programming, and if I'm honest there's only a couple dozen of these, I will spend my time with this helpful timesaver as I do the rest by hand~ |
| 03:13 | <@Reiv> | I have none of these! |
| 03:13 | <&ToxicFrog> | fair enough! |
| 03:13 | <@Reiv> | I have Windows 10 on a work PC. |
| 03:13 | <@Reiv> | I have a limited selection of software I install freehand, but there are, erh |
| 03:13 | <@Reiv> | If I was super confident in Cygwin or something I could probably get away with installing it? |
| 03:13 | <&ToxicFrog> | Might be worth installing wsl at some point if you anticipate doing stuff like this in the future, assuming it's not just built in to win10 now |
| 03:14 | <@Reiv> | I do this... well, a few times a year, probably |
| 03:14 | <@Reiv> | So yeah, enough for it to be annoying |
| 03:14 | <&ToxicFrog> | (that being Windows Subsystem for Linux, MS's official linux-running-inside-windows feature) |
| 03:14 | | * ToxicFrog nods |
| 03:14 | <@Reiv> | But not usually often enough for me to tool up |
| 03:14 | <&ToxicFrog> | Fair |
| 03:14 | <&ToxicFrog> | My threshold for repetition is basically 0, so I would start installing WSL as soon as there's more than one file to do this to~ |
| 03:15 | <&ToxicFrog> | But I realize that most people would not |
| 03:15 | <@Reiv> | This is "We are decomissioning this server. Please go look at the historic views people still theoretically use for archival data, and pull and dump those output into our new server so we Have The Data in case anyone asks"~ |
| 03:15 | <@Reiv> | I do it more than never, but it is not BAU work either, if that parses |
| 03:15 | <@Reiv> | Maybe I should go learn WSL. |
| 03:15 | <&McMartin> | It's Literally Ubuntu But Launched By Windows with a kernel emulator |
| 03:16 | <@Reiv> | ... neat |
| 03:16 | <@Reiv> | If also terrifying |
| 03:16 | <@Reiv> | oh |
| 03:16 | <@Reiv> | cool |
| 03:16 | <@Reiv> | I have Powershell? |
| 03:16 | <@Reiv> | HEY MAHAL |
| 03:16 | <@Reiv> | How clever is Powershell~ |
| 03:16 | <&McMartin> | Oh yeah PS is standard install now |
| 03:16 | <@Reiv> | Because shit if I could PowerShell this sucker |
| 03:16 | <@Reiv> | IT would mean I could spend the time, write the scripts once |
| 03:17 | <@Reiv> | And then run this on /every/ .asp sucker we have to deal with |
| 03:17 | <@Reiv> | And more importaintly when it comes to Crafting Bespoke Tools |
| 03:17 | <@Reiv> | I could argue with a stragiht face that this is maintainable, aka usable by anyone other than me in the future |
| 03:17 | <&ToxicFrog> | My understanding is that Powershell is quite, well, powerful |
| 03:17 | <@Reiv> | (Which is the issue with Linuxesque style solutions) |
| 03:17 | <&ToxicFrog> | I have not personally used it |
| 03:18 | <@Reiv> | (Not that I object to them, but I can't as easily argue "Oh, just install WSL on your machine and..." when it's Quite A Bit More Complicated) |
| 03:18 | <@Reiv> | Wheras "Yeah, spent some time, here's the Powershell script, have fun" is... |
| 03:18 | <@Reiv> | It can be put In The Toolbox and be expected to be usable by The Next Muggle~ |
| 03:23 | < Mahal> | Powershell is _very_ clever |
| 03:23 | < Mahal> | I am afraid that the things above are not my forte though |
| 03:39 | <@Reiv> | Alrighty, worth asking, cheers |
| 03:54 | | Netsplit Deepthought.Nightstar.Net <-> Krikkit.Nightstar.Net quits: @PinkFreud |
| 04:08 | | Netsplit over, joins: @PinkFreud |
| 04:08 | | mode/#code [+o Kindamoody[zZz]] by ChanServ |
| 04:19 | | celticminstrel [celticminst@Nightstar-5pqf1t.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 04:28 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has quit [Connection closed] |
| 04:28 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has joined #code |
| 04:28 | | mode/#code [+ao VirusJTG VirusJTG] by ChanServ |
| 04:34 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has quit [[NS] Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] |
| 05:13 | | Vorntastic [uid293981@Nightstar-h2b233.irccloud.com] has joined #code |
| 05:13 | | mode/#code [+qo Vorntastic Vorntastic] by ChanServ |
| 05:13 | | gnolam_ [lenin@Nightstar-ik80lk.priv.bahnhof.se] has joined #code |
| 05:16 | | gnolam [lenin@Nightstar-ik80lk.priv.bahnhof.se] has quit [Ping timeout: 121 seconds] |
| 05:24 | | Pink [uid208117@Nightstar-h2b233.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 05:34 | | himi [sjjf@Nightstar-1drtbs.anu.edu.au] has quit [Ping timeout: 121 seconds] |
| 05:36 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has joined #code |
| 05:36 | | mode/#code [+o Reiv] by ChanServ |
| 05:58 | | 720AALZH4 [catalyst@Nightstar-ao6osp.dab.02.net] has joined #code |
| 06:01 | | catalyst [catalyst@Nightstar-4dpc87.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 06:11 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has quit [[NS] Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] |
| 06:12 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has joined #code |
| 06:12 | | mode/#code [+o Reiv] by ChanServ |
| 06:17 | | Pink [uid208117@Nightstar-h2b233.irccloud.com] has joined #code |
| 06:33 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has quit [[NS] Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] |
| 06:43 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has joined #code |
| 06:43 | | mode/#code [+o Reiv] by ChanServ |
| 06:50 | | gnolam_ is now known as gnolam |
| 06:50 | | mode/#code [+o gnolam] by ChanServ |
| 06:54 | | Reiv [NSkiwiirc@Nightstar-ih0uis.global-gateway.net.nz] has quit [[NS] Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] |
| 07:47 | | mac [macdjord@Nightstar-re5.7if.45.45.IP] has joined #code |
| 07:47 | | mode/#code [+o mac] by ChanServ |
| 07:50 | | macdjord|slep [macdjord@Nightstar-re5.7if.45.45.IP] has quit [Ping timeout: 121 seconds] |
| 07:53 | | ErikMesoy1 [Bruker@Nightstar-hcpkod.bb.online.no] has joined #code |
| 07:53 | | ErikMesoy [Bruker@Nightstar-hcpkod.bb.online.no] has quit [Ping timeout: 121 seconds] |
| 08:14 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has joined #code |
| 08:14 | | mode/#code [+o himi] by ChanServ |
| 08:36 | | Kindamoody[zZz] is now known as Kindamoody |
| 09:17 | | Pink [uid208117@Nightstar-h2b233.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 09:33 | | Kindamoody is now known as Kindamoody|afk |
| 10:04 | <@gnolam> | FLIP TABLE postgres; |
| 10:09 | | 720AALZH4 is now known as catalyst |
| 10:59 | | FLHerne [flh@Nightstar-6tv.748.10.86.IP] has joined #code |
| 11:14 | < Yossarian> | err, in irssi or weechat, how do you access a buffer beyond 0-9? if the buffer has double digits? |
| 11:16 | <&[R]> | https://irssi.org/documentation/startup/#basic-user-interface-usage |
| 11:17 | <&[R]> | https://weechat.org/files/doc/stable/weechat_quickstart.en.html#buffer_window |
| 11:21 | < Yossarian> | irssi behavior isn't working in weechat which would be preferable |
| 11:22 | <&[R]> | .... which is why I linked both |
| 11:23 | < Yossarian> | irssi's default bindings would have been preferable for weechat's defaults in this regard |
| 11:26 | <&[R]> | Can't you make bindings in weechat? |
| 11:26 | <&[R]> | Yeah, the /key command |
| 11:34 | < Yossarian> | I'll refrain from asking questions, then, I suppose. |
| 11:38 | < Yossarian> | I have a distinct sense your answers are pissy as if I should have RTFM; btw, discovered in weechat you can find bindings to specific keys with meta-k (key combo you want to see if bound) |
| 11:39 | < Yossarian> | In the input bar is a safe place I guess, useful to see if something is already pre-bound or not. |
| 11:40 | < Yossarian> | which meta-k /input grab_key_command - saves the trouble of typing |
| 11:42 | <&[R]> | Only the elipses answer was snarky |
| 11:55 | < Yossarian> | I can't find anything in the documentation about 'meta2' or 'jmeta', perhaps these bindings are pseudo-bindings for scripts? or extended characters like F1-F10, etc |
| 11:57 | < Yossarian> | ah yes, the delete key is `meta2-3~` |
| 12:31 | | catalyst [catalyst@Nightstar-ao6osp.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 12:32 | | catalyst [catalyst@Nightstar-3j1002.dab.02.net] has joined #code |
| 13:29 | | celticminstrel [celticminst@Nightstar-5pqf1t.dsl.bell.ca] has joined #code |
| 13:29 | | mode/#code [+o celticminstrel] by ChanServ |
| 14:28 | < ErikMesoy1> | I have seemingly outwitted the Spyder IDE. It's giving an undefined name error on code that runs fine. >_> |
| 14:29 | < ErikMesoy1> | There was a variable called "max" and then I did "del max" so I could use the max() builtin function and Spyder thinks that max() can't exist at that point. I should probably rename the variable. |
| 14:32 | | catalyst [catalyst@Nightstar-3j1002.dab.02.net] has quit [[NS] Quit: -a- IRC for Android 2.1.59] |
| 14:32 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has quit [Connection closed] |
| 14:32 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has joined #code |
| 14:32 | | mode/#code [+ao VirusJTG VirusJTG] by ChanServ |
| 14:38 | <~Vorntastic> | Probably |
| 15:18 | | catalyst [catalyst@Nightstar-3j1002.dab.02.net] has joined #code |
| 15:25 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 15:25 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 15:34 | | Kindamoody|afk is now known as Kindamoody |
| 15:47 | | Pink [uid208117@Nightstar-h2b233.irccloud.com] has joined #code |
| 16:51 | | Emmy [Emmy@Nightstar-l49opt.fixed.kpn.net] has joined #code |
| 17:23 | | Vorntastic [uid293981@Nightstar-h2b233.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 17:50 | <&[R]> | Windows is Best OS |
| 17:50 | <&[R]> | https://tech.slashdot.org/story/20/11/16/218209/microsoft-is-showing-windows-10-users-full-screen-ads-for-its-edge-browser |
| 18:00 | | Kizor [moryok@Nightstar-e0a4sm.utu.fi] has quit [Connection closed] |
| 18:01 | <@sshine> | heh |
| 18:01 | <@sshine> | I just asked a guy politely to stop upvoting every single StackOverflow answer I ever made. |
| 18:01 | <&[R]> | https://answers.microsoft.com/en-us/windows/forum/apps_windows_10-winapps-appscat_social/how-to-disable-edge/2a99ad5b-237e-4b20-ad0a-20d731d5ee0a <-- here's a patch apparently |
| 18:02 | <@sshine> | this reminds me of a Danish sketch show concept character, "gentleman nazi". he was such a nice guy, but he got his values all upside down. |
| 18:12 | <@sshine> | [R], wow, so not only does the OS cost money, you even get ads. |
| 18:14 | | Kizor [moryok@Nightstar-e0a4sm.utu.fi] has joined #code |
| 18:41 | < FLHerne> | *and* it collects tracking data |
| 18:41 | < FLHerne> | Hey, at least it doesn't have microtransactions yet |
| 18:44 | < FLHerne> | You may view up to 12 folders per day. Extra folder views are one of many bonuses you might find in a Super Productivity Crate. This week, get 10% off all crates! |
| 19:01 | | catalyst [catalyst@Nightstar-3j1002.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 19:07 | | catalyst [catalyst@Nightstar-6p24lf.dab.02.net] has joined #code |
| 20:10 | <&[R]> | sshine: IKR |
| 20:10 | <&[R]> | Best value |
| 20:16 | <@sshine> | I've seen some articles on Hacker News circulating around how to express why Linux is a better choice for desktops than Windows, as if there's a kind of right argument to be had to trump the lesser, highly subjective ones. |
| 20:16 | <@sshine> | fortunately it's been a long time since I was in the business of convincing people of OS choice. |
| 20:17 | <&McMartin> | Every time I walk up to a freshly-installed Linux VM desktop I'm really happy for about five minutes |
| 20:17 | <@sshine> | I can only say that ever since I realized Steam works on Linux for *ALL MY GAMES*, I don't need Windows. |
| 20:17 | <&McMartin> | And then I remember why I'm not. |
| 20:18 | <&McMartin> | I'm still, notionally, tri-platform, but Windows 10 has taken a solid advantage over GNOME 3 in usability for me, and macOS is now used only under duress. |
| 20:18 | <@sshine> | I ran MacOS for half a year this year. I'd go back if another job offered a MacBook. but in spite of the JVM cracking up my Linux once a day if I switch Spotify track and run Scala tests simultaneously, my main operating systems are still: my editor, and my browser. the terminal, I'm happy, is mostly defined by its text selection and copy-paste behavior, which is easy to mimic across platforms. |
| 20:19 | <&[R]> | I couldn't deal with Windows constantly rebooting while I was working, and I'm not paying 100$ for crap that's already pissing me off |
| 20:19 | <@sshine> | it's not $100. I recently bought a Windows 10 license for a lot less on one of those sketchy websites. |
| 20:20 | <&[R]> | Yeah... not a fan of sketchy |
| 20:20 | <&[R]> | Linux has improved my work productivity significantly anyways |
| 20:21 | <&[R]> | What used to be a bunch of hunt-and-peck shit is now "plug in drive, run script" and just get to work |
| 20:21 | <@sshine> | the alternatives seemed to be paying a premium at Microsoft, or downloading a pirated copy and risk malware. at least I suspect there isn't malware in the activation code. |
| 20:21 | <&McMartin> | Fedora wants to reboot approximately eight times as often as Windows does, these days. |
| 20:21 | <@sshine> | Linux sort of destroys my productivity if I don't stick to the default settings as much as possible. |
| 20:21 | <&[R]> | The malware is that you installed Windows |
| 20:22 | <@sshine> | oh, my GNOME just crashes. I suspect I run out of RAM because 8GB isn't a lot any more. |
| 20:22 | <&[R]> | McMartin: I don't run systemd at all, so that's not even a concern |
| 20:22 | <&McMartin> | I'm talking the cadence of kernel upgrades here. |
| 20:22 | <&McMartin> | Systemd has nothing to do with it. |
| 20:22 | <@sshine> | McMartin, yep. I naively installed a dedicated /boot that ran out of space after a few weeks due to constant upgrades. |
| 20:22 | <&McMartin> | Fedora packages and releases Every Patch Release |
| 20:22 | <&[R]> | Right, but it's a distro that runs systemd, and I don't ever have forced reboot with Linux at all |
| 20:23 | <@sshine> | McMartin, but do you reboot every time there's an upgrade? |
| 20:23 | <&McMartin> | GNOME would like you to! |
| 20:23 | <&McMartin> | But you don't have to. |
| 20:23 | <&[R]> | That's annoying |
| 20:23 | <&McMartin> | If I'm going to bother updating the kernel, though, yes, I'm going to reboot into the new kernel. |
| 20:23 | <@sshine> | MacOS wants you to reboot *constantly*. those messages are so annoying. they only let you postpone one day. |
| 20:23 | <&[R]> | Yeah |
| 20:24 | <@sshine> | surely we are ready for hotswapping kernels. |
| 20:24 | <&McMartin> | As of Windows 7 SP1 or so MS has been pretty solid about only trying to do reboot-necessary upgrades once a month |
| 20:24 | <&[R]> | sshine: it is implimented |
| 20:24 | <&[R]> | And you don't need to reboot all of userspace when you do it apparently |
| 20:24 | <@sshine> | isn't Windows 7 unsupported now? |
| 20:25 | <&McMartin> | Yes. This has been a thing for longer than their support window. |
| 20:26 | <@sshine> | I had to update my mom's computer a few months ago, and I realized Windows 7 wasn't a good choice any more. and Windows 10 simply didn't support her hardware. so I got her a ThinkPad x250 and put Windows 10 on it and made it look like classic Windows. |
| 20:27 | <&McMartin> | Also: Fedora 33 runs pretty well on a 2015-era NUC, which is running a Celeron-class Atom CPU and only has 8GB of RAM. |
| 20:27 | <&McMartin> | *Firefox* chugs |
| 20:27 | <&McMartin> | And if I try to do scaled video playback my framerate will die |
| 20:27 | <&McMartin> | But pretty much everything else is Fine (tm). |
| 20:28 | <&[R]> | I think we really have to blame advertisers for the web becoming so RAM hungry |
| 20:28 | <&[R]> | Know what I don't miss? Flash. Flash and it's "lol 100% CPU at idle"ness |
| 20:28 | <&[R]> | Gotta peg that one core at idle man |
| 20:29 | <@sshine> | I have quite a lot of stuff disabled by default. |
| 20:29 | <&McMartin> | Checks out. I say many bad things about JS but making it the standard was a step forward |
| 20:30 | <&McMartin> | ... also, of course, nowadays we're cheerfully building flash emus and now Newgrounds is better preserved that the mobile app store games~ |
| 20:30 | <@sshine> | fortunately most of my browsing preferences tell me that if I'm on a website that informs me I've disabled too many features, I'm on the wrong website. |
| 20:30 | <@sshine> | in spite of this, Firefox is leaking memory like hell. |
| 20:30 | <&McMartin> | *than the |
| 20:30 | <&[R]> | Yeah, FF does not do well when it's long-running |
| 20:30 | <@sshine> | McMartin, also, WASM is getting close. |
| 20:31 | <&McMartin> | I consider WASM a speed hack at this point |
| 20:31 | <&McMartin> | I'm aware of the reasons this is unfair, but. |
| 20:31 | <&[R]> | I know a guy that compiled PHP for WASM |
| 20:31 | <@sshine> | [R], someone told me that Google is keeping Firefox alive-and-crippled by funding it enough so that it never really competes with Chrome. |
| 20:31 | <&[R]> | So he can run PHP in the browser |
| 20:31 | <@sshine> | McMartin, I somehow assumed that WASM would remove some sources of memory leaks in browsers... |
| 20:32 | <&[R]> | lol, nope |
| 20:32 | <&[R]> | Could probably add new ways to leak memory |
| 20:33 | <@sshine> | as in, GC in JS is hard because when is something referenced. |
| 20:33 | <&McMartin> | Even emscripten will cheerfully and carefully emulate all your buffer overflows :) |
| 20:33 | <&[R]> | I'm pretty sure the leaks are mostly handles and UI stuff |
| 20:33 | <&McMartin> | But WASM will improve the results of stuff jammed through Emscripten that were intended for non-web platforms. |
| 20:33 | <@sshine> | whereas with WASM, quality of GC, I presume, becomes a function of the many other languages that get compiled, not just JS. |
| 20:33 | <&[R]> | Stuff that JS doesn't really touch |
| 20:34 | <@sshine> | McMartin, as in, the many things you transpile into JS primitives that aren't meant for being targetted at compilers and will work but generate so much garbage? |
| 20:34 | <@sshine> | s/at/by/ |
| 20:35 | <@sshine> | [R], "leaks are mostly handles and UI stuff": I'm totally out of my leak, I'm just writing WASM softcore fantasy. |
| 20:35 | <&McMartin> | The uses of WASM I've had to deal with the most are actually using C or C++ as the fontend language. |
| 20:35 | <@sshine> | s/leak/league/ |
| 20:36 | <&McMartin> | So I'm pretty sure that what's happening is, with some fluffy cloud handwaving here, that a big array is being allocated and presented to the program as "its memory" |
| 20:36 | <@sshine> | yes. I'm waiting for Haskell's Asterius compiler to mature. |
| 20:36 | <@sshine> | I wanna write Haskell in the front-end, too. |
| 20:36 | <&McMartin> | And so from the point of view of the browser and the OS, the C program is using A Big Old Chunk Of Memory, but it's constant and permanently referenced. |
| 20:36 | <@sshine> | ah. |
| 20:37 | <&McMartin> | But that also means that even in your bounds-checked WASM implementation you're always In That Array so overflows are legal within it |
| 20:38 | <&McMartin> | ... of course, your return-oriented programming exploits aren't gonna work ;) |
| 20:41 | <&[R]> | Is it a VM of sorts, or is it actually allowed to touch the CPU? |
| 20:43 | <@sshine> | it would be nice if websites CPU/RAM usage was generally confined to some fixed amount so you know approx. how many active tabs you can have before your computer bugs out. of course, that's only possible if your programs have expectable behavior. sort of like how blockchain contracts come with a very explicit cost model bounded by gas price. |
| 20:43 | <&McMartin> | That's up to the implementation, I think. If you JS your JIT, you're touching the CPU... |
| 20:43 | <&McMartin> | That sounds like an RTOS |
| 20:44 | <&McMartin> | Also as is usual, if something is being hyped by Blockchain Technology Enthusiasts, make sure somebody didn't solve the problem in a much better way 30 years previously because they probably did |
| 20:44 | < Pink> | That is excellent advice. |
| 20:46 | <&McMartin> | Programming/CS in general is infamous for disdaining domain-specific knowledge and relying on Golden Hammers that are poor fits for all kinds of things |
| 20:46 | <&McMartin> | ... and I'm pretty sure that the blockchain types are what happens when that process is applied recursively |
| 20:47 | < Pink> | But... have you considered blockchain voting? |
| 20:50 | <&McMartin> | For as long as it took for them to describe what they think the problem is that they are solving, and that was enough to bury them in the desert |
| 20:50 | <&McMartin> | Fortunately, we have a recent paper whose draft is out on this topic! https://people.csail.mit.edu/rivest/pubs/PSNR20.pdf |
| 20:50 | <&McMartin> | Yes. That Rivest. |
| 20:51 | <@gnolam> | Eh, what does that guy know about crypto?~ |
| 20:51 | < Pink> | That was the joke. |
| 20:52 | <&McMartin> | I'm being professional today |
| 20:52 | <&McMartin> | If I *weren't*, I'd suggest that many of these "serial entrepreneurs" believe that they have invented lying, and then, as a corrolary, believe that everyone else is telling the truth at all times. |
| 20:54 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has quit [Ping timeout: 121 seconds] |
| 21:13 | <&ToxicFrog> | > I'm still, notionally, tri-platform, but Windows 10 has taken a solid advantage over GNOME 3 in usability for me, and macOS is now used only under duress. |
| 21:13 | <&ToxicFrog> | McMartin: I'm curious if you've tried KDE5/plasma or if it's been strictly GNOME |
| 21:13 | <&ToxicFrog> | > as in, GC in JS is hard because when is something referenced. |
| 21:13 | <&ToxicFrog> | When it's a global, a data member of something referenced, an upvalue of a closure that is referenced, or a local on the stack? |
| 21:14 | <&ToxicFrog> | I'm pretty sure modern JSVMs all use mark-and-sweep variants and that's a very well understood and reliable tech. But that won't help (in any language) if the code being run keeps generating huge amounts of garbage and holding references to it. |
| 21:22 | | ErikMesoy1 is now known as ErikMesoy |
| 21:22 | | mode/#code [+o ErikMesoy] by ChanServ |
| 21:27 | <&McMartin> | I've only tried Plasma very briefly, and it was early enough that I'm not sure it was the KDE5 version. |
| 21:28 | <&McMartin> | That said, one of the reasons I stick to GNOME 3 generally and Fedora's GNOME 3 specifically is because it's the "default" windowing environment for people using Linux desktops these days, and Fedora modifies stock the least. |
| 21:28 | <&McMartin> | And a big part of why I use it is to make sure that things like UQM and VICE still work right when GNOME breaks everything. |
| 21:29 | <&McMartin> | (This also allowed me to meaningfully make sure that stuff kept working right under Wayland instead of X11, etc; Fedora is the most reliable bleeding-edge distro I've worked with and that makes it ideal for this use case) |
| 21:30 | <&ToxicFrog> | Fair enough |
| 21:31 | <&ToxicFrog> | (although note that whether it's "the default" depends on your distro; KDE5-plasma/wayland is the default on SUSE, for example, and GNOME is included on a "best effort" basis) |
| 21:31 | <&McMartin> | Yeah. As of October of 2018, though, both Ubuntu and Fedora shipped GNOME 3 as the default. Prior to that I kept an Ubuntu machine running to make sure Unity was OK with everything too. |
| 21:32 | <&McMartin> | Fedora tracks latest stable GNOME in a way that no other distro does, though. |
| 21:33 | <&McMartin> | I believe but am not positive that KDE Plasma 4 was the KDE/Plasma I tried and I remember bouncing off of it extremely hard |
| 21:33 | <&McMartin> | But also being told that Plasma 5 was unrecognizably different and thus actually usable. |
| 21:33 | <&ToxicFrog> | I don't think i used kde4 ever, I went straight from 3 to 5 |
| 21:34 | <&McMartin> | It is quite possible that this was Extreme Wisdom. |
| 22:03 | <&[R]> | I remember being willing to kill to have a feature from KDE4 in Windows |
| 22:03 | <&[R]> | But that might also be because our workflow is a bit wonky |
| 22:03 | <&[R]> | They had a desktop gadget that let you have a file explorer "window" to any directory |
| 22:04 | <&[R]> | I remember being upset that Win8? didn't have it (whatever version had the gadgets) |
| 23:00 | | himi [sjjf@Nightstar-1drtbs.anu.edu.au] has joined #code |
| 23:00 | | mode/#code [+o himi] by ChanServ |
| 23:35 | <&McMartin> | Like, throw mouse to lower right, get the build directory sliding out? |
| 23:35 | | Emmy [Emmy@Nightstar-l49opt.fixed.kpn.net] has quit [Ping timeout: 121 seconds] |
| 23:56 | <&[R]> | Like there's literally a box on the desktop that's an entirely different directory than the Desktop directory |
| 23:56 | <&[R]> | And you can have more than one |
| 23:56 | <&McMartin> | Oh I see |
| 23:57 | <&[R]> | So you could have "working" "needs report" "ready for archiving" etc... |
| 23:57 | <&McMartin> | ... come to think of it, part of what threw me about KDE4 may have been that I seem to recall that its notion of the desktop metaphor was nonstandard. |
| 23:57 | <&[R]> | Our system was literally to cluster them in various parts of the desktop |
| 23:57 | | * McMartin nods |
| 23:57 | <&McMartin> | Kind of a tiling WM under your proper desktop WM. |
| 23:57 | <&[R]> | Which worked great until Windows decided to chose a different resolution for whatever reason and just jumble all the stuff on the desktop |
| 23:58 | <&[R]> | Except instead of a window it's just icons on the desktop |
| 23:58 | <&[R]> | But they're in categorized boxes |
| 23:58 | | Kindamoody is now known as Kindamoody[zZz] |
| 23:58 | <&[R]> | Or windows rebooted wrong and forgot things, like settings, or where the various icons on a desktop were supposed to go |
| 23:59 | <&[R]> | Now I just have a bunch of aliases to cd to the right directory, a multi-line prompt that gives me a detailed breakdown of the status of a job when I'm in a job directory (each job gets one directory) |
| 23:59 | <&[R]> | Scripts to move things around |
| --- Log closed Wed Nov 18 00:00:27 2020 |