00:18:48 WG9s: indeed.. thank you fro that tidbit 00:19:24 while I will have visual studio 2022 installed on here when I reformat i really don't want to ahve to have that as a requirement.. 00:20:02 most people who want to build the code never open visual studio and i want some additional choices for building so if the tools alone plus redist is sufficient.. then wonderful 00:30:30 i build using clang-cl 00:30:41 which momes with mozilla build 00:31:24 you also need the sdk 00:31:41 from VS 00:32:13 i only eve opn visual studio to update it bt it usually only updates tings i do not use 06:29:57 WG9s: FYI: The L64 build announced by email as "completed Tue Jan 31 18:26:12 EST 2023" has a date of 2023-01-29 22:00 on the HTML page http://www.wg9s.com/comm-253/ but its bz2 file length is different from what I downloaded in the morning of 30 January. 06:39:28 WG9s: P.S. Build ID 20230129220006 10:32:34 tonymec|away: not sure why it shoul have a different length. It should have been built from the identical source. 10:35:35 there was not supposed to be new build last night but I screwed up again. 11:15:08 tonymec|away: was supposed to only be doing a test macOS build 12:18:49 ah, OK. 19:06:54 real question WG9s is how can I alter the build system so it will accept build tools if full vs isn't there 19:07:17 vswhere isn't actually coded to consider the build tools AS visual studio for some reason they say is a feature not a bug 19:07:28 tho i did get a successful build 19:09:03 hmm not sure I have always isntalled vs even thought i really do not use it. 19:11:34 well i am thinking more along the lines of using a clean VM to actually produce my shipped binaries again 19:12:02 i used to do that 19:12:09 but other things became more important 19:12:55 my system is always clean but it was that extra level of assurance.. by having the VM isolated from the host to a good degree and upload direct from vm to server 19:13:06 i mena i have always done that for linux 19:13:08 but winders 19:13:46 and if i can get away with JUST the build tools on a windows 7 vm that means my overhead is only a gig above libevil linking 19:13:57 if i have to use windows 10 it will be twice that 19:14:16 yes i could use server core but then i have a crippled experience 19:16:18 and if i remove a bunch of components.. yes the base windows system works but most of the shit is now tied into twinui tied into dwm so even if you go as far as explorer not being the shell.. you literally can't access half the shit in windows 10.. and 75% in latest windows 11.. in fact if you want to do ANYTHING important outside of an mmc snapin or the registry or the tabbed dialog control panel applets of old you simply can't without explorer 19:16:18 running IN SHELL MODE 19:16:39 so yes you can TECHNICALLY EVEN IN 11 BYPASS ALL THE MODERN HORSESHIT after inital setup 19:16:56 BUT you loose almost complete control over your system in the process 19:17:00 cause there ain't much left 19:17:04 not infected 19:18:01 this is why what I am shifting to will be important least in my view 19:18:50 rejoice i won't be in the market of competing with you with a huge chunk of your own frontend application anymore! 19:53:36 CaptainTobin there were some changes in the build system where it now calls the external vswhere. I am not sure if I ported the detection of the build tools too. Usually just install selected components blow away all vctip.exe binaries and everything in the folder servicehub. Then vs2019 quiets down. IDE no longer works but don't care. 19:57:04 as far as I know the build tools are a supplement license so to comply fully you technically need the base license. Not that anyone cares if you don't get one for community and only do open source I suspect. 20:03:00 i think MCP might be violating some terms by putting vswhere inside of the codebase 20:03:12 at least not putting it in other-licenses 20:03:29 but he doesn't want to maintain a mozillabuild package 20:03:38 i am gonna because i also want additional tools 20:14:51 Bug 1522560 20:15:27 yeah 20:16:09 still i would like to make it capable of using both the tools and the full vs 20:17:26 if you consider the windows crazy looking to compile him self is MORE likely doing it for them selves not actual development.. so just having the tools, modified mozillabuild package, build and go and don't even need to worry about redist in that case cause they already have the cpp runtime on their machine 20:17:55 but there has to be a better solution that makes it as easy as it was before but allows latest installed vs and vsbuildtools 20:18:35 sucks it is early in mozconfigure and not just ac-configure.. i'd have that sorted in 20 minutes 20:22:39 CaptainTobin: sorry like I say I install VS but do not actually ever run it. 20:22:53 i just sue the SDK and redist dir 20:22:57 i use it OCCASIONALLY 20:23:00 VERY OCCASIONALLY 20:23:25 but it would be nice to not have to choose either or 20:23:30 i should be able to use both 20:24:03 hell except for crash dumps visual studio is pretty fuckin useless for us as mozdevs unless you use the ide to code in of course 20:25:12 i hate mozconfigure 20:26:35 so if it has to be there for a build to work I was not ever even aware of that 20:27:21 i mena it LOOKS like i can simply add another for loop just dup and alter.. but this flow is confusing me.. 20:27:29 guess will just have to TRY AND SEE 20:28:04 not in the def but how this relates to anything else 20:28:23 it is why i hate mozconfigure and python by people who think they are smart but aint 20:29:05 i copy the resdist dir to aanother location a refeence it from there i guess it might be using the vs install directory to find the sdk to use 20:29:13 why the hell can't the batch file call vswhere and just populate the same damned vars it always did 20:29:53 i will likely just bundle the redist files with vswhere in my custom mozillabuild package 20:30:01 and put the mapi includes there too 20:30:22 well maybe not that 20:30:24 but still 20:30:28 i copy the mapi includes and redist not sure how it finds the sdk 20:30:48 the mapi includes are now in the codebase for everyone but mcp i believe 20:30:57 tho a200 might have done that i haven't looked 20:31:27 WG9s: how it finds the SDK depends on the codebase 20:31:34 differences between versions 20:31:43 and mcp isn't the same as 52 anymore 20:31:53 even more so now since they did the vswhere shit 20:32:02 and i use the clang-cl that is installed by mach bootstrap 20:32:20 so it isn't as simple as following the mozbugs for me i have to deal with what and how they did relating to what I did in the past 20:32:28 well i simply won't use clang period 20:33:00 unless someone fully funds me a powerful apple workstation whatever they have that is closest to that class anyway 20:33:09 then I will use clang.. on macos 20:33:24 i believe in prevailing system compilers being best 20:33:46 gcc broadly on linux, msvc on winders, and whatever apple has 20:34:02 that IS a matter of experience informing opinion 20:34:10 and one that ain't universal 20:34:49 and as we know WG9s a significant enough change or even a build system rejigger can throw out current day stats on whatever compiler it is atm 20:34:55 well using clang makes debugging and crahreporter better. clang and rust are just both different front end parser things for the actual code generator for which they both use llvm so for rut and c and c++ code in one executable better to use clang because then using the same code generator and linker for everything 20:35:08 one change may make it compile great on clang everywhere another change may favor platform compilers across the board 20:35:30 it just comes down to what you are producing, how, and just what the real world result actually is in that regard 20:35:38 just saying if some of the code is in rust better to use clang 20:36:03 there is no rust 20:36:18 if i have to include rust.. then the project all projects end 20:36:30 I will NOT work with that shit. 20:36:33 clang and rust are just the front end parser part of the compiler the real comiling happens in llvm and better to use that for both rust and c 20:36:45 betetr 20:36:50 better based on what? 20:37:06 what makes clang better for what I specifically want to do 20:37:19 why is it better than visual studio and gcc for my specific case 20:38:12 and tell me WG9s just how many times have you compiled an evolved 52 on windows and linux with clang 20:38:48 and where do i get this all inclusive package that I can just use without changing anything? 20:39:52 becuase you are now using the same compiler for both the rust and c code so debugging works better becuase the debugger only needs ot understand what llvm builds vs what llvm builds for rust and what gcc builds for c same with crashreporter and definitily works better for PGO but then I have no use for PGO so I don;t care about that 20:40:01 what rust code? 20:40:06 what crashreporter 20:40:11 what pgo 20:40:32 the large protions of mozilla that is written in rust 20:40:43 good for mozilla 20:40:50 luckily I escaped that fate 20:40:51 i cont really care about crashreporter but others do 20:41:00 not unscaved but still pretty well intact 20:41:04 i thought it was seamonkey you wre trying to build? 20:41:04 who? 20:41:09 who cares about crash reporter 20:41:14 other than Mozilla them selves 20:41:17 large prtions are in rust 20:41:33 http://xr.binaryoutcast.com/aura-central/ 20:41:35 show me 20:41:52 parts or the mozilla code that is written in rust is actully chromium code that is written in rust 20:42:03 well show me dude 20:42:10 where in my codebase is the rust code 20:42:21 aside from left over stubs that most never did anything anyway 20:42:45 ipc is a nasty piece of code btw 20:43:22 WG9s: http://xr.binaryoutcast.com/aura-central/find?string=\.rs 20:43:24 ;) 20:44:03 http://xr.binaryoutcast.com/aura-central/search?string=MOZ_STYLO&find= 20:44:07 leftovers nothing more 20:44:48 *I* Do not require nor employ rust nor have any intention to do so. 20:45:03 CaptainTobin: I never liked the ipc code has to do with why mozilla dropped the old type extensions. they wanted to to soemthing like ipc and what google did did not work with the legacy extensions. 20:45:26 they STARTED on their own IPC loosely based off a decent little project 20:45:31 but then scrapped it for google 20:45:32 but mozilla had no clue how to do this on their own so just snagged the google code 20:45:39 and scrapped their other shit for google 20:45:46 yes 20:46:13 and all of this happened between 2008 and 2009 20:46:57 the UX Team taking over firefox, firefox becoming the ONLY priority, mobile bullshit, literally everything comes down toa fundamental shift in mozilla that basically lead us all to today 20:48:06 yes making the desktop interface look more like the iPhone 20:48:22 but regardless yeah my point going forward is NOT making a web browser that can actually compete with google.. that is a lost cause for the moment .. i want to focus on the classical technologies and all they can do in ALTERNITIVE to the OpenWeb agenda 20:48:48 and shit as a service operating bullshit 20:49:26 I am not building an integrated internet suite either.. what I am doing is far more grand in vision than one kind of client or class of suite 20:49:54 I am going for broke.. a core extensable operating environment .. initally tabbed-based. 20:49:58 well you know google chrome only exists becuase mozilla pissed off Ben Goodger He was the lead engineer on Firefox form back when it was still call Phoenix. and when they pissed him off he left to go to google and then they have CHrome 20:50:24 i was not aware of that specific tidbit interesting 20:53:59 well early chrome was a good little browser served a particular couple of core reasons to exist.. spite, web platform testing, basic minimal browser to chip away at IE some more.. seemed like it was fine 20:54:09 THEN CHROME BECAME THE GOOGLE AS MUCH AS SEARCH AND ADS ARE 20:55:19 Back in the day were about 4 of us worked on Thunderbird in the early days. Scott McGreggor was the lead but mostly worked on the UI. Dave mose worked on the mailnews backend I cknd of did a bit of each and also was trying to help Dave to get the impa code to work with more imap server the original code only really worked with one impa server that once and actual imap standard came about... 20:55:21 ...ws noreally stadard comliant. 20:55:43 and then there was arvid who did the icons for the theme 20:55:45 but the shit that google did then and mozilla did soon after and that microsoft and apple got to do more of.. brings us all here where i have to dump my plans and start with a new idea to basically create 90% of a shell or desktop environment in extensiable xul 20:56:57 WG9s: can you explain why Thunderbird despite wanting to chase Firefox UX never really focused on making their application specific code cohesive.. it is a very bizare mix of pseudo-xpfe, overlays, protojsms, etc 20:57:03 it's a god damned mess 20:57:38 so then mozilla decided they did not want to support this as a tier 1 product and wanted to make it more community supported and gave use 3 choices for how we wanted to proceed asked us to vote and we unamioulsy pcied one of the 3 choices they presented and they said tough shit you picked the wrong one. Scott immediately quit mozilla and he and Ben were good friends and I think this is why... 20:57:38 then again unlike Firefox Thunderbird actually DID Decend from the mailnews ui 20:57:40 ...he left soon after. 20:57:46 Firefox was constructed using bits and pieces 20:58:08 WG9s: have you written about this stuff 20:58:25 so the guy who i am not sure he is still at google but was the chrome guy when it came about was the former firefox guy at mozilla. 20:58:44 cause someone needs to really construct a look into what really happened during that era when everyone was obsessed with finally getting their iphones after the first few years 21:00:02 WG9s: anyway it is my belief that rust servo clang and continued google is what has nearly obliterated a once great set of technologies 21:01:02 and i am not satisfied to try and hold the line simply to keep what i have going.. i want to create.. create new things.. new interpretations.. new concepts like succeeding where mozilla constantly botched the same few concepts repeatedly but everyone else seems to do fine.. 21:01:06 rust and servo are the same thing. servo is the codename for rewtiing gecko in rust 21:01:07 not mobile tho 21:01:36 rust is the language servo is the gecko port/rewrite, stylo is the component to replace nsLayout 21:01:46 and there is spidermonkey infection too now 21:01:48 i know dude 21:02:07 well layout style 21:02:25 webrender seems to deal with the other layout bits as well as layers and accel 21:02:26 but not sure clang is bad gcc is kind of clunky and never desigend to do c++ 21:02:32 little fuzzy on that tho 21:02:45 i am not saying clang is bad 21:03:22 but if it isn't it will be .. eventually.. but regardless.. i just feel that one compiler across disperate operating systems isn't as good as using the compiler that specialozes in that target os 21:04:11 especially when dealing with pre-quantum code which largely hasn;t seen clang outside osx and has no real testing in that configuration nor can I provide any nor do I have time to make it happen regardless 21:04:15 also i dont wanna 21:04:16 and clang seems more suited to being able to do cross compiles 21:04:17 WG9s: lol 21:04:58 which isuspect is why you say ok for osx becuase you are cross compiling for that? 21:05:06 no 21:05:14 which is why I can't properly support mac 21:05:27 i don't have the hardware nor a viable virtualization solution 21:06:18 y es so with clang i can use my linux clang compiler and set the target and get it to compile and link for macos 21:06:36 jusing gcc i would need a seperate compiler to do this 21:06:50 WG9s: compile the aura runtime environment with clang for windows by tomorrow morning and I will agree with you. 21:06:54 deal? 21:06:54 I like choice and so clang fits the bill. Not the first time one compiler is buggy and produces bad code. The reason clang is currently used for official Windows x86 is just this. But would still like to get rid of 32 bit builds :) 21:07:26 frg_Away: 32bit windows is becoming increasingly problimatic 21:07:30 across the board 21:07:42 and it is still only due to artifical limitations 21:08:06 if microsoft would release a patch for 7 and up removing the ram restriction from 32bit and wow64 .. we'd be fine 21:08:16 but they want it dead as much as everyone else does so they won't 21:08:38 windows just gets slower and slower becuase of the anti-virus overhead to the point were in my opinion 32-bit windows is too slow to be usable 21:09:35 as such next year when EL7 is dropped I will also drop gcc lower than 10, any possibility of vs older than what i use, and 32bit windows builds. Also i will NOT be introducing any NEW 32bit builds.. Also likely gonna drop gtk2 then as a produced build but leave it capable for a while longer 21:09:49 WG9s: what anti-virus overhead lol 21:09:53 i don't have windows defender 21:09:57 gone 21:09:58 lol 21:10:10 i got something else i can just run on demand 21:10:19 and i do manual scans 21:10:28 but the issue is the average user is so if you are doing win32 builds who is the actual customer? 21:11:07 people who would simply use the last version regardless because their old and insecure computer is important to them because their desktop icons are in a specific order 21:11:19 stuck on 3 gig of memeory max if your system supports it otherwise 2giig max system memory 21:11:23 WG9s: or did you want an answer that was less truthful but less offensive 21:11:24 lol 21:11:47 i did not find that offensive 21:11:48 Windows Defender is actually the best. Just get rid of the cloud crap stuff in it. Death to Smart screen and block on first sight! 21:12:11 and the other phone home tools in it. 21:12:16 Defender is fine but if i have it it finds a way to reactivate even tho i lock out the reg key, files, etc 21:12:22 i have to REMOVE it from the package store 21:12:25 so it CAN'T come back 21:12:46 group policies do it. 21:12:58 i don't even HATE all the modern shit i HATE lack of control and choice 21:13:04 frg_Away: not in windows 11 21:13:16 tamper protection can override group policy 21:13:20 bjt still i find 8 gig memory systems as slow as i want so sin32 can only use 2 gig if you have 4 unless your system supports the 3gig kludge thing but still e gig only p[er app 21:13:22 it did in the betas 21:13:27 so they can re-enable it 21:13:30 2 gig only per app 21:13:33 That is why I run Server 2019 :) 11 is crap on a new level. 21:14:09 I am gonna have to either stay on LTSC 2021 go back to 2019 or go linux and just write eveyrthing else in xul 21:14:13 i mean 21:14:15 at this point 21:14:21 it is coming to that 21:14:43 Server 2022 when you can get it cheap next. 21:15:08 And as a bonus: No Windows store in the server versions! 21:15:23 i was HOPING i could just get away with the free hyperv server core and remove the hyperv role put bblean on there and call it a day 21:15:39 i had an old 32 bit machine and gave up on running winows on it switched to linux and then gnome 3 showed up so I had to witch to a different than default desktop to get something that actually ran on it at a useful speed 21:15:45 BUT hyperv server core has its OLE crippled and without DWM gdi can BARELY function 21:15:59 it is in a trash heap someplace now. 21:16:17 nut yeah can you believe they crippled drag and drop and most of the clipboard 21:16:27 the ONLY drag and drop works is actually in winfile 21:16:35 cause i think its DnD predates OLE 21:16:41 so it does it its self 21:17:14 i dunno if normal server core is that crippled 21:17:23 back on win32 when it was slow i rember we called this drag until you drop becuase it was so slow you thought you would die before the drop actually completed. 21:17:35 I think all core versions are discontinued now but unsure. 21:17:46 plug and pray 21:18:19 yes another pne 21:18:22 one 21:18:43 frg_Away: well at this point none of the GUI is really functional even if present in server core at 11 so there is little point unless they include dwm and hack twinui to run without explorer its self 21:19:30 hmm so a nice bitchfest, but was there something useful we were discussing before we got off on this tangent? 21:19:37 oh btw without explorer as a shell.. windows 11 explorer goes back to windows 10 real explorer mode with a ribbon but gets badly busted 21:20:09 so eyus the new windows 11 shit is on top of previously good but now crippled interfaces 21:20:13 what insanity is that 21:20:38 windows is a doomed platform 21:20:43 that is my conclusion 21:21:02 I see nothing in windows 11 that is an improvement over windows 10 21:21:21 few kernel improvments.. same as 10 same as 8.1 same as 8 21:21:29 but everything else just makes it harder and harder 21:22:13 how does a company that produces very little that is actually new (tho often changed or crippled) can they still be in such power 21:22:16 unually i say overall is see more negative than positive in a new version and eventi=ually that kind of gets sorted but this time is every change seems to have just made things works going from win10 to win11 21:22:20 subserviant to google yes but why 21:22:48 i BARELY have Windows 10 LTSC in a state i can use it nearly identically as Windows 7 21:22:59 i don't think I will get Windows 11 in the same state 21:23:06 hell the control panel is crippled 21:24:13 and because the windows 11 UX is dynamically generated differently and overlayed/replaced on the windows 10 ux still using largely traditional means.. altering the traditional means doesn't eman you will SEE any difference when Windows can just decide to replace entire swaths of UI on the fly with the compositor 21:24:17 WG9s: consider that 21:24:32 Well 2022 is just like 10 with a higher build level and support till 2031 https://imgur.com/a/3oBOtsH 21:24:33 a compositor that as of Windows 8 breaks winstation session boundries 21:24:38 TELL ME HOW THIS IS GOOD FRG 21:25:35 windows.. can now dyanmically and transparently replace client area UI bits wholesale through the desktop compositor 21:25:57 and i thought the Firefox UX Tour that breaks boundries between content and chrome was bad 21:26:43 oh and don;t tell me about windows composiros was a big thing i had to fix back in the firefox 4 days ahving to do with getting it to work correctly with multiple monitors. 21:27:16 and isn't that one of the things used to shit on xul.. remote chrome, altering ui.. They want the technology as much as we do.. they just want to be the ONLY ones that have it and benefit 21:27:52 it's like the radical portions of any cause 21:28:46 the real issue is they came up with this compositor thing that ws kind of a windows manger for applications, but for the auto hide taskbar thing to work correclty should have been doing the subtrack on for the number of lies the app could use but did not so every app had to figure out how to figure out what to do to make sure the auto hide and unhide the tasbe thing work but stood have been... 21:28:48 ...taken care of by the compositor 21:29:07 they don't want equality or fairness.. they want supremecy .. they don't want peace or disarmerment or acceptabce they want to dominate as they have been dominated. 21:29:37 was epecially tricky if you had mutiple monitors and they had differeing screen resolutions and even worse if you had moved the taskbar to the secondary monitor 21:29:53 yeah 21:30:34 getting this all to wrok was my big contribution to getting firefox to work correctly on mutiple monitore under windows vista (which was ms latest debacle os at the time) 21:30:59 And the credgeview crap is now used internally and can't be uninstalled in 11. But doesn't matter. Linux is a 1000 different shards and macOS becoems an even smaller walled garden with apple proprietary hardware the future. So we can bitch all day and it will not even make a small dent in the universe. 21:31:26 so that is why i have a firsfox 4 beta team tshirt from the mozilla folk from back before they baned me from bugzilla 21:33:05 btw 21:33:10 webview2 21:33:15 comes with widevine 21:33:53 so there is that .. one person at microsoft got them to include it with it so devs that use it can get eme without having to beg google 21:34:07 so heh.. i have an idea 21:34:10 about that 21:34:15 not an orginal one mind you 21:34:21 but it could be good 21:34:34 a stupid thing where you can put the taskbar at diffent places on different screens but for some stupid assed readon hiding and unhiding only works if the taskbars is at screen bottom if you prevent the app from using the bottom row on the screen 21:35:05 i have noticed that windows is becoming really glitchy about z-order 21:35:35 applications change z-order strangely for no reason occasionaly when you hover over them and no I don't have that thing tat explicitly does that 21:35:44 taskbar vs fullscreen youtube fights 21:35:54 whereas even 4 months ago i never saw that 21:35:59 and this again is on LTSC 21:36:04 the so called STABLE one 21:39:14 With the concept of "xul applets" i could create such an applet that used ietab technology to call webview2 and bring that feature in as a browser applet as opposed to a navigator applet.. 21:39:34 the webview2 21:40:52 in effect webview2 wrapped in an applet extension would allow modern web apps and general modern web browsing while navigator does what it is best at.. classic web and whatever modern shit can be added 21:41:02 again not a NEW concept but a new interpretation 21:41:15 in the face of a world where PWA is a thing that isn't Mozilla Prism 21:41:50 that is why this concept is so good.. it is a core .. that can do anything but isn't bent too far in any specific direction 21:42:00 THAT is what I will do to make my mark in XULvolution