Friday, 2023-08-25

*** gnucode <gnucode!~gnucode@user/jab> has joined #libre-soc00:03
*** gnucode <gnucode!~gnucode@user/jab> has quit IRC02:58
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC07:42
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc07:43
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC08:33
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc08:34
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC08:50
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc08:51
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has quit IRC09:28
*** octavius <octavius!> has joined #libre-soc09:30
lkcloctavius, you need to put "bug #xxx" not "#xxx" in order to trigger bugzilla to create an easy-to-use hyperlink.  "#xxx" will not create that hyperlink, resulting in inconvenience and pretty much wasting your time having gone to the trouble of putting the bug numbers in.09:59
lkclcan you please also put that insight into the issue tracking procedures.10:00
lkcleverything everything everything is about information management, convenience, and ease of making incremental forward progress regardless of knowledge or skillset10:01
lkcland about accelerating the iterative process of that incremental progression10:02
lkclprogrammerjake: remove that paragraph. "Because there are too many instructions to fit within this budget"10:15
lkclimplicitly it is known that we do the best progress that we can.10:15
lkclthe final wrap-up in the bugreport at the end is to either alter the description to state what was achieved or to put a comment just as it is closed "this is what was achieved"10:16
lkcl"<From here down doesn't go in MOU>"10:16
lkclremove ALL of those and replace them with the standard markdown line-break "--" that i instructed you to use10:17
lkcla "trigger-sentence" is just f*****g irritating. especially as it is undocumented, non-intuitive, and will get very tiresome to both type and see.10:20
lkcland if not typed 100% correctly will not do its job10:21
lkclplease DO NOT waste my time or yours saying "but you can just cut/paste it" - (a) that is also time wasted (b) it ignores the other irritations and time-wasting aspects of using an undocumented non-intuitive trigger-sentence10:22
lkcl"--" (<hr /> in HTML) is intuitive, short, obvious, and easy to explain to people.10:23
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has joined #libre-soc10:23
lkcloctavius, see how i put "bug #1128" here?
lkclall completed for bug #999's sub-tasks10:26
lkclnicely done sadoon10:28
lkcloctavius, again
lkcl"bug #976" not "#976"10:34
lkcl**TODO: 1) the 1st sentence does not mention binutils. 2. it needs to say "continuation of bug #976"10:34
lkcl 3. a "--" is needed. 4. the paragraph "this is an umbrella task" is unnecessary.  we already know it's an umbrella task, as it has child10:34
lkclsubtasks.  5. the last sentence which repeats for the *third* or fourth time "this is a task" can be removed. 6. again "bug #976" not "#976".**10:34
openpowerbot[irc] <sadoon[m]1> <openpowerbot> "[irc] <lkcl> nicely done sadoon" <- Thank you :)10:44
lkclit only needed 2-3 lines rather than 2-3 paragraphs but hey :)10:56
lkclthink in terms of what an EU Auditor (non-technical) would read. they would want to see less (because they're busy) and would leave it up to michiel and the team to assess the detail10:57
lkclit'll do - we have enough else to get on with.10:57
openpowerbot[irc] <sadoon[m]1> I'll try to make it a bit shorter and leave the rest of the details in the later comments11:21
*** psydroid <psydroid!~psydroid@user/psydroid> has joined #libre-soc11:27
*** psydroid2 <psydroid2!~psydroid@user/psydroid> has joined #libre-soc11:28
*** psydroid <psydroid!~psydroid@user/psydroid> has quit IRC11:29
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has quit IRC11:58
octaviusOk, lkcl I fixed bug #100312:11
openpowerbot[irc] <sadoon[m]1> octavius I want to link to the progress of running microwatt/libre-soc softcore on an fpga, which bug should I link to?12:17
octaviushang on a sec, Ill check12:22
octaviusThere are two:12:43
octaviusBug 1037 - improvements of Libre-SOC core support on FPGA boards12:43
octaviusBug 1129 - Tasks required for ls2 soc peripheral interconnect on FPGA12:43
octaviusYou should probably link to 1037, as 1129 is a dependency12:43
openpowerbot[irc] <sadoon[m]1> perfect thanks12:45
openpowerbot[irc] <sadoon[m]1> comment 0 on bug1128 is for another bug, will fix now12:49
openpowerbot[irc] <sadoon[m]1> lkcl  1128 is for VSX testing and 1132 is for qemu SFFS, sorry for the confusion12:56
openpowerbot[irc] <sadoon[m]1> All that's left now is to finish up glibc for debian and write the summary for qemu12:56
octaviusUpdated discussion page:
lkcloctavius, brilliant.
lkclthat leaves removal of the paragraphs from 1025 and 1026 and the removal and replacement of long-winded "trigger" sentences with the simple intuitive "--"13:14
octaviuslkcl, So you want these paragraphs removed?14:08
octaviusbug 1025, paragraphs commencing with "The first NLnet grant..." and "Because there are too..."14:08
octaviusbug 1026, paragraphs commencing with "Unit tests..." and "Because there are too..."14:08
octaviusAnd as for "replacement of long-winded "trigger" sentences with the simple intuitive "--""14:09
octaviusIs this something Jacob can do? I have no idea of the context14:09
markos_lkcl, ooc, why the 80char limit? is anyone reading this on a VT terminal?14:11
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC14:11
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc14:11
octavius80char limit is fine, means that you can have many terminals open side by side markos_14:11
markos_could we remove that limit? I don't see how it's still relevant on today's computers14:11
markos_I have lots of terminals open and none are 80c :)14:12
octaviusDo you want source code where lines are 300 chars long? ;)14:12
markos_I'm not wrapping code at 80chars I'll you that :)14:12
octavius"I'll you that"?14:13
markos_sorry missed the verb :)14:13
markos_it doesn't reach 300 chars wide, I guess I probably stop at around 120-15014:13
markos_well there are a couple of projects which use clang-format to auto-wrap code at 80c, but I personally dislike that14:14
openpowerbot[irc] <sadoon[m]1> Triple monitor 1000 char terminal14:14
octaviusThe problem is probably, once the limit is extended to 150, then someone will request it to be extended further still14:14
openpowerbot[irc] <sadoon[m]1> You roll in your chair to read the rest14:14
markos_I prefer the code to look as I intend it to, not how the formatter expects it14:14
octaviusWell, we're working as a project, which means there must be code formatting guidelines in place14:15
octaviusAnd I prefer to leave it as it is, there are already too many things too worry about14:15
markos_I don't disagree, I don't even mind a limit, it's just the 80 chars limit that made me curious14:16
octavius80 has been a standard for decades14:16
markos_yes and people now have 4k monitors and stuff :)14:16
octaviusand it means even someone with a potato computer could easily read our code files14:16
markos_I don't mind the limits as long as there is a reasoning, even "because I say so" is a reason14:17
markos_but yes, making sure it's readable on old computers is good enough14:18
markos_octavius, now that you mention it, some projects I've contributed to actually enforce the compliance to certain formatting using CI14:19
markos_if your code is not formatted CI will reject your patch14:19
octaviusThis second link gives some good reasoning14:20
octavius"To improve readability. Narrow code can be read quickly without having to scan your eyes from side to side."14:20
markos_I disagree, but ok14:21
markos_esp when it comes to logfiles, having a single line per entry instead of wrapping on 80 cols (which is tiny for really log entries) is much easier on the eye14:22
markos_anyway, I was just curious14:23
octaviusI'd say for source code, 80 lim is good.14:23
octaviusBut I agree that logs are a bit different14:23
octaviuswith logging it doesn't matter as much14:23
openpowerbot[irc] <sadoon[m]1> Even the microwatt patches are not working for glibc14:23
octaviusbut logs are also auto-generated (and thus don't need to be committed anyway)14:24
markos_even for code, I'm often cursing when I'm above those limits :-/14:24
markos_the problem is that with C and eg tabs=4, when you are in a deep nest of branches, you've already wasted 25% of the available space on tabs and the formatter wraps your code to a ridiculous level14:25
markos_with C++ and templates it's even worse14:25
markos_I guess it depends on the language14:26
octaviusYes indeed14:29
octaviusOne thing I noticed in Python (after Luke pointed it out), is that refactoring an if statement check, I could save myself from indenting a whole bunch of code (at the expense of an extra return statement, which some safety critical code specs wouldn't allow)14:31
*** octavius <octavius!> has quit IRC14:45
*** octavius <octavius!> has joined #libre-soc14:48
openpowerbot[irc] <sadoon[m]1>
openpowerbot[irc] <sadoon[m]1> Should I mark qemu task as resolved?14:55
openpowerbot[irc] <sadoon[m]1> glibc without altivec vsx et al is not doable atm, I think it needs a separate task because even the patches need patching it seems14:58
openpowerbot[irc] <sadoon[m]1> Here's what I'm proposing for the next tasks: patching glibc and making linux-microwatt-5.7 reproducible for our uses (optionally either backporting the patches to 6.1 or 5.4 to be on LTS)14:59
markos_sadoon, you don't *need* glibc without vsx14:59
openpowerbot[irc] <sadoon[m]1> 2- getting the damn thing running on softcore FPGA, which I can now attempt since I have two that use FOSS tools15:00
markos_you only need qemu without vsx15:00
openpowerbot[irc] <sadoon[m]1> 3- trying to add cpu feature flags to qemu for ppc to disable vsx and altivec et al15:00
openpowerbot[irc] <sadoon[m]1> Yeah markos_ I understand but qemu without vsx is being a huge pain that it is well worth it to run on FPGA first, it's actually more doable15:01
markos_well, we may need qemu in the future anyway15:02
markos_not referring to this task btw15:02
openpowerbot[irc] <sadoon[m]1> Yes, it's not canceled or anything15:02
openpowerbot[irc] <sadoon[m]1> It just needs a lot more work15:02
openpowerbot[irc] <sadoon[m]1> I'll leave this to next week's meeting so we can discuss if we should move on to FPGA stuff + linux-microwatt to finally be able to test stuff15:03
markos_most likely IBM will have to do it on their own anyway15:03
markos_qemu that is15:03
openpowerbot[irc] <sadoon[m]1> I doubt they will, there are still no commercially available v3.0 SFFS chips afaict15:04
openpowerbot[irc] <sadoon[m]1> Unless they plan to make some15:04
openpowerbot[irc] <sadoon[m]1> I'll include instructions on building vanilla glibc for now so the chroot actually works15:05
openpowerbot[irc] <sadoon[m]1> This only affects debian, gentoo already ignored my CFLAGS for glibc, sigh15:05
markos_I wonder if it would be worth using a simpler libc for this project, eg. musl15:13
markos_it's already built for ppc64le15:15
markos_and it can be used in place of glibc, with LD_PRELOAD15:15
markos_ah it's not for replacement of glibc15:16
markos_but for building binaries against musl15:17
lkclthere was some serious, *serious* problem with i think musl.15:24
lkclsomeone posted about it - the approach that they took was to duplicate entire headers from the linux kernel, rather than do the sane thing of #including them15:25
markos_for a PoC that won't be a problem15:26
markos_I'm not talking about a long term solution eg for Debian15:26
markos_but only to make sure that we get something that boots on eg. FPGA or qemu and brings us to a prompt without having to fix half of glibc15:27
lkcltoshywoshy did actually get it to work, with a backport. he's done the work already, it's just not published / public15:30
markos_even better15:30
lkcl(which is why i asked for NGI POINTER work to be made public at the time it was being written)15:30
lkcltoshywoshy said (summarising) on a call a few days ago that the "latest" version worked fine because the work by tulio upstreamed was in the version used by "latest" powerel (something like that)15:31
openpowerbot[irc] <sadoon[m]1> I am worried about the rest of the system not playing well with another libc15:32
openpowerbot[irc] <sadoon[m]1> Look at void, alpine, etc and the massive number of patches they needed15:32
lkclbut that when trying the "stable" version of powerel (which has an older version of libc6) it required back-porting of tulio's patches "nixing #ifdef POWER9 using #ifdef HAVE_VSX"15:32
lkclsadoon, it's that bad, ehn?15:33
lkclwell, at least gentoo worked. that's "good enough"15:33
openpowerbot[irc] <sadoon[m]1> Gentoo's glibc still has VSX in it, no way to know until we can test on real hardware, which I want to do asap :)15:33
openpowerbot[irc] <sadoon[m]1> It just didn't fail when I supplied my cflags15:33
openpowerbot[irc] <sadoon[m]1> which is a good sign15:34
lkcli'd actually like to put say openembedded/yocto and buildroot on the next grant's TODO list15:34
openpowerbot[irc] <sadoon[m]1> and contrast to debian, gentoo does have musl experimental support which we can try as well if need be15:34
lkclparticularly as... who was it... joel shenki already had buildroot working15:34
openpowerbot[irc] <sadoon[m]1> They have their patches done in gentoo for musl so we don't have to mess around that15:34
openpowerbot[irc] <sadoon[m]1> (for some stuf)15:34
lkclreally good idea to read the online post about the maintenance nightmare of musl, first15:35
lkclit's a quite reasonable and rational "roasting" but the responses are very revealing15:35
lkclindicating (once again) that the musl team are, sadly, in that "empathy / mirror-neuron-atrophy" trap15:36
markos_I wouldn't go there15:36
openpowerbot[irc] <sadoon[m]1> bah15:36
openpowerbot[irc] <sadoon[m]1> I'd rather we don't waste our time15:36
markos_it doesn't matter how the project operates internally, we only want something minimal that doesn't use VSX so that we get a simple prompt15:37
openpowerbot[irc] <sadoon[m]1> Ok.. generating the debian repo finally, if this works I can call this done15:37
lkclthe guy who posted had some really important technical insights, and the responses were "i fail to see why that's important" - everything you'd expect if someone lacked mirror-neurons15:37
lkclmarkos_, mmmmm good point15:37
lkclsadoon, well even if you can't that's also fine15:37
markos_lkcl, it's not like we plan to develop a product around musl15:37
markos_but if glibc becomes too much of a problem to get it to work without VSX, we have to have an alternative15:38
lkclthere's no point bashing your head against a brick wall for the same amount of money, but i do understand and appreciate that you'd prefer to reach a "completed" goal :)15:38
openpowerbot[irc] <sadoon[m]1> :)15:38
openpowerbot[irc] <sadoon[m]1> I want the next step to start with a clean objective, "ok we know what failed now let's try to fix it"15:39
lkclmarkos_, over time (as the version of libc6 in debianNNN.NN goes up to cover tulio's patches) that should disappear15:39
openpowerbot[irc] <sadoon[m]1> so btw should I add the new tasks for fpga, linux, and glibc stuff?15:39
openpowerbot[irc] <sadoon[m]1> I want to start on them asap15:39
lkcli _say_ should - as long as the ppc64/libc6 maintainers don't bloody well introduce more "#ifdef POWER9/POWER10" patches, sigh15:39
markos_lkcl, that's great, if our timeplans overlap :)15:39
lkclsadoon: new tasks? que? oh, you mean "future plans?"15:40
openpowerbot[irc] <sadoon[m]1> yes15:40
openpowerbot[irc] <sadoon[m]1> For the next funding and to start working on soon15:40
lkclmmm yyyeees... probably a good idea, *in brief*, and link them as "depends on" the relevant ones15:40
markos_#ifdef POWER10 // I'll just use some MMA code15:41
lkcland shove them under the "Future" milestone15:41
openpowerbot[irc] <sadoon[m]1> alright sounds good15:41
lkclmarkos_, sigh15:41
lkclbut don't spend too long on them - make them real brief placeholders15:41
lkcloh i updated the wiki page on SFFS/qemu
openpowerbot[irc] <sadoon[m]1> alright15:42
lkclif you're good to close it then i am too15:42
* lkcl afk15:42
openpowerbot[irc] <sadoon[m]1> alright, great15:43
openpowerbot[irc] <sadoon[m]1> debootstrap fails at some point probably missing packages, but the chroot works fine, yay15:43
openpowerbot[irc] <sadoon[m]1> vim not building, neither is udev, I'll add mc and nano to the chroot so that we can have editors at least15:44
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC15:51
lkclgood idea15:58
lkclbtw can youuuuu.... drop the rootfs-tarball onto your server just like last time, i'll keep updating it15:58
lkcli actually want to try it out15:59
openpowerbot[irc] <sadoon[m]1> One step ahead of you ;)15:59
lkcl(under qemu-ppc64le-user)15:59
lkcl(which i thiiink i already have working....)15:59
openpowerbot[irc] <sadoon[m]1> deboostrap luckily has an option to make a tarball automatically15:59
openpowerbot[irc] <sadoon[m]1> apt will not work for now, stupid udev15:59
openpowerbot[irc] <sadoon[m]1> I can add vanilla udev if you want apt to work on non sffs machines16:00
lkcli'm used to embedded systems and using /sbin/MAKEDEV :)16:05
lkcli can handle it myself but others might not be able to.16:05
openpowerbot[irc] <sadoon[m]1> but apt isn't :D16:05
lkclplease tell me the fucking virus known as "potterton's insanity" hasn't also taken over apt16:06
openpowerbot[irc] <sadoon[m]1> it requires udev if that's what you're asking16:06
lkclfucking hell16:06
lkclthis has to stop16:06
lkclthat makes apt *critically dependent* on the fucking systemd package16:07
openpowerbot[irc] <sadoon[m]1> we can use devuan's apt16:07
openpowerbot[irc] <sadoon[m]1> if it comes down to that and systemd is absolutely failing to build whatsoever16:07
lkclgive me a second to get the source code, i usually build debian packages with debian/control and debian/rules *REMOVING* systemd dependency16:09
openpowerbot[irc] <sadoon[m]1> if it's that easy then great16:10
openpowerbot[irc] <sadoon[m]1> it should be fine16:10
lkclbut now do you see why i said "use debian/10"?16:11
openpowerbot[irc] <sadoon[m]1> debian 10 also has the same issues I assume16:12
lkclno it won't because the "infection" had not propagated, there16:12
lkcllet me check - apt 2.7.3 depends on                libsystemd-dev [linux-any],16:13
lkcl               libudev-dev [linux-any],16:13
* lkcl checking 1.8.316:14
lkclno you're right, it does indeed depend on libudev-dev and libsystemd-dev16:15
openpowerbot[irc] <sadoon[m]1> we'll use the devuan apt source package16:15
openpowerbot[irc] <sadoon[m]1> easy stuff16:15
openpowerbot[irc] <sadoon[m]1> (for another day)16:15
lkclfascinating - there's no *actual* dependency listed16:16
openpowerbot[irc] <sadoon[m]1> why systemd fails to build without VSX is beyond me, but not surprising16:16
lkcli am not in the LEAST bit interested. at all.16:16
openpowerbot[irc] <sadoon[m]1> apt needs something16:16
lkcli never run with systemd (another reason why i said use debian/10 because less of debian/10 has been "infected")16:17
openpowerbot[irc] <sadoon[m]1> if we do a full system devuan might actually be a viable option thanks to no systemd and the fact that there's no systemd means less precious ram eaten16:18
lkclsad to say we may have to abandon debian as it's just too infected at this point16:18
openpowerbot[irc] <sadoon[m]1> (cue joke about rams being eaten)16:18
lkcland devuan is not "inclusive to all init systems" because it's "anti-systemd"16:18
openpowerbot[irc] <sadoon[m]1> Better than nothing though16:18
lkclexcept for when you want to use debian/testing packages to get at least some way towards a "rolling-release" distro16:19
openpowerbot[irc] <sadoon[m]1> They do that though no?16:19
lkcldevuan is so small they can't possibly put the time/effort into *also* covering debian/testing16:20
lkclno they do not16:20
openpowerbot[irc] <sadoon[m]1> there's devuan testing afaict16:20
openpowerbot[irc] <sadoon[m]1> ah16:20
openpowerbot[irc] <sadoon[m]1> debian testing in a chroot :)16:20
lkclor at least, they weren't in any way able to keep up when i last spoke to them16:20
lkclfound that they were "anti-systemd" in direct violation of their stated goal, weren't interested in fixing that, at which point i backed away fast16:21
openpowerbot[irc] <programmerjake> > [irc] <lkcl> "--" (<hr /> in HTML) is intuitive, short, obvious, and easy to explain to people.16:24
openpowerbot[irc] <programmerjake> it is totally non-obvious that it means "the rest of this comment doesn't go in the MOU", it just looks like random formatting that means nothing of the sort16:24
octaviusIt's the first time I've seen '--' being used like that, but I also don't care16:25
openpowerbot[irc] <programmerjake> imho we should use something that's more self-documenting than "----" but less error-prone than the whole sentence16:25
openpowerbot[irc] <programmerjake> how about </MOU>16:25
octaviusWe just need to get the final changes done to submit the MoU16:25
octaviusthat can be thought about later16:26
openpowerbot[irc] <programmerjake> since anyone who knows any xml/html knows what an end-tag looks like16:26
octaviusAre you suggesting having <MOU>text text</MOU>?16:27
openpowerbot[irc] <programmerjake> technically, changing to "----" isn't a requirement to generate the correct MOU since the script also triggers off that trigger phrase16:27
openpowerbot[irc] <sadoon[m]1> lkcl16:27
openpowerbot[irc] <sadoon[m]1> nano works, mc is giving me trouble, ignore it16:28
openpowerbot[irc] <programmerjake> no, i'm suggesting: blah, blah, text in MOU16:28
openpowerbot[irc] <programmerjake> </MOU>16:28
openpowerbot[irc] <programmerjake> text not in MOU16:28
openpowerbot[irc] <sadoon[m]1> I'll work on improving this a lot once we bootstrap an fpga16:28
*** octavius <octavius!> has quit IRC16:31
*** octavius <octavius!> has joined #libre-soc16:31
openpowerbot[irc] <programmerjake> lkcl:;a=commitdiff;h=cec00456818c6437de30cecc8c2bb0746a6cfb3c16:32
openpowerbot[irc] <programmerjake> we still need to look for blank lines before/after the "---" line otherwise we'd trigger on ascii-art16:32
openpowerbot[irc] <programmerjake> e.g. we'd trigger on:16:33
openpowerbot[irc] <programmerjake> ---16:33
openpowerbot[irc] <programmerjake> | |16:33
openpowerbot[irc] <programmerjake> ---16:33
openpowerbot[irc] <programmerjake> so i want to partially revert that commit16:33
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc16:38
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has joined #libre-soc16:38
openpowerbot[irc] <programmerjake> \    this ascii-art wire also incorrectly triggers it16:42
openpowerbot[irc] <programmerjake> ---16:42
openpowerbot[irc] <programmerjake> \16:42
openpowerbot[irc] <sadoon[m]1> I marked 999 and all subtasks as done, woo16:46
octaviusNice one sadoon :)16:47
openpowerbot[irc] <sadoon[m]1> thanks!16:47
markos_lkcl, systemd preferences aside, we will *have* to move debian 12 soon17:19
markos_I neither like or dislike systemd, it's just a tool, I can do the same things with systemd that I did with init scripts before17:20
markos_some things are easier others are a pita17:20
markos_s/move debian 12/move *to* debian 12/17:20
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has quit IRC17:22
markos_programmerjake, </MOU> seems too XML-ish but missing the <MOU>, if you need to use just one tag, just add a combination, eg  ---MOU---17:24
markos_or ---/MOU--- rather17:24
markos_or !MOU, etc17:24
markos_you get the point17:24
markos_but definitely not </MOU>17:25
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has joined #libre-soc17:25
openpowerbot[irc] <programmerjake> hmm, how about <end MOU>17:26
markos_why do you need the <> at all?17:29
markos_eg. for text files it's common to use separator of the form ---TEXT--- all the time17:29
markos_with less or more dashes, that's your choice17:30
openpowerbot[irc] <programmerjake> ok, with dashes sounds fine, but not just `end MOU` with nothing but the words17:30
openpowerbot[irc] <programmerjake> lkcl: what do you think?17:31
openpowerbot[irc] <programmerjake> since `--` by itself is totally non-obvious since it's not self-documenting17:32
*** tplaten <tplaten!~tplaten@> has joined #libre-soc17:49
tplatencontinue working on gram, where I left a month ago.17:50
*** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has quit IRC17:57
*** sauce <sauce!> has joined #libre-soc18:13
lkclsadoon: when you create bugs don't set "comment editing" or change any defaults that could make bugs anything other than private. don't check *any* tickboxes in other words18:14
lkclprogrammerjake: no. leave it. move on.18:15
lkcldo not waste further time thinking about or discussing it.18:15
openpowerbot[irc] <programmerjake> well, I don't like leaving a trap for noobs when it's easy to fix18:16
lkclplease read what i just wrote, respect what i have written, and do not waste further time.18:19
markos_I think Jacob is right in this one and this is trivial enough to fix, I think ---end MOU--- is good enough and unlikely to trigger anything18:19
openpowerbot[irc] <programmerjake> this violates the principle of least surprise18:19
lkclthe task that has *zero budget* is good enough18:19
lkclplease drop the discussion18:19
openpowerbot[irc] <programmerjake> ok, but this will come back to bite us...18:20
openpowerbot[irc] <programmerjake> sadoon: checking the "comment editing" box means you're making the bug only visible to those with the privilege to edit comments, which is not desired18:22
markos_this is the sort of micro-management that I dislike, it's not like we're designing a full blown protocol here, tbh, I wouldn't even ask what to choose in this case and just pick something sensible18:22
openpowerbot[irc] <programmerjake> hence why luke said to not check it18:22
octaviusGuys, can we save this for another time? Let's just finish what's needed for the MoU already18:23
octaviusbugs 1025 and 102618:23
lkclprogrammerjake: not important. far higher priority tasks to focus on.  please don't keep putting roadblocks and trying to repeat what you have already repeated and i have already stated clearly is not important18:24
lkcloctavius, precisely. it's not important, we have to *get the damn MoU in*. screw future users, right now i'm under financial pressure, ok?18:25
lkcli could care less about future users18:25
openpowerbot[irc] <programmerjake> 1025/6 are now ready for the MOU18:26
octaviuslkcl, you haven't answered my questions on:
openpowerbot[irc] <programmerjake> oh, i forgot the paragraph you wanted removed, sec...18:27
octaviusBut don't they need to be replaced with something else? Won't there be a blank comment 0 problem18:28
openpowerbot[irc] <programmerjake> no, because we're just removing the paragraph about too many insns for budget, everything else stays18:29
octaviusAh ok18:29
octaviusthanks for clarifying18:29
openpowerbot[irc] <programmerjake> removed that paragraph from 961 and 1025/6, those should be ready for RFP now18:30
openpowerbot[irc] <programmerjake> luke, just found why we *have* to not have the trigger be just `---`: 961 uses that to separate subtasks18:32
openpowerbot[irc] <programmerjake> so, can i remove `---` from the script and add `---end MOU---`?18:33
octaviusI just noticed that in bug #961 as well. Does it impede current RFP?18:33
octaviusIf we can get away with it, make sure we're ready for submission first18:34
openpowerbot[irc] <programmerjake> if `---` is the trigger than 961's text will end up as just "Schedule A, based on" and nothing else18:34
openpowerbot[irc] <programmerjake> yes, it impedes the RFP18:34
octaviusAdded '+' on either end for now? Separate by +------+ etc.18:35
octaviusThen ----------- at the bottom18:35
openpowerbot[irc] <programmerjake> it takes 1 min to change the script, 961 is not the only top-level bug with --- formatting18:36
openpowerbot[irc] <programmerjake> i just need the go-ahead from luke18:37
octaviusOh I see. Then lckl, it *will be faster* to add a terminator `---end MOU---` than it would to change every top-level bug18:37
octaviusBut how many bugs will we need to change for the `---end MOU---` format programmerjake?18:38
openpowerbot[irc] <programmerjake> 2, 1025/618:38
openpowerbot[irc] <programmerjake> luke, your commit doesn't even work, it uses \ instead of :18:40
openpowerbot[irc] <programmerjake> i'm reverting it18:41
openpowerbot[irc] <programmerjake> lemme know and i'll push with it changed to `---end MOU---`18:42
openpowerbot[irc] <programmerjake> ok, I have the json and it looks fine, lkcl I'm waiting on you18:50
octaviusI need to go chaps, speak to you later18:52
openpowerbot[irc] <programmerjake> all i need to do is git push and upload the JSON...18:52
openpowerbot[irc] <programmerjake> bye octavius18:52
*** octavius <octavius!> has quit IRC18:53
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has quit IRC18:53
*** ghostmansd[m] <ghostmansd[m]!~ghostmans@> has joined #libre-soc18:54
openpowerbot[irc] <programmerjake> I already changed 1025/618:54
openpowerbot[irc] <programmerjake> I'll just go ahead and upload the JSON then...19:04
openpowerbot[irc] <programmerjake> uploaded json and pushed to the fix-mou-splitting branch, now you don't need to wait on me (unless more task descriptions need to be changed?)19:08
openpowerbot[irc] <programmerjake> note that the Schedule A is changed, removing the too many insns for budget paragraph, so lkcl you'll need to re-copy that to the email or just link it19:09
openpowerbot[irc] <sadoon[m]1> <openpowerbot> "[irc] <lkcl> sadoon: when you..." <- Oh my bad, thought that allowed others to edit comments19:41
openpowerbot[irc] <programmerjake> yeah, it's misleading, but too annoying to figure out how to change19:41
lkclsadoon: i was caught out by that as well when i started. i had to hand-patch a *25 year* old version of bugzilla with a "comment editing" patch that someone wrote *15* years ago19:46
lkclprogrammerjake: no, i won't. part of the responsibility of writing the MoU (Schedule A) is to duplicate/simultaneously-update the Schedule A if the comment#0s are changed19:47
lkcland i am not doing day-to-day admin any more, that's down to you (all)19:48
openpowerbot[irc] <programmerjake> I did simultaneously update it...I'm just saying it's different now so if you're submitting the email to nlnet you need the correct contents19:48
lkclbottom line: please don't attempt to assign tasks to me or make *any* assumption that i *will* do anything - at all19:48
lkclahh ok19:49
openpowerbot[irc] <programmerjake> ok, then should I email nlnet?19:49
lkclthe email(s) i send are simply to notify NLnet "here's the bugtracker comments, here's the attachment again in the bugtracker"19:49
openpowerbot[irc] <programmerjake> ok, sounds good!19:49
lkclno because i haven't reviewed it, and it's my responsibility as project leader under whose name the MoU is signed, to do that check19:50
openpowerbot[irc] <programmerjake> k19:51
lkclprogrammerjake: *please follow my instructions*.19:59
lkcli did NOT authorize the use of anything other than "-----" for task ending20:00
openpowerbot[irc] <programmerjake> but ---- is broken because the main mou tasks use ---- everywhere20:02
openpowerbot[irc] <programmerjake> the json is correct, can we just submit it and figure out the rest later?20:02
lkclthe top-level MoU Schedule A is *not* included in the RFPs20:04
openpowerbot[irc] <programmerjake> k20:04
lkcland even if it was included, the use of "---" is fantastic because it subdivides the "description of the top-level milestone" from the Schedule A (an analog of the same use for milestones/subtasks)20:05
openpowerbot[irc] <programmerjake> in any case, the json is correct and we can change to ----- later20:06
lkclno: please *re-run* it.20:06
lkcli have deleted the branch that you created.20:06
openpowerbot[irc] <programmerjake> did you change something on bugzilla that needs rerunning?20:06
lkclthis will teach you to do what is asked.20:07
lkclSchedule A, based on
lkclthat's perfect for exactly what should go into the JSON file20:07
lkcljust that first line20:07
lkclif you had listened instead of wasting my time by thinking that you were right, i would not be so angry20:08
lkcland this task would be completed20:08
openpowerbot[irc] <programmerjake> k, did something change that ends up in the json?20:08
lkclRERUN it and check FOR YOURSELF20:09
lkclthat is your responsibility to check, given that you have screwed up.20:09
lkclthis is in the Charter20:09
openpowerbot[irc] <programmerjake> i'm asking if you changed anything, that should be trivial to answer20:09
lkclif you screw up, *you* own the screw-up and *you* get to fix it20:09
lkcli am going out now20:10
lkcli am not well20:10
lkclyou are taking up my personal time doing tasks that make me healthy20:10
lkcltherefore dealing with this is MAKING ME ILL20:10
lkclplease STOP.20:10
lkcland TAKE RESPONSIBILITY for your screw-up from NOT LISTENING20:10
lkclhave you got the message yet?20:11
lkclthis is terribly frustrating for me, makes me really angry and i LITERALLY cannot afford to do that, it has severe damaging health consequences20:11
openpowerbot[irc] <programmerjake> i'm listening. i'm asking you if you changed anything because that means we don't need a new json file20:11
lkcli am going out now: please do what i have asked.20:11
lkclyou're still not getting it20:12
lkcli am not going to answer20:12
lkclRIGHT NOW20:12
*** lkcl <lkcl!> has quit IRC20:12
openpowerbot[irc] <programmerjake> i'm working on that...20:12
*** gnucode <gnucode!~gnucode@user/jab> has joined #libre-soc20:15
openpowerbot[irc] <programmerjake> just wasted time fixing the code and running it, only to discover you changed it too, it'd be nice if you told me next time20:30
openpowerbot[irc] <programmerjake> made json20:36
openpowerbot[irc] <programmerjake> lkcl ^20:37
*** tplaten <tplaten!~tplaten@> has quit IRC21:22
*** psydroid2 <psydroid2!~psydroid@user/psydroid> has quit IRC23:04

Generated by 2.17.1 by Marius Gedminas - find it at!