| --- Log opened Fri Apr 24 00:00:42 2015 |
| 00:45 | | Meatyhandbag [sebastianfe@Nightstar-dk0.5eb.224.136.IP] has joined #code |
| 00:48 | | Turaiel is now known as Turaiel[Offline] |
| 00:49 | | thalass [thalass@Nightstar-h1qmno.eastlink.ca] has joined #code |
| 00:49 | | mode/#code [+o thalass] by ChanServ |
| 02:06 | | kourbou [kourbou@Nightstar-pg5.ec2.121.38.IP] has joined #code |
| 02:21 | | kourbou is now known as kourbou|willdc |
| 02:28 | | Derakon [chriswei@Nightstar-jl39jl.ca.comcast.net] has quit [Ping timeout: 121 seconds] |
| 02:30 | | Derakon [chriswei@Nightstar-jl39jl.ca.comcast.net] has joined #code |
| 02:30 | | mode/#code [+ao Derakon Derakon] by ChanServ |
| 02:31 | | kourbou|willdc [kourbou@Nightstar-pg5.ec2.121.38.IP] has quit [[NS] Quit: Bye] |
| 02:48 | | thalass [thalass@Nightstar-h1qmno.eastlink.ca] has quit [Ping timeout: 121 seconds] |
| 02:53 | | Derakon [chriswei@Nightstar-jl39jl.ca.comcast.net] has quit [Connection closed] |
| 02:54 | | Derakon [chriswei@Nightstar-jl39jl.ca.comcast.net] has joined #code |
| 02:54 | | mode/#code [+ao Derakon Derakon] by ChanServ |
| 03:13 | | Kindamoody[zZz] is now known as Kindamoody |
| 04:44 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has quit [[NS] Quit: Leaving] |
| 04:49 | | Checkmate [Z@Nightstar-484uip.cust.comxnet.dk] has quit [Ping timeout: 121 seconds] |
| 04:53 | | Kindamoody is now known as Kindamoody[zZz] |
| 04:56 | | Derakon [chriswei@Nightstar-jl39jl.ca.comcast.net] has quit [[NS] Quit: Leaving] |
| 06:32 | | Meatyhandbag [sebastianfe@Nightstar-dk0.5eb.224.136.IP] has quit [Client exited] |
| 06:48 | | celticminstrel [celticminst@Nightstar-9t3mu7.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 07:05 | | gnolam [quassel@Nightstar-2bq.qcr.248.89.IP] has joined #code |
| 07:06 | | gnolam is now known as NSGuest58858 |
| 07:07 | | NSGuest58858 is now known as gnolam_ |
| 07:07 | | gnolam_ is now known as gnolam |
| 07:07 | | mode/#code [+o gnolam] by ChanServ |
| 07:51 | | gnolam [quassel@Nightstar-2bq.qcr.248.89.IP] has quit [[NS] Quit: Boarding] |
| 09:23 | | Kindamoody[zZz] is now known as Kindamoody |
| 09:27 | | Kindamoody is now known as Kindamoody|out |
| 09:53 | | AverageJoe [evil1@Nightstar-2ofrtr.ph.cox.net] has joined #code |
| 10:34 | | AverageJoe [evil1@Nightstar-2ofrtr.ph.cox.net] has quit [[NS] Quit: Leaving] |
| 10:34 | | gnolam [lenin@Nightstar-t1tbf0.cust.bahnhof.se] has joined #code |
| 10:34 | | mode/#code [+o gnolam] by ChanServ |
| 11:23 | | Xires is now known as ^Xires |
| 12:01 | | Checkmate [Z@Nightstar-484uip.cust.comxnet.dk] has joined #code |
| 12:01 | | mode/#code [+o Checkmate] by ChanServ |
| 12:09 | | ^Xires is now known as Xires |
| 12:49 | | thalass [thalass@Nightstar-h1qmno.eastlink.ca] has joined #code |
| 12:49 | | mode/#code [+o thalass] by ChanServ |
| 13:41 | | Orthia [orthianz@Nightstar-mhk.3v7.224.119.IP] has quit [Ping timeout: 121 seconds] |
| 13:46 | | Orthia [orthianz@Nightstar-8e3.prq.224.119.IP] has joined #code |
| 13:46 | | mode/#code [+o Orthia] by ChanServ |
| 14:53 | | thalass is now known as Thalass|TF2 |
| 15:22 | | kourbou [kourbou@Nightstar-deqg8j.fbx.proxad.net] has joined #code |
| 15:30 | | Meatyhandbag [sebastianfe@Nightstar-dk0.5eb.224.136.IP] has joined #code |
| 15:45 | | Checkmate [Z@Nightstar-484uip.cust.comxnet.dk] has quit [Ping timeout: 121 seconds] |
| 16:19 | | Checkmate [Z@Nightstar-ev6.6um.94.83.IP] has joined #code |
| 16:19 | | mode/#code [+o Checkmate] by ChanServ |
| 16:27 | | celticminstrel [celticminst@Nightstar-9t3mu7.dsl.bell.ca] has joined #code |
| 16:27 | | mode/#code [+o celticminstrel] by ChanServ |
| 17:18 | <@celticminstrel> | Is make smart enough to automatically figure out header dependencies? Probably not, right? |
| 17:24 | <@Tarinaky> | iirc there's a set of compiler switches that makes gcc produce a list of header includes that can be read by make. |
| 17:24 | <@Tarinaky> | But it's pretty arcane. |
| 17:25 | | Turaiel[Offline] is now known as Turaiel |
| 17:37 | <&ToxicFrog> | celticminstrel: no, not even a little |
| 17:38 | <&ToxicFrog> | celticminstrel, Tarinaky: gcc actually has specific support for make; look for the -M and related flags. |
| 17:40 | <@celticminstrel> | Do you know if clang supports that too? |
| 18:09 | <@Tarinaky> | ToxicFrog: I said it was for gcc and make. I just couldn't remember it was -M |
| 18:26 | <&ToxicFrog> | Tarinaky: right, but I mean, it's not just "outputs dependencies", it's "specifically outputs Makefile source code with lots of tuning options so it can be used with zero postprocessing" |
| 18:26 | <&ToxicFrog> | It's not particularly arcane and is super common among those poor doomed souls who are still using Make for complicated C/++ projects |
| 18:30 | <@Tarinaky> | Make is inherently arcane. |
| 18:32 | <@Tarinaky> | Although it is a decent output format for a better build manager~ |
| 18:59 | | * kourbou puts on coding hat. |
| 18:59 | < kourbou> | Let's do this. |
| 19:06 | <@froztbyte> | celticminstrel: make does exactly what you tell it |
| 19:06 | <@froztbyte> | celticminstrel: nothing more, nothing less |
| 19:07 | <@froztbyte> | if you want dependency declaration and such, I believe the pain you want is automake |
| 19:07 | <&McMartin> | You can build it out of macros though |
| 19:07 | <&McMartin> | I forget if make macros are turing-complete. Probably. |
| 19:07 | <@froztbyte> | yeah but that's still telling it |
| 19:07 | <@froztbyte> | just half a meta level up |
| 19:09 | <@froztbyte> | (also: (opinion) I don't really know if you'd want make to do /more/ than it already does) |
| 19:10 | <&McMartin> | (I've worked with some nice systems where you defined, say, PROGRAMS=foo bar and then foo_SOURCES=a.c b.c foo_LIBS=m z and it assembled appropriate linking steps) |
| 19:10 | <&McMartin> | (And that was pretty much just a striaght application of macros) |
| 19:10 | <&McMartin> | If you don't have the PROGRAMS step then it's just patsubst for the rest |
| 19:25 | <@froztbyte> | McMartin: I think that's extruded niceness, though |
| 19:26 | <@froztbyte> | if you /start/ with the intent of making that, you're basically engaging in voluntary stockholm behaviour |
| 19:26 | <@froztbyte> | (at least as far as make goes) |
| 19:27 | <&McMartin> | It's always easier to fit the man to the machine than the other way around, that's not stockholm behaviour~ |
| 19:27 | <&McMartin> | More seriously, people throw absolute shitfits whenever "./configure && make" doesn't work and "it uses a different build system" demonstrably cuts no ice |
| 19:41 | < kourbou> | You guys know how I can do this? public void SendRaw(byte[] Data, int Offset = 0, int Length = Data.Length) |
| 19:42 | < kourbou> | of course Data does not exist. |
| 19:42 | < kourbou> | So I can't use it. |
| 19:42 | < kourbou> | Any alternatives? |
| 19:44 | < kourbou> | Lemme reformulate: Data.Length isn't available because "Data" does not exist in the context. |
| 19:45 | < abudhabi> | Make it exist? |
| 19:45 | < kourbou> | :P |
| 19:45 | < kourbou> | This is method declaration. |
| 19:45 | <@Namegduf> | default it to -1, check for -1 in the method. |
| 19:45 | < kourbou> | What do you mean? |
| 19:45 | <@Namegduf> | Or, use an overloaded method. |
| 19:45 | <@Namegduf> | Overloading would be the better way. |
| 19:45 | < kourbou> | Why not the -1? |
| 19:46 | <@Namegduf> | The -1 is a bit kludgy. |
| 19:46 | < kourbou> | So how would I go about overloading it? |
| 19:47 | < kourbou> | Use an abstract? |
| 19:47 | <@Namegduf> | public void SendRaw(byte[] Data, int Offset = 0) { SendRaw(Data, Offset, Data.Length); } |
| 19:47 | <@Namegduf> | And a second version which has no defaults at all. |
| 19:47 | < kourbou> | Oh >.> |
| 19:47 | < kourbou> | Of course. |
| 19:47 | < kourbou> | Thanks :D |
| 19:48 | <@Namegduf> | If this is C#, by the way, idiomatically your local variables should have lowercase names. |
| 19:48 | < kourbou> | Oh. |
| 19:48 | < kourbou> | :/ |
| 19:48 | < kourbou> | Okay. |
| 19:48 | < kourbou> | Only Caps for method names? |
| 19:49 | <@Namegduf> | And public properties. |
| 19:49 | < kourbou> | alright. |
| 20:03 | <@froztbyte> | McMartin: you just made me think of (and adopt) a horrible habit |
| 20:03 | | kourbou is now known as kourbou|foodz |
| 20:03 | <&McMartin> | froztbyte: Oh? |
| 20:04 | <@froztbyte> | McMartin: whenever I have a not-really-./configure project (that might end up in the hands of people used to such things), there'll be a ./configure that'll cheat things |
| 20:04 | <@froztbyte> | McMartin: and then when they wonder how this wondrous invention pips all the pythons with such ease, they'll dig |
| 20:04 | <@froztbyte> | McMartin: and learn zshian horrors. |
| 20:04 | <&McMartin> | Heh |
| 20:04 | <&McMartin> | Yeah, when you get right down to it, this is basically how dpkg works too |
| 20:05 | <&McMartin> | (And it's because not having configure/make working means that debhelper's tools require extra configuration that makes them so cranky) |
| 20:05 | <@froztbyte> | how/in which way do you mean that remark? |
| 20:05 | <&McMartin> | I mean that debian/rules is executable and its first line is #!/usr/bin/make |
| 20:06 | <@froztbyte> | well |
| 20:06 | <&McMartin> | And the tutorials for novice packagers suggest use of various tools to make that do the right thing |
| 20:06 | <@froztbyte> | it doesn't /have/ to be |
| 20:06 | <@froztbyte> | but yes, I guess it pretty much "is" |
| 20:06 | <@froztbyte> | hysterical raisins are the worst kind of joke. |
| 20:06 | <&McMartin> | Which usually forwards to the debhelper/dh library, which does all the usual configure/make/make install stuff |
| 20:06 | <@froztbyte> | ja |
| 20:07 | <@froztbyte> | I'm quite familiar with those things, fwiw |
| 20:07 | <&McMartin> | So if the project being packaged uses those, then they don't have to do anything, and this is proof of how great a packager the packager is |
| 20:07 | <&McMartin> | And anyone who does anything else is clearly maliciously insane |
| 20:07 | <@froztbyte> | at job[:-2] I helped run an inhouse fork of debian |
| 20:07 | <&McMartin> | Ah so |
| 20:07 | <@froztbyte> | more or less an "overlay' distro |
| 20:07 | <&McMartin> | Yeah |
| 20:07 | <@froztbyte> | ~130 packages |
| 20:07 | <&McMartin> | We had people here with a similar task |
| 20:07 | <@froztbyte> | so I know it well. |
| 20:07 | <&McMartin> | I actually may want to pick your brain about this at some other point |
| 20:07 | <@froztbyte> | I reaaaaally need to put the time in and just fucking finish my DD status |
| 20:08 | <&McMartin> | As there are at least two packages I've been thinking about walking that road myself |
| 20:08 | <&McMartin> | Specifically, since UQM's package was abandoned, I'd like to update it, and I've been publicly whining about gambc being preposterously out of date for literally years now |
| 20:09 | <@froztbyte> | forks to version-push, or something else? |
| 20:09 | <&McMartin> | My understanding is that Something Is Wrong with gambc but I'm not sure what it is |
| 20:09 | <@froztbyte> | oh |
| 20:09 | <@froztbyte> | right |
| 20:09 | <@froztbyte> | sure, ping me sometime |
| 20:09 | <&McMartin> | Not now, though, have stuff to do, but since you bring it up, I should at least state an intent |
| 20:09 | <@froztbyte> | weektime is kinda turdy because TZ and rarara, but weekends are good |
| 20:09 | | Thalass|TF2 [thalass@Nightstar-h1qmno.eastlink.ca] has quit [Ping timeout: 121 seconds] |
| 20:09 | <@froztbyte> | if you're okay with higher latency, mail me |
| 20:09 | <&McMartin> | (IIRC there were issues building certain architectures with gambc, and I'm not sure how dpkg-build and friends work with cross-compilation) |
| 20:10 | <&McMartin> | Yeah, I figure this is a play-it-by-ear thing |
| 20:10 | <&McMartin> | There's certainly no rush! |
| 20:10 | <@froztbyte> | more or less based on target decl in your control file |
| 20:10 | <@froztbyte> | (release etc defined there) |
| 20:11 | <@froztbyte> | naturally, it all being the One Platform That More Or Less Can Run Everywhere, it accounts for near every possibility |
| 20:11 | <@froztbyte> | so there are some deviations possible. |
| 20:11 | <@froztbyte> | (holy crap, this latency is awful) |
| 20:12 | <&McMartin> | OK. That gives me some ideas for experiments I could try, then |
| 20:12 | <@froztbyte> | haha |
| 20:13 | <@froztbyte> | also, tangent |
| 20:13 | <&McMartin> | (Basically, swipe the debian/ stuff from their ancient gambc and see where it fails when working with the most recent one) |
| 20:13 | <@froztbyte> | debhelper, debconf, and preseed |
| 20:14 | <@froztbyte> | are some fucking /deep/ magic |
| 20:14 | <@froztbyte> | like *reaaal* deep |
| 20:14 | <@froztbyte> | the preseed parsing stuff, by itself, is crazy insane |
| 20:14 | <@froztbyte> | (iirc, *all* of that is shellscript at the end of the day) |
| 20:15 | <&McMartin> | elch |
| 20:15 | <@froztbyte> | debian-installer is eldritch in sheep's clothing, dude |
| 20:15 | <@froztbyte> | srsly |
| 20:15 | <@froztbyte> | a couple years back I was trying to do some crazy crazy raid/lvm/etc layout |
| 20:15 | <&McMartin> | I have never pretended that apt and friends are some kind of gem of design~ |
| 20:15 | <@froztbyte> | partially due to a misunderstanding in what someone wanted |
| 20:16 | <@froztbyte> | I did, however, figure out how to do it |
| 20:16 | <@froztbyte> | which a) shows just how fucking potent that shit is, b) has given me a /very/ thorough appreciation of its depth and insanity |
| 20:16 | <@froztbyte> | McMartin: apt is, more or less, basically the best |
| 20:16 | <&McMartin> | That is a low bar |
| 20:17 | <@froztbyte> | chosen solver (and such) aside, I don't think that anything else comes /close/ |
| 20:17 | <@froztbyte> | at the other end of the spectrum, we have the jokes of npm |
| 20:17 | <&McMartin> | Haven't heard of that one |
| 20:17 | <@froztbyte> | pip is getting better, but can only live within "python userland" (as in, no systems package info) |
| 20:17 | <&McMartin> | Right |
| 20:17 | <@froztbyte> | cpan is pretty good. I like cpan. if I have to deal with perl, it's pretty good. |
| 20:18 | <&McMartin> | I don't conside cpan a "package system" in the same sense as apt, any more than I consider maven to be one |
| 20:18 | <@froztbyte> | full-featured, wide depth, but DANGER WILL ROBINSON DANGER if you don't know about perl things |
| 20:18 | <&McMartin> | or cabal |
| 20:18 | <@froztbyte> | McMartin: well. kinda. |
| 20:18 | <&McMartin> | Yeah, I realize I'm splitting a real fine hair here |
| 20:18 | <&McMartin> | While also conflating at least three other major things |
| 20:19 | <@froztbyte> | I find it convenient to draw a distinction between "dependency manager" and "package manager" |
| 20:19 | <@froztbyte> | mentally, that is |
| 20:19 | <@froztbyte> | some of those are a little bit more than "automated library acquirer and linker" |
| 20:20 | <&McMartin> | Right |
| 20:20 | <&McMartin> | So, you've got dependency manager, package manager, package curation (both as process and as result) and system update mechanism |
| 20:20 | <@froztbyte> | (apt does, in some cases, have some downsides. managing language libraries with it is a fairly hard thing to do at "release" level, for instance) |
| 20:20 | <@froztbyte> | McMartin: as discrete components of process? |
| 20:21 | <&McMartin> | As distinct problems to solve |
| 20:21 | <@froztbyte> | right |
| 20:21 | <@froztbyte> | yes, I could agree with that |
| 20:21 | <@froztbyte> | there is also an emergent problem from that, btw, but it's not universal |
| 20:21 | <&McMartin> | Basically all of my complaints about apt as a user are on the "package curation" side |
| 20:21 | <@froztbyte> | mmmmmm |
| 20:21 | <@froztbyte> | I guess that depends on Things |
| 20:21 | <@froztbyte> | (which distro, how well you actually know how to drive things, etc) |
| 20:21 | <&McMartin> | The closest thing to a technical solution I can imagine for the problem I usually have |
| 20:22 | <&McMartin> | "how well you actually know how to drive things" |
| 20:22 | <&McMartin> | This has a preposterous learning curve, basically |
| 20:22 | <@froztbyte> | so, for context: my home desktop is a debian sid, my home server is jessie, and all work servers are wheezy + wheezy-backports |
| 20:22 | <@froztbyte> | McMartin: it *really* doesn't |
| 20:22 | <&McMartin> | And the only fix I can think of is "there needs to be a way to identify the metapackages that you are actually supposed to install" |
| 20:22 | <@froztbyte> | I could, in 3 lines, review what you need to know |
| 20:22 | <&McMartin> | I literally used debianoids as a developer for five years before I learned that "build-essential" was a thing |
| 20:23 | <&McMartin> | Because it never came up on apt-cache search in a way that looked like a thing I should maybe be installing! |
| 20:23 | <&McMartin> | So I manually installed the 75-odd packages you need to compile C++ and complained a lot |
| 20:23 | <@froztbyte> | line 1: debian packages come in releases; packages.debian.org is the easiest way to find what you may ever care about |
| 20:23 | <&McMartin> | And probably missed a few so man was always a crapshoot |
| 20:24 | <@froztbyte> | line 2: apt-get/apt-cache/aptitude all come with '-t', which is a weirdly chosen (in my mind) toggle for "release". if you want to select a specific release of package, that's how |
| 20:26 | <@froztbyte> | line 3: there's actually a package priority system at work in the solver, which is mostly hidden from people (because simplicity). if you need to do anything "non-normal", package pinning is the way |
| 20:26 | <&McMartin> | Right, so, this is where the learning curve comes in |
| 20:26 | <&McMartin> | (a) how do you know which packages to install |
| 20:26 | <@froztbyte> | that's a very high-level skim, I guess, so feel free to feel lost and ask questions :) |
| 20:26 | <&McMartin> | (b) how do you know when you are doing something "non-normal" |
| 20:27 | <&McMartin> | (c) how do you distinguish between the two things "suggested" packages means |
| 20:27 | <@froztbyte> | expand somewhat on (a)? |
| 20:27 | <@froztbyte> | I'm not sure which way to parse it |
| 20:27 | | Thalass|TF2 [thalass@Nightstar-h1qmno.eastlink.ca] has joined #code |
| 20:27 | <&McMartin> | "It took me five years to learn about build-essential when my primary use case for debianoids was compiling C++ programs. That sounds off to me." |
| 20:27 | <@froztbyte> | also, back in 5~8 probably, need to get home |
| 20:27 | <&McMartin> | Like, to discover that that is the thing you install if you would like to compile C++ |
| 20:28 | <&McMartin> | If I had to guess, for my first Many years with Debian and Ubuntu and Mint, I kept accidentally pinning packages when I should have been installing something with a completely different name that was the metapackage that you are "supposed" to install in the first place. |
| 20:34 | <&McMartin> | And even now the experience for installing something new is generally apt-get update; apt-cache search $WHATEVER | less and then looking through five screens of output for something plausible |
| 20:35 | <&McMartin> | It's just that now I usually get the correct answer the first time after years of experience |
| 20:35 | <&McMartin> | I can't help but think that there ought to be a way to make that list both shorter and more actually correct |
| 20:36 | | kourbou|foodz is now known as kourbou |
| 20:38 | <&McMartin> | And to expand on (c), AIUI a suggested package is one that isn't strictly required for the package you named to run |
| 20:38 | <&McMartin> | However, some of those are handy optional plugins/ancillary support, and some of them are "while this is *technically* optional, basically everyone uses them" and dealing with those absent knowing there's a metapackage somewhere else tends to be like sorting out the exact set of 473 mods you need to get Oblivion to work the way you want |
| 20:39 | <&McMartin> | And whenever I've complained about this the answer is invariably "that's a feature!" and if my job is designing APT rather than curating an OS repository I'll even admit that's the right answer |
| 20:40 | <@froztbyte> | McMartin: mmmmm |
| 20:41 | <@froztbyte> | okay, so, minor bit of history: my first debian was a sarge (from a Linux Format cover disc) and I fucking /hated/ it |
| 20:41 | <@froztbyte> | because I had really shit internet, and this thing seemed to be incredibly hamstrung without it |
| 20:41 | <@froztbyte> | (without internet) |
| 20:41 | <@froztbyte> | it took about 3~4 years for me to try things again |
| 20:42 | <@froztbyte> | and I do agree; discoverability of "things" could be easier |
| 20:43 | <@froztbyte> | I'm not exactly sure how, wrt implementation. just know that my gut says "yes, it could be" |
| 20:43 | <&McMartin> | Yeah |
| 20:43 | <&McMartin> | And this is the part where I think curation matters |
| 20:43 | <@froztbyte> | aaalso I'm not an apologist for the apt/apt-cache/apt-get/aptitude/... scenario |
| 20:43 | <&McMartin> | Because I have, for instance, never seen anyone prefer yum to apt |
| 20:43 | <@froztbyte> | debian has some remarkably good build etc systems |
| 20:43 | <@froztbyte> | and I reaaaally respect just how far the folks have gotten all of that |
| 20:43 | <&McMartin> | But Fedora's repos I was able to get up to speed with wrt discoverability *much* faster |
| 20:44 | <&McMartin> | And that's what I was trying to get at with "curation" |
| 20:44 | <@froztbyte> | but goddamn. add a fucking "this is old, check out $otherthing" notice to some of it, ffs. |
| 20:44 | <@froztbyte> | McMartin: I have had the literal inverse experience, fwiw |
| 20:44 | <&McMartin> | maven/cabal/cpan/etc, which are solving a much smaller problem, are so good at this that they're practically invisible |
| 20:44 | <@froztbyte> | (and I have a good number of rhel/centos/fedora years under my belt too) |
| 20:44 | <&McMartin> | Yeah |
| 20:44 | <@froztbyte> | I find it hard to articulate that |
| 20:45 | <@froztbyte> | my general way of saying it is that everything rh* feels "just" a bit "poorer" |
| 20:45 | <&McMartin> | My usual go-to for "what the Hell dudes" in the Fedora repos is actually subversion |
| 20:45 | <&McMartin> | Which is the *dedicated subversion server* which you almost never want |
| 20:45 | <&McMartin> | Generally you wanted to install subversion-client |
| 20:45 | <@froztbyte> | package archives aren't as wide, syntax is a bit more arcane, the exact "oh, right, /that/ config" type of approach is a bit wtf, etc |
| 20:46 | <@froztbyte> | but "yum groups" are a fine little touch. |
| 20:46 | <@froztbyte> | debian does, in fact, have these things too. they just don't show as easy. |
| 20:46 | <@froztbyte> | :/ |
| 20:46 | <&McMartin> | I'm not 100% sure that "package archives aren't as wide" isn't why I got up to speed so much faster >_> |
| 20:46 | <&McMartin> | What's the debian equivalent of yum groups? |
| 20:46 | <@froztbyte> | (for the purposes of discussion, I'm basically ignoring ubuntu's existence) |
| 20:47 | <@froztbyte> | McMartin: "tasks" |
| 20:47 | <@froztbyte> | the pool isn't as wide |
| 20:47 | <&McMartin> | That sounds like it would make my life a lot easier |
| 20:47 | <@froztbyte> | just running an anaconda install (vs tasksel) will immediately tell you that, fwiw |
| 20:48 | <@froztbyte> | RH have put a lot of time into making those things a bit more visible |
| 20:48 | <@froztbyte> | (by "a lot", I mean "something like 10~15 years") |
| 20:49 | <&McMartin> | Right |
| 20:49 | <&McMartin> | OK, so, looking into tasks I see they're one level up from metapackages. |
| 20:49 | <&McMartin> | metapackages, given the things I do, are usually the things I want |
| 20:49 | <&McMartin> | Is there an apt-cache option to search only metapackages? |
| 20:51 | <@froztbyte> | this is the part where I express a sad face |
| 20:51 | <&McMartin> | OK then. |
| 20:51 | <@froztbyte> | because, yes, basically |
| 20:52 | <@froztbyte> | but it's "perl-style sigil" fuckery :/ |
| 20:52 | <@froztbyte> | like I said, I'm no apologist |
| 20:52 | <&McMartin> | (Because yeah, my user story is usually something like "I've become interested in this project over here, and now I need to get mercurial and $obscure-language installed so I can grab and build it, and it turns out those are six packages each, WHICH ONES DO I NEED" |
| 20:52 | <@froztbyte> | the system is incredibly powerful, I just wish parts of it could have better intro interfaces :( |
| 20:52 | <&McMartin> | ( |
| 20:52 | <&McMartin> | )) |
| 20:53 | <&McMartin> | Yeah, and I'm not sure you can get a good intro interface without extra metadata |
| 20:53 | <&McMartin> | And it looks like there might be places for that metadata! |
| 20:53 | <@froztbyte> | (also, only in about the last ....... couple of months, I guess, have I gained a good understanding of just how far my knowledge gap has grown) |
| 20:55 | <@froztbyte> | McMartin: at work I'm presently playing Accidental Reviewer to a lot of people who are doing Babby's First Debian Package kind of things |
| 20:55 | <@froztbyte> | McMartin: it's proving to be an educational experience, and I'm taking a lot of notes. |
| 20:55 | <&McMartin> | Also, I might have meant virtual packages, not metapackages, except that often one is the other |
| 20:56 | <@froztbyte> | (my "tutor", as it were, was a perlhead unix dude. the tooling he swore by was basically already a good cycle out of date. in the last 2 years I've had to replace that knowledge properly. and then there's the aforementioned issue of old docs without clear superceding indication) |
| 20:56 | <@froztbyte> | McMartin: so, "metapackages" are more or less what you could call a task, I gues |
| 20:56 | <@froztbyte> | guess* |
| 20:56 | <@froztbyte> | "give me a GNOME DE" |
| 20:57 | <&McMartin> | Yeah, it seems formally metapackages are packages with no files that you install for their dependencies. |
| 20:57 | <&McMartin> | They bundle up your 473 oblivion mods, etc. |
| 20:57 | <&McMartin> | virtual packages are what the Provides: fields are for |
| 20:57 | <@froztbyte> | virtual packages are either of two ways; a package can "provide" a virt, or a group of things can consist of a virt so that "apt-get install $foo-webserver" will Do The Right Thing |
| 20:58 | <@froztbyte> | ALSO: virts are how you can resolve something like "sensible-editor" in a pretty damn sane way |
| 20:58 | <&McMartin> | Yeah |
| 20:58 | <@froztbyte> | this is, however, knowledge that is buried slightly further than my liking |
| 20:58 | <&McMartin> | But I have no idea how to know what to install or searchf or |
| 20:58 | <@froztbyte> | have you seen Twisted FTW before? |
| 20:58 | <&McMartin> | nope |
| 20:58 | <@froztbyte> | (tangent explainable in moments) |
| 20:59 | <@froztbyte> | right, so |
| 20:59 | <@froztbyte> | twisted suffers from a mildly related problem |
| 20:59 | <@froztbyte> | it's a goddamn incredible framework/library |
| 20:59 | <@froztbyte> | but you gotta have some of the koolaid first |
| 20:59 | <&McMartin> | Heh |
| 20:59 | <@froztbyte> | because $reasons |
| 20:59 | <&McMartin> | Yeah, I've poked at it a little |
| 20:59 | <&McMartin> | Because concurrent reactive programming |
| 20:59 | <&McMartin> | I've walked the ivory tower, I know that risk >_> |
| 20:59 | <@froztbyte> | so, the reason I mention this |
| 21:00 | <&McMartin> | re: virtual packages, I just found https://www.debian.org/doc/packaging-manuals/virtual-package-names-list.txt |
| 21:00 | <@froztbyte> | Twisted FTW was a nice attempt at trying to show "hey, twisted's got all this cool shit, look how easy it is!" |
| 21:00 | <@froztbyte> | not sure how far it got (and that's part of the problem, I think) |
| 21:00 | <@froztbyte> | debian is /very/ close to that same problem |
| 21:00 | <@froztbyte> | the UI shouldn't be under-estimated. |
| 21:01 | <@froztbyte> | McMartin: okay so. to let that point sink in for you. I've been building packages and such (in varying capacities) since ~'09. I only saw (or needed to find it and went looking for it) that page 2 weeks ago. |
| 21:01 | <@froztbyte> | that, by itself, should also be telling :/ |
| 21:02 | <&McMartin> | That sounds a great deal like my "build-essential" story, really |
| 21:02 | <@froztbyte> | yup |
| 21:02 | <&McMartin> | So yeah |
| 21:02 | <@froztbyte> | I believe that debian could really benefit from a Twisted FTW |
| 21:02 | <&McMartin> | In the taxonomy above, this is what I was calling package curation issues |
| 21:02 | <@froztbyte> | current $work stuff has had me thinking a bit harder on how such a thing could be realized |
| 21:02 | <&McMartin> | Which means "I'm blaming the Debian Project for this state of affairs, not apt" |
| 21:03 | <@froztbyte> | haha |
| 21:03 | <@froztbyte> | McMartin: I.....mm, sec |
| 21:03 | <@froztbyte> | see pm just now |
| 21:03 | <&McMartin> | Not a joke, really. I mean it's a political/social/organization problem |
| 21:03 | <@froztbyte> | just need to find this quickly |
| 21:03 | <&McMartin> | From what I've seen of apt it either has or could be trivially and retroactively extended with the capabilities to address the issues |
| 21:04 | | Kindamoody|out is now known as Kindamoody |
| 21:04 | <&McMartin> | Also, I like how doom-engine is at the same level of officialdom as x-display-manager |
| 21:04 | <&McMartin> | I didn't get a PM |
| 21:05 | <@froztbyte> | yeah sorry, had to locate it quickly |
| 21:05 | <@froztbyte> | and censor out some stuff |
| 21:06 | <@froztbyte> | so, that's a preseed file. ~= "kickstart". from a previous employer (which I'm sure you'll quickly deduce) |
| 21:06 | <@froztbyte> | I share it for 2 reasons |
| 21:06 | <@froztbyte> | 1) to show some of the flexibility in driving the system (note the expert-recipe section, for instance), and how extensive /just the install/ can actually get |
| 21:07 | <@froztbyte> | 2) as a prime example of some fucking deep voodoo that I believe could *really* be better explained/dealt with |
| 21:11 | <&McMartin> | Yeah |
| 21:11 | | kourbou is now known as kourbou|film |
| 21:11 | <&McMartin> | Thinking about it more, a lot of my objections could be resolved with a package naming policy |
| 21:11 | <&McMartin> | But that would also require making decisions that I suspect Debian would not want to make. |
| 21:12 | <@froztbyte> | so |
| 21:12 | <@froztbyte> | that is actually a thing |
| 21:12 | <@froztbyte> | this is why you can almost-sorta guess the name of most debian-packaged cpan entries, for instance |
| 21:13 | <@froztbyte> | debian isn't so much "a set of processes" as "a set of policies" |
| 21:13 | <@froztbyte> | however. |
| 21:13 | <&McMartin> | I am imagining that for any given command line program you would have metapackages program-core and program-default (or whatever), where the first is the minimal running package and the second a metapackage that includes the core and the extensions that are part of what you'd normally install with it |
| 21:13 | <@froztbyte> | https://www.debian.org/doc/manuals/maint-guide/ is not the most approachable thing. |
| 21:13 | <&McMartin> | I've tried to read that |
| 21:13 | <&McMartin> | That's a level of abstraction below the policies I think they'd need to address the issue I have, and that's a level I don't think they want to play on |
| 21:14 | <@froztbyte> | mm |
| 21:14 | <@froztbyte> | well |
| 21:14 | <&McMartin> | It would be like dictating people's configurations, and as we all know, once you do that you might as well be the GNOME 3 project |
| 21:14 | <@froztbyte> | if, in the coming weeks, you find yourself distilling your thoughts some more |
| 21:14 | <&McMartin> | (Some bitter sarcasm here, but an acknowledgement that it is the failure mode) |
| 21:14 | <@froztbyte> | mind sending me a semi-concrete version? |
| 21:14 | <@froztbyte> | (mail or whatever) |
| 21:14 | <&McMartin> | I'll see what I can do |
| 21:14 | <@froztbyte> | failure and hate is how the internet develops |
| 21:14 | <&McMartin> | I'm not sure it's not just the equivalent of bitter n00b ranting though. |
| 21:14 | <@froztbyte> | (unfortunately) |
| 21:15 | <@froztbyte> | McMartin: I don't really think that's a bad thing. |
| 21:15 | <&McMartin> | ("But you've been at this for five years." "Like I said, bitter n00b ranting.") |
| 21:15 | <&McMartin> | Yeah |
| 21:15 | <&McMartin> | I mean, the two immediate touchpoints for me are actually in the "let's try to make the system work" part |
| 21:15 | <@froztbyte> | McMartin: if a pretty bright individual has the same gripes as someone with 3h on the system, it's way well worth heeding. |
| 21:15 | <&McMartin> | Even if I don't have the chops or time to be a full-scale DD, letting a couple packages take six years of updates at once might be worthwhile. |
| 21:16 | <@froztbyte> | ye |
| 21:16 | <@froztbyte> | I need to pick up the slack on a couple of packages; have people who'd be willing to sponsor |
| 21:16 | <@froztbyte> | after that it's probably a couple of months and then I can sign at DD level |
| 21:17 | <&McMartin> | Yeah, here's where things were on the one I cared enough to actually shout |
| 21:17 | <&McMartin> | https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=618273 |
| 21:18 | <@froztbyte> | is there an INTP for that? |
| 21:18 | <@froztbyte> | (intent to package) |
| 21:19 | <@froztbyte> | (might be ITP, I don't recall) |
| 21:19 | <&McMartin> | I see two there; when I commented lo these many years ago it had been in "pending upload" state for two and a half years |
| 21:19 | <&McMartin> | Then someone comes in and says "I'm taking it over" and is never seen again |
| 21:19 | <&McMartin> | And the package link to mentors.debian.net doesn't actually have any evidence of it anymore |
| 21:20 | <@froztbyte> | jessie and sid (presently) still have the same version |
| 21:20 | <@froztbyte> | McMartin: okay, well |
| 21:21 | <&McMartin> | I had communicated out-of-band with Mr. Thomas there, and ISTR him saying something was screwy with the ARM compiles for the latest package. |
| 21:21 | <@froztbyte> | McMartin: if you want to puppet this, I'd be fine to do an ITP (and get a friend of mine to sponsor it) if you could tell me in which direction to poke things for packaging |
| 21:21 | <&McMartin> | Heh |
| 21:21 | <@froztbyte> | I don't have the first clue what gambc is |
| 21:21 | <&McMartin> | Well, I was hoping to learn some stuff myself |
| 21:21 | <&McMartin> | It's a Scheme compiler |
| 21:21 | <&McMartin> | And a pretty good one |
| 21:21 | <@froztbyte> | sure, I could show you the whatwherehowetc |
| 21:21 | <&McMartin> | Yeah, I had to pick up the basics last month |
| 21:22 | <@froztbyte> | every time I go over the WNPP list I get depressed |
| 21:22 | <@froztbyte> | mostly because time |
| 21:22 | <@froztbyte> | (WNPP is "things that need work") |
| 21:22 | <&McMartin> | p. much |
| 21:22 | <@froztbyte> | (aka "potentially low hanging fruit for any potential new DD") |
| 21:23 | <@froztbyte> | but they're just not the right shape for my time/skillset :/ |
| 21:23 | <&McMartin> | Which is also why I need to make sure that any complaints I have are not basically "hey, what's up with you guys not having infinite resources" |
| 21:53 | | Thalass|TF2 [thalass@Nightstar-h1qmno.eastlink.ca] has quit [Ping timeout: 121 seconds] |
| 22:34 | | Turaiel is now known as Turaiel[Offline] |
| 23:00 | | kourbou|film [kourbou@Nightstar-deqg8j.fbx.proxad.net] has quit [[NS] Quit: Got to go. Bye.] |
| 23:03 | | Meatyhandbag [sebastianfe@Nightstar-dk0.5eb.224.136.IP] has quit [Client exited] |
| 23:08 | | kourbou [kourbou@Nightstar-deqg8j.fbx.proxad.net] has joined #code |
| 23:15 | | Vornicus [vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 23:15 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 23:38 | | Turaiel[Offline] is now known as Turaiel |
| 23:49 | | kourbou is now known as kourbou|zzz |
| 23:52 | | Thalass|TF2 [thalass@Nightstar-h1qmno.eastlink.ca] has joined #code |
| 23:56 | | Kindamoody is now known as Kindamoody[zZz] |
| --- Log closed Sat Apr 25 00:00:58 2015 |