**** BEGIN LOGGING AT Thu Aug 01 02:59:58 2013 Aug 01 04:52:58 sgw_: i was thinking base-files, but it doesn't seem to create it on its own Aug 01 04:53:52 so would recipe be preferred over base-files.bbappend? Aug 01 06:53:38 morning Aug 01 06:53:41 is Beth around? Aug 01 08:02:24 morning all Aug 01 08:34:14 abelloni: about network config for sato, you should have a network icon at the top that will configure basic stuff (its a UI to connman) Aug 01 08:59:38 rburton: I didn't see anything Aug 01 08:59:52 is connman part of core-image-sato ? Aug 01 09:00:37 yes Aug 01 09:00:43 there's a network icon in the titlebar thingy Aug 01 09:00:47 well, should be! Aug 01 09:03:56 rburton: you once said sato could maybe fit on qvga screens with a bit of work? Aug 01 09:05:25 ant_work: depends how much you want to use Aug 01 09:05:32 ant_work: the desktop and panel should scale fine Aug 01 09:05:47 just don't expect other gtk+ widgets to work on qvga Aug 01 09:06:04 there's an ancient patch against gtk2 to make a qvga-friendly file selector Aug 01 09:06:12 incredibly bitrotted though Aug 01 09:06:28 but if you just want to use the panel or desktop, changing the font and icon size should do enough Aug 01 09:08:07 tbh my last try was done long ago...back then the desktop did not fit on 320 Aug 01 09:08:38 well, exactly I've rotated qvga so th etopbar is rendered in 240pix Aug 01 09:09:28 ant_work: fwiw, i'm porting the desktop and panel to gtk3 right now Aug 01 09:09:44 ant_work: so if you have any feature requests tell me now while the code is still changing :) Aug 01 09:09:58 (the panel's positioning code was just simplified massively) Aug 01 09:17:41 I'll send you couple of screenshots Aug 01 09:25:55 hi yocto. How to clean all packages compiled even downloads and start a fresh build? Aug 01 09:35:26 hello guys Aug 01 09:35:53 is there any plan to replace qt4 with qt5 in yocto layer for 1.5 release Aug 01 09:46:01 Noor: not for 1.5, no Aug 01 09:46:19 rburton: thanks Aug 01 09:46:24 Noor: feel free to start that debate when 1.5 freeze for 1.6 though :) Aug 01 09:46:33 so we have to use meta-qt5 layer then Aug 01 09:46:37 yes Aug 01 09:47:01 that's what we'd be pulling directly in anyway Aug 01 09:47:06 hooray layer model :) Aug 01 09:47:11 rburton: is there any link which describe what is planned for the release Aug 01 09:48:08 https://wiki.yoctoproject.org/wiki/Yocto_1.5_Schedule Aug 01 09:48:23 target features are all in bugzilla Aug 01 09:50:47 rburton: thanks it helps Aug 01 09:50:53 hello everybody. i got a little problem with gnutls and libtasn on branch master... Aug 01 09:51:25 gnutls provides libgnutls1.so.6 but libtasn needs libgnutls1.so.3 Aug 01 09:52:11 creating a softlinks lets it compile, but the build fails at do_rootfs() because bitbake realises no one provides libgnutls1.so.3 Aug 01 09:52:22 dRp3pP3r: just rebuilding tasn should solve that Aug 01 09:52:58 the question is why didn't it rebuild itself :( Aug 01 09:53:23 dRp3pP3r: so try cleaning and rebuilding tasn Aug 01 09:53:37 on it. thanks in advance. :) Aug 01 09:58:46 ok, how exactly do i do that? (whats the name of the package?) refetching and compiling gnutls didn't work, even though the recipes are in the same directory... Aug 01 09:59:14 got it, libtasn1, not libtasn Aug 01 10:05:26 well, didn't work. gnutls compiles, as always, but compiling wpa_supplicant fails due to missing dependencies of libgnutls (libtasn1.s0.3) Aug 01 10:10:33 the exact error is: .../ld: warning: libtasn1.so.3, nedded by ..../yocto/build/tmp/sysroots/imx6qsabrelite/usr/lib/libgnutls.so, not found Aug 01 10:11:23 ah the dep was the wrong way around Aug 01 10:11:28 rebuild gnutls Aug 01 10:11:44 have you been switching versions up and down? Aug 01 10:11:52 that can cause issues like this if you're not careful Aug 01 10:12:38 not up and down, only up :) Aug 01 10:12:57 gnutls rebuilded when i refetched and builded libtasn Aug 01 10:20:30 seems to have worked when i cleand gnutls, libtasn and wpa-supplicant and rebuilt. thanks. :) Aug 01 10:27:29 dRp3pP3r_: hopefully it doesn't happen again. linkage gets noticed and will cause re-builds, so something *odd* happened Aug 01 10:28:45 hm. can't really explain that one to me, because the package that was updated was wpa-supplicant. the linkage between gnutls and libtasn1 should not have been touched... Aug 01 10:29:20 lets just pretend nothing happened ;) Aug 01 10:31:13 happened? did i miss sonething? ;) btw: got the next one with tslib (no one provides libts >= 1.0+gitr0+) and cleaning won't work. can one compare version numbers including a git-hash using relative operators? Aug 01 10:31:57 yes Aug 01 10:32:03 the gitr0 will be gitr1 on the next build Aug 01 10:32:34 though, what platform are you using that is using tslib? there was discussion here yesterday about what is remaining that still uses it. Aug 01 10:32:56 ah. than that's odd. tslib-calibrate wants libts, which is provided by libts. both in the same version. Aug 01 10:33:27 my platform is a Freescale i.Mx6Quad on a BoundaryDevices Sabrelite Eva-Board Aug 01 10:33:37 its at times like this that i blow away my deploy directory and relevant stamps Aug 01 10:33:42 qt on framebuffer? Aug 01 10:33:50 or what ui stuff Aug 01 10:34:25 qt 4.8.x on a framebuffer, yes. matchbox-desktop on X-Server... Aug 01 10:37:19 doesn't matter, kicked the touchscreen out of the image, don't have one here right now. problem for future-dRp3pP3r :) Aug 01 10:38:27 ha Aug 01 11:34:43 rvurton: correction. shows up again. Aug 01 11:34:52 im meant rburton Aug 01 11:34:59 -.- Aug 01 11:35:31 i mean libtasn Aug 01 11:41:18 i'll fix it and come back with the solution, but for now, i need to change the line... Aug 01 11:41:20 zombie in your kitchen? Aug 01 14:31:27 is there a convenient way of installing RPMs on my target device? I'm trying to get build-essentials installed (on Gumstix Overo), but chasing down each needed RPM with all its dependencies is hopeless Aug 01 14:36:16 ErkiS: packagegroup-core-buildessential should contain what you want Aug 01 14:36:31 well, s/contain/pull in/ Aug 01 14:38:17 bluelightning, I should bitbake packagegroup-core-buildessential? Aug 01 14:38:34 getting "ERROR: Nothing PROVIDES 'packagegroup-core-buildessential' " Aug 01 14:39:22 ErkiS: you can, or you can just add it to IMAGE_INSTALL if you want it in the image by default Aug 01 18:55:39 * kergoth mutters Aug 01 19:11:56 i pretty much do that every day... Aug 01 19:12:30 if you get close enough to hear, i kinda sound like popeye... Aug 01 19:15:47 hehe Aug 01 19:16:00 one of those frustrating days.. Aug 01 19:20:36 Oh, kergoth. I may be about to send a flurry of suggested patches to meta-sourcery. Where "about-to" means a month from now I'll be saying "haha no I haven't got that working yet". :) Aug 01 19:21:42 heh :) Aug 01 19:22:02 what will they do? Aug 01 19:22:31 Well, long story there! We're now getting CS toolchains which also have a layer that provides bbappends to build "normally" (for Yocto) using the CS patches. Aug 01 19:22:55 Our plan is to drop all the local stuff we had for supporting partial rebuilds, and messing around with binutils on targets, and just use that layer for all those things. Aug 01 19:23:31 But, there is a cleanup pass I'd done to external-csl-toolchain once, which I'd like to recreate, which did a better job of separating out "copy the multilib files we care about somewhere useful" from "install these files and package them". Aug 01 19:24:09 Which the leads to support for an alternative distribution model, where instead of a completely unpacked libc directory, we have a selection of cpio.gz archives which correspond to multilibs. Aug 01 19:24:48 For MIPS, the libc directory for our unweildy selection of multilibs is about 1.8GB as-shipped, and bloats to over 5GB if it goes through a git server (breaking all the hard links). Aug 01 19:25:59 In our distribution, we're distributing about 750MB of compressed archives, with the useful trait that any *given* install only has to decompress about 100MB of that, and it ends up being faster in total than unpacking the original archive would have been. Aug 01 19:26:43 So basically, it checks for actual unpacked files, and does the same thing external-csl-toolchain did (roughly) if it finds those, but also supports a faster/smaller unpacking mechanism when appropriate. Aug 01 19:27:02 … unwieldy. I can spell, really. Aug 01 19:28:28 I also had stuff to make the toolchain wrappers (which exist separately) become enabled mostly-automatically with binary toolchains, which simplifies life for packages which end up failing to specify the right CCARGS, although I think all the known cases got fixed, so I might just drop that. Aug 01 19:30:29 * mr_science adds USE="aspell" to seebs irc client Aug 01 19:30:43 I can't use spell checkers, they drive me batty. Aug 01 19:31:05 Their failure rate is usually higher than mine, at least for what I write. Aug 01 19:31:45 i wasn't suggesting anything like M$ autocorrect Aug 01 19:32:05 just some red squiggly highlights... Aug 01 19:32:24 ha Aug 01 19:33:31 oh, is that whta those are for ;-) Aug 01 19:33:40 s/whta/what/ Aug 01 19:34:54 I hate the red squiggles. They're … not very reliable. Also, I occasionally do real proofreading, and apparently most proofreaders find that if you get in the habit of having those on, you stop being able to proofread as reliably. Aug 01 19:35:17 I will occasionally run a spell checker over a document, but I don't like having it always-on, because they really aren't very good at it. Aug 01 19:37:50 only as good as their dictionaries... Aug 01 19:38:47 also kinda humorous how firefox defaults to crazy language that's completely unrelated to my locale settings Aug 01 19:38:53 *some Aug 01 19:39:06 currently cuban spanish... Aug 01 19:53:05 huh, kind of surprised 'tac' isn't posix/sus Aug 01 20:06:43 I think it may be because tac is obviously not possible to generically implement for stdin. Aug 01 20:13:10 true, can't treat everything as a file if that's the operation you want to do Aug 01 20:13:11 good point Aug 01 20:13:21 still, quite handy :) Aug 01 20:16:45 I wrote an unsort(1) utility. Aug 01 20:17:19 I have, on my very low priority list of things to do: making it correctly parse, and warn about errors in the specification of, all the sorting options available to the POSIX sort(1). Aug 01 20:18:33 seebs: stop, you're making me all goose-pimply... Aug 01 20:18:55 Obviously, implementing the functionality would be much, much, easier than parsing the options. Aug 01 20:19:14 Option parsing is one of those problems which sounds REALLY easy until you do it. Aug 01 20:19:49 … argh Aug 01 20:20:12 I was thinking about that, and now I've realized that there is an obvious way in which I could make my option parser more useful in an interesting way. Aug 01 20:20:43 It supports getopt-like strings, with the extension that, just as "f:" means "option f takes an argument", "f#" means "option f takes a numeric argument". Good so far. Aug 01 20:21:23 if it's anything above "trivial" i tend to use getopts Aug 01 20:21:26 I have now realized that I might find life easier if it also supported "f::" meaning "option f takes exactly two arguments". Aug 01 20:21:55 In sh, I use getopts. Or, in some cases, a very clever portable near-reimplementation of it in sh I got from Gary Vaughan during the work on the shell book. Aug 01 20:22:17 but i'll admit to writing a lot of shell script with custom check_input functions and no getopts... Aug 01 20:22:35 But being a sort of compulsive programmer, and playing a game that has a Lua-based addon development environment, I also maintain a getopts-alike for parsing commands in that. Aug 01 20:22:41 * mr_science should take his own advice more often Aug 01 20:22:43 And there I feel some freedom to experiment. Aug 01 20:23:03 I actually maintain an options library in C, although most of its features turned out to be ill-considered. But it was an interesting idea. Aug 01 20:23:45 but i'll bet you learned something from your "ill-considered" implementation... Aug 01 20:23:47 In particular, it has the fancy feature of letting you save options to a .foorc file, or load them from such a file, so you can trivially set default options. This seemed like SUCH a good idea at the time. I think I've used it twice. Aug 01 20:24:15 Oh, heck yeah. This is why I am always happy to make an excuse to go write something. I know people who write prose, and they insist that no words you write are a waste. I have found this to be true of code, also. Aug 01 20:25:09 the main difference being the concept of throwing out implementation #1 Aug 01 20:25:49 since *not* throwing out rough drafts of prose is much less likely to bite you in the ass later Aug 01 20:27:05 Someone, and I forget who, but… Might have been Joel Spolsky? Argued recently, and moderately persuasively, that sometimes refactoring may be better than throwing away the first implementation. Aug 01 20:27:16 Because some significant amount of the "cruft" is actually handling real problems. Aug 01 20:27:23 http://www.joelonsoftware.com/articles/fog0000000069.html <-- yup Aug 01 20:27:43 i wouldn't disagree with that Aug 01 20:27:47 much... Aug 01 20:28:15 probably depends a lot on the individual person/code Aug 01 20:29:27 It does. Also on scale, timing, scheduling, and how far off the original requirements were. Aug 01 20:29:39 many variables... Aug 01 20:29:53 In general, my experience has been that all the clear rules as to what to do are wrong, and you are better off having multiple factors and weighing them. Aug 01 20:30:12 This is a good strategy especially because it is easy to explain how to do it so novices can start using it right away. *nods wisely* Aug 01 20:34:33 seebs: its a good article, he makes a good point about the implicit knowledge carried in the code. i know i've attempted rewrites on things before and ended up re-rolling the workarounds for the ocrner cases again which i didn't pull over at first thinking they weren't necessary Aug 01 20:37:44 This is highly relevant to the task I've just started on of trying to figure out whether to (1) uprev my wr-toolchain layer, or (2) try to get the fixes in it that I actually still need into meta-sourcery. Aug 01 20:37:47 I think I am going with #2. Aug 01 20:44:42 that's one of the points i've tried to explain repeatedly over the years... Aug 01 20:45:03 usually not very successfully for one reason or another Aug 01 20:46:32 Regarding the Joel on Software article, there are also industry examples of market-leading companies having their lunch handed to them because they *didn't* rewrite their software. Aug 01 20:46:35 it just seems harder than it should be to get some people to understand that Aug 01 20:47:34 ie, that writing code is fundamentally a knowledge-capturing activity Aug 01 20:47:59 maybe i just need a bigger clue stick Aug 01 20:49:36 mr_science: I worked for a while with a jr. engineer who seemed really weak at figuring out how to do things. After talking with him for a while, it became apparent to me that his main problem was that he went out of his way to avoid *reading* code. Aug 01 20:50:15 How people expect to *write* coherent things in a language they're not adept at *reading* first is beyond me. Aug 01 20:50:19 yikes Aug 01 20:51:34 but yeah, Joel's mom didn't raise no dummies Aug 01 20:51:52 not that i couldn't argue with over some things... Aug 01 20:51:59 *him even Aug 01 20:52:30 I guess it doesn't help much that low-level languages like C make it very easy to write highly imperative code, in which the actual code provides a lot of tiny explicit details without ever making clear *why* those operations are being ordered. Aug 01 20:53:24 yeah, some tools (like that example) require extra thought/work/effort to mitigate those kinds of things Aug 01 20:53:59 which is why i like helpful languages like Ada... Aug 01 20:54:23 * mr_science waits for the groans and jaw-popping Aug 01 20:54:33 * kergoth chuckles Aug 01 20:57:17 mr_science: I really need to learn Ada. I mostly enjoy statically-typed functional languages (Haskell, OCaml) for exploratory/modeling programming, because it helps you get the design hashed out and coherent before you get around the the implementation details. Aug 01 20:58:43 From what I've seen of Ada, it's highly verbose but does provide a flexible set of tools for describing the way you expect things to work and having the compiler verify that you're operating within those constraints. Aug 01 20:59:24 Ada has some lovely features that are still missing from most other languages Aug 01 20:59:32 since Ada 83 even Aug 01 20:59:44 why they're still missing is very good question Aug 01 21:00:11 * fray suspects because of the reputation of Ada Aug 01 21:00:29 probably goes along with the whole "software engineering do as we say not as we do" thing Aug 01 21:00:37 and specifically the reams of documentation when Ada is used in gov't contracted software Aug 01 21:01:28 reams of govt documentation software docs are pretty much independent of the implementation language... Aug 01 21:01:45 er, "software documentation" Aug 01 21:01:49 ya.. but Ada was designed around that Aug 01 21:01:58 not really Aug 01 21:02:01 I think they're primarily missing because no one has done any R&D at the level of C/Ada languages in a long time. You just use C or C++ and deal with it, and those can't change much due to massive amounts of code and trained programmers that can't be obsoleted. Aug 01 21:02:17 the "myths" of Ada revealed... Aug 01 21:02:18 from what I saw it was.. (I saw it from a Honeywell perspective in the late 1980s) Aug 01 21:03:01 the wall of Ada manuals was insane.. the reams of "you must do it this way" docuemnts from the gov't were insane.. Aug 01 21:03:06 * mr_science spent 20+ years doing IV&V on DoD launch range systems Aug 01 21:03:13 Ada was designed around the idea of reliable, deterministic systems. Aug 01 21:03:18 it led to a bad taste in new programers, who said I won't work on system slike that Aug 01 21:03:30 levi, absolutely.. and I think it had it's place Aug 01 21:04:16 Ada was designed to handle several different problem domains, but mostly the design is meant to iprove software quality Aug 01 21:04:24 The current 'state of the art' way to do that is apparently gigantic software systems like UML modeling and code extraction tools. Aug 01 21:04:26 nothing specific to DoD about that Aug 01 21:04:55 ya, I'm not saying what we have today is any better frankly.. Aug 01 21:05:36 if there's one thing i've seen over those 20+ years, it's that almost nobody knows the Ada standards or design criteria Aug 01 21:05:47 but they all likt to talk about it... Aug 01 21:07:15 i also get that some people just don't like things like range variables that won't allow out-of-bounds indeces and things like that Aug 01 21:07:36 personally, i need all the help i can get to eliminate bugs Aug 01 21:07:51 heh, from my experience language usage is very cyclical.. Aug 01 21:07:53 Amen to that. Aug 01 21:08:10 there seems to always be the language of the month behavior.. and then longer term languages.. Aug 01 21:08:16 (C, Ada, Fortran, Cobol...) Aug 01 21:08:27 let's port INTERCAL Aug 01 21:08:33 As Alan Kay observed, we have a programming 'pop culture'. Aug 01 21:08:34 cross-intercal... Aug 01 21:08:55 i think Fortran wins the geriatric language prize Aug 01 21:09:16 but wow, it's still being used for new development.. :/ Aug 01 21:09:27 (I'd rather program Fortran then cobol though) Aug 01 21:09:38 very useful for certain things, yet still a 50's language... Aug 01 21:09:50 It's easier to write high-performance numeric code in Fortran than C, though. Aug 01 21:09:57 yup Aug 01 21:10:13 You don't have to write a bunch of aliasing annotations all over the place. Aug 01 21:10:42 I was in the supercomputing industry for a few years, and Fortran is very much alive and well there. Aug 01 21:10:46 I have a hated for AS/400 & Cobol due to a course I took at my university.. Aug 01 21:11:05 the only real knock on Ada i can think of is that they just did a crappy job selling it Aug 01 21:11:15 Cobol was just kind of a bad idea that gets re-discovered every so often. Aug 01 21:11:21 HAHA Aug 01 21:11:27 given the prevalence of myths still floating around... Aug 01 21:11:52 "Let's write a programming language that untrained managers can understand!" Aug 01 21:12:31 s/programming language/magazine ad/ Aug 01 21:12:32 levi (and in my university case) be sure to tell all of the "programmers in the room" that you are going to show them the 'right way to program'.. Aug 01 21:12:57 * fray is still annoyed (or worse) at that professor... and that was about 18 years ago Aug 01 21:13:03 mr_science: Well, compared to other languages of its vintage, it was a relatively large and complex language. Also, from my understanding there was a long period where there were no free compilers available for it. Aug 01 21:13:49 levi: the latter period was mostly while the original 83 standard was still under development Aug 01 21:13:58 heh CDC's uTutor was better then Cobol IMHO Aug 01 21:14:23 I don't think it's possible to understate the importance of the open, freely-sharable nature of Unix and C to their current ubiquity. Aug 01 21:14:46 it's probably still the only language with a compiler validation test suite *or* a standard before compiler implementation Aug 01 21:14:48 absolutely Aug 01 21:15:06 both of which are Good Things Aug 01 21:15:43 They are not ubiquitious because they're the *best*, but because they're relatively small, simple, portable, and freely available. Aug 01 21:16:14 yeah, but have you tried any of the "free" non-GNU C compilers? Aug 01 21:16:31 pretty easy to find compiler bugs in them Aug 01 21:16:39 at least it used to be... Aug 01 21:16:44 If you haven't read the UNIX-HATERS Handbook, it's entertaining reading even if you like Unix. Aug 01 21:17:10 levi: why Aug 01 21:17:43 i pretty sure i read that stuff a loooong time ago, along with the original BOFH stories... Aug 01 21:17:53 *i'm Aug 01 21:18:04 mr_science: On the other hand, it's also not uncommon to find bugs in obscure parts of gcc. I ran into one or two many years ago while doing work on a NIOS2-based system. Aug 01 21:18:20 maybe it's time to skim it again in the cold light of ... something... Aug 01 21:19:16 levi: didn't say gnu code was bug-free Aug 01 21:19:22 lpapp_: It's filled with humorous rants, many of which come from users of other contemporary systems that were more advanced in a lot of ways. Aug 01 21:19:33 just orders of magnitude cleaner usually... Aug 01 21:20:20 My initial experience with C code bases was actually from reading the code for MUD-style games. Aug 01 21:20:25 okay, i think i finished this bash script about 25 minutes ago... Aug 01 21:20:29 Specifically the Tiny- variety. Aug 01 21:21:01 oops, forgot my die() function Aug 01 21:21:04 I mostly learned C from roguelikes. Aug 01 21:21:31 It was very enlightening to see how a couple of programs that appeared very similar to users could be implemented so differently, and how much clearer one was than the other. Aug 01 21:21:51 Oddly enough, the clearer one tended to crash a lot less... Aug 01 21:22:22 funny how that works... Aug 01 21:22:45 can you say "obfuscated C contest"? Aug 01 21:26:03 just the fact that you can actually do some of that stuff with C is pretty scary Aug 01 21:32:23 when i weigh making a .emacs (again) against paying $79 for sublime Aug 01 21:32:48 i'm thinking the $79 is a good investment... Aug 01 21:33:16 since i haven't made a custom .emacs in years Aug 01 21:34:42 i picked it up, but i do too much remotely over ssh to make much use of it :( Aug 01 21:35:04 works okay over ssh -Y Aug 01 21:35:55 damn, it's like those guys read my mind... Aug 01 21:36:07 * mr_science gets out his foil hat again Aug 01 21:36:12 just in case... Aug 01 21:51:09 I like sublime. The editor GUI is very slick/optimized, and it's extensible. Aug 01 21:51:28 On the other hand, emacs is very familiar, so I mostly just use emacs. Aug 01 21:51:59 There are a couple of pre-made .emacs directories that are nice. Aug 01 21:52:29 Like the Emacs Starter Kit Aug 01 22:04:47 yeah, i specific things in mine so i had to make my own Aug 01 22:04:54 *wanted even Aug 01 22:05:31 sublime seems to have almost exactly what i need working out-of-the-box Aug 01 22:05:57 plus other stuff i haven't even played with yet Aug 01 22:06:11 For me, it's not the nice features but the muscle memory for standard emacs bindings that's got me stuck with emacs. :) Aug 01 22:06:20 yup Aug 01 22:06:46 Although I sometimes make extensive use of org-mode Aug 01 22:07:16 like kergoth, i ended up doing so much work over shell connections i got used to simple Aug 01 22:08:10 i need to break out of that mold... Aug 01 22:10:11 Okay, before I spend a huge amount of time tracking this down… Had a build fail because was absent. Using binary toolchain which may well not have it. Just wondering whether that's "report to vendor", or "don't be silly, turn on rpc in DISTRO_FEATURES", or what... Aug 01 22:11:25 heh, my sdk build failed last night with a missing /bin/sed in the native sysroot Aug 01 22:11:53 i filed a bug, but it sort-of felt wrong... Aug 01 23:11:07 so now that i look again, sublime does fall short of vi in several areas Aug 01 23:15:23 I seem to recall having tried sublime some, and not been especially happy, but… It is hard for me to learn editors. Aug 01 23:16:33 it looks pretty nice at first, and i do like the fact that certain things Just Work without any fiddling Aug 01 23:16:43 but... Aug 01 23:17:33 when you look a little closer, it's short on language sensitivity and a few other things Aug 01 23:18:26 I mostly don't use language-sensitive features in editors, and actually go around turning them off. I never got used to things like syntax highlighting, and tend to find them distracting. Aug 01 23:28:23 mr_science: If you do much C programming, there's an interesting integration of clang with sublime to catch errors on the fly. Aug 01 23:28:41 depends on the color scheme, but i *do* find syntax highlighting helpful Aug 01 23:29:05 levi: now yer talkin' Aug 01 23:29:36 more analysis tools should be "built-ins" Aug 01 23:30:18 of course, with Ada the compiler pretty much does that Aug 01 23:31:20 personally, i would much rather fix compile errors than chase runtime bugs... Aug 01 23:51:59 rburton: I am running into an issue which boils down to this fix https://bugzilla.yoctoproject.org/show_bug.cgi?id=2597 Aug 01 23:52:00 Bug 2597: normal, Medium+, 1.3 M5, constantinx.musca, VERIFIED FIXED, intltool is missing intltool.m4 Aug 01 23:52:24 rburton: why is RDEP on a -dev package added ? Aug 01 23:52:44 IMO needed files should have been moved into gettext main package Aug 01 23:52:58 and dependency on gettext be created Aug 01 23:53:32 the description in bug its not clear what it fixed and what was broken Aug 02 00:21:21 khem: I just tested that pcc/floor issue with 1.6-rc1 qemu and got a failure so something changed between 1.4 and 1.5 versions of qemu, I guess I could try a bisect, but maybe you have other ideas? Aug 02 02:07:40 khem: I am building gcc fix Aug 02 02:42:10 sgw_: are you able to get same failure on 1.6rc1 or a different one ? Aug 02 02:42:30 otavio: ok. Cool add your results to patch thread Aug 02 02:42:51 1.6? Aug 02 02:43:05 otavio: thats the qemu version Aug 02 02:43:08 not YP Aug 02 02:43:24 oh **** ENDING LOGGING AT Fri Aug 02 02:59:59 2013