| 01:02.40 | *** join/#edev miltonm (~miltonm@71.22.127.106) |
| 01:35.46 | *** join/#edev kergoth_ (~kergoth@ip24-251-173-232.ph.ph.cox.net) |
| 02:41.14 | *** join/#edev roxfan (dunno@91.180.8.123) |
| 02:42.14 | *** join/#edev prpplague (~prpplague@ppp-70-242-126-3.dsl.rcsntx.swbell.net) |
| 03:08.34 | *** join/#edev djerome (~djerome@ip24-251-138-141.ph.ph.cox.net) |
| 03:22.19 | *** join/#edev djerome (~djerome@ip24-251-138-141.ph.ph.cox.net) |
| 03:31.53 | prpplague | man i never expected it to be so hard to get rid of my intronix logicport |
| 03:34.04 | alex4nder | haha |
| 03:34.07 | alex4nder | albatross hardware |
| 03:37.00 | prpplague | alex4nder: indeed |
| 03:37.18 | prpplague | alex4nder: the manufacturer is very anti linux |
| 03:37.36 | prpplague | alex4nder: and no one wants it because it isn;t supported under linux |
| 03:37.57 | prpplague | atleast no one i know wants it |
| 03:38.31 | *** mode/#edev [+o prpplague] by ChanServ |
| 03:40.51 | alex4nder | yah, that's why I don't have one. |
| 03:41.18 | alex4nder | it's too bad |
| 03:41.48 | prpplague | have a LA or have an intronix la? |
| 03:42.07 | alex4nder | intronix |
| 03:42.20 | prpplague | ahh |
| 03:42.26 | alex4nder | it seems OK if you use it exactly as they sell it, under Windows |
| 03:42.36 | prpplague | very limited capture |
| 03:42.47 | alex4nder | yah |
| 03:43.03 | alex4nder | it's definitely not like a current-model Agilent LA |
| 03:43.23 | alex4nder | speaking of which, I kind of want one, their pricing is actually almost reasonable. |
| 03:43.28 | prpplague | yea well even the chronovu is better |
| 03:51.03 | prpplague | alex4nder: if you know of anyone that might be interest, i'd appreciate a mention |
| 05:03.14 | *** join/#edev mbuf (~user@nat/redhat/x-nmmazotjjjrfmrrm) |
| 06:04.00 | alex4nder | prpplague: will do |
| 06:54.25 | *** join/#edev kirabo_ (~kirabo@LCaen-156-54-47-233.w80-11.abo.wanadoo.fr) |
| 07:16.08 | *** join/#edev florian (~fuchs@port-217-146-132-69.static.qsc.de) |
| 07:16.08 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 07:53.37 | *** join/#edev gvandenbosch (~gvandenbo@86.81.67.223) |
| 09:00.31 | *** join/#edev JoeLlama (snork@unaffiliated/joellama) |
| 09:09.01 | *** join/#edev amarsman (~marsman@52488909.cm-4-1c.dynamic.ziggo.nl) |
| 10:09.19 | *** join/#edev _abc_ (~user@unaffiliated/ccbbaa) |
| 12:08.12 | *** join/#edev _abc_ (~user@unaffiliated/ccbbaa) |
| 12:51.57 | *** part/#edev TemporalBeing (~bmeyer@12.52.185.66) |
| 13:35.04 | *** join/#edev bgamari (~ben@gamari.physics.umass.edu) |
| 13:53.18 | *** join/#edev miltonm (~miltonm@32.97.110.51) |
| 14:23.41 | *** join/#edev GPSFan (~kenm@64.92.145.112) |
| 14:23.41 | *** mode/#edev [+o GPSFan] by ChanServ |
| 14:26.25 | *** join/#edev kergoth (~kergoth@ip24-251-173-232.ph.ph.cox.net) |
| 14:37.02 | *** join/#edev CosmicPenguin (~nobody@soa.codeaurora.org) |
| 14:43.59 | *** join/#edev bgamari (~ben@physicsnat56.physics.umass.edu) |
| 14:46.45 | *** join/#edev sjhill (~sjhill@home.bethel-hill.org) |
| 14:46.45 | *** mode/#edev [+o sjhill] by ChanServ |
| 14:58.15 | *** join/#edev bgamari (~ben@gamari.physics.umass.edu) |
| 15:01.56 | landley | Morning. |
| 15:19.38 | CosmicPenguin | Two days of landley in a row. We've got a streak going |
| 15:19.44 | *** join/#edev kergoth (~kergoth@ip24-251-173-232.ph.ph.cox.net) |
| 15:23.11 | *** join/#edev jayabharath (~jayabhara@nat/ti/x-xaddniriqpspfntf) |
| 15:46.37 | landley | Mostly just "I set up the thing again". |
| 15:55.21 | *** join/#edev prpplague (~danders@nat/ti/x-cvotjngfnljuyxos) |
| 16:08.21 | *** join/#edev miltonm (~miltonm@32.97.110.53) |
| 16:26.38 | *** join/#edev kergoth (~kergoth@ip24-251-173-232.ph.ph.cox.net) |
| 16:29.42 | *** join/#edev kergoth (~kergoth@ip24-251-173-232.ph.ph.cox.net) |
| 16:41.20 | *** join/#edev causticsand (~quassel@pool-71-114-243-164.austtx.dsl-w.verizon.net) |
| 16:42.15 | *** join/#edev sndcrb (~sndcrb@dsl-hkibrasgw2-fef8de00-126.dhcp.inet.fi) |
| 16:50.06 | *** join/#edev kergoth (~kergoth@ip24-251-173-232.ph.ph.cox.net) |
| 16:55.37 | *** join/#edev kergoth (~kergoth@ip24-251-173-232.ph.ph.cox.net) |
| 17:19.50 | *** join/#edev kergoth (~kergoth@ip24-251-173-232.ph.ph.cox.net) |
| 17:39.16 | landley | So llvm/clang built a working Linux kernel. Cool. |
| 17:39.57 | landley | The question is, when I can I use it as a drop-in replacement for gcc in a larger build... |
| 17:40.27 | landley | There's no obvious list for just that. |
| 17:40.33 | landley | wants a clang-linux-announce. |
| 17:46.12 | landley | Sigh. What I really want is a compiler written _in_C_ to be able to compile the kernel. |
| 17:46.37 | landley | Moving forward off of gcc 4.2 you need C++ to have a self-bootstrapping system, which was never the case before the FSF lost its last vestiges of restraint. |
| 17:46.53 | landley | (I was going to say sanity, but that's not correct. They lost that in the 90's.) |
| 17:48.14 | landley | http://bsdfund.org/bundle is nice but not a result. |
| 17:52.56 | landley | Ah right, I tried engaging them back at http://comments.gmane.org/gmane.comp.compilers.pcc/1996 and the thread petered out with them not replying to my first actual patch... |
| 18:00.11 | *** join/#edev dfarnsworth (~dfarnswor@nat/montavista/x-xunwsvvnhgjpaduk) |
| 18:36.57 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 19:01.56 | _abc_ | hi landley |
| 19:02.01 | _abc_ | been reading what you wrote |
| 19:02.09 | _abc_ | I am frustrated by avr-gcc not building |
| 19:02.31 | _abc_ | ranting in #gcc made me meet a troll who was pushing VS values in #gcc and nothing else |
| 19:02.46 | _abc_ | avr-c is also moving towards llvm I think |
| 19:03.03 | _abc_ | Just right now building avr-gcc simply fails |
| 19:03.43 | _abc_ | All versions I tried, binutils build, gcc barf. It revolves around some poorly integrated patches and really broken makefiles |
| 19:03.54 | _abc_ | I tried to patch the offending stuff out but failed |
| 19:04.15 | _abc_ | So I came to #avr and #gcc to rant and nada. No-one responsible showed up. |
| 19:20.44 | landley | _abc_ what are VS values? |
| 19:20.55 | landley | What's avr-gcc? |
| 19:21.12 | landley | avr is a hardware platform? |
| 19:26.16 | _abc_ | avr is Atmel's microprocessor line which has more members than a dog has fleas. #avr is its freenode channel, avr-gcc is gcc built to cross compile for avr target. |
| 19:26.47 | _abc_ | So building avr-gcc consists in building the cross compiler (in the gcc source tree). |
| 19:27.39 | *** join/#edev sndcrb (~sndcrb@dsl-hkibrasgw2-fef8de00-126.dhcp.inet.fi) |
| 19:27.50 | _abc_ | The breakage comes from substandard testing and integration which causes totally nonsensical make generated command lines while building the toolchain. This and badly tested architecture selection flags make that building a nightmare. |
| 19:31.51 | prpplague | landley: a small 8-bit microcontroller line, in terms of arm, its kind of like an arm7tdmi |
| 19:32.52 | landley | _abc_ There are something like 5 active compiler projects I'm aware of. I'm forgetting the name of one of them but only because it's not very interesting. |
| 19:33.37 | landley | One is tinycc, which did tccboot in 2004 but then its founder went on to found qemu and stopped having time for it, and it's languished ever since. Its current maintainer is a windows guy who totally does not grok linux. |
| 19:33.54 | landley | One is pcc, which is done by the openbsd and netbsd guys. |
| 19:34.17 | landley | It's fairly promising and is written in pure C. But building the Linux kernel is a "gee, that would be nice" thing rather than an active todo item. |
| 19:34.28 | landley | They did build the whole of BSD with it and have a 1.0 release, though. |
| 19:34.39 | landley | The other is LLVM/clang, which is sponsored by Apple. |
| 19:34.51 | landley | Basically when gcc went gplv3 lots of people gave up on it. |
| 19:35.22 | wmat | lots? |
| 19:35.25 | landley | I stuck with the last gplv2 release (4.2.1), the bsd guys revived pcc from the 1970's and turned it into a modern c99 compiler, and Apple sponsored a graduate student project and threw money at it to turn it into a real compiler. |
| 19:35.25 | *** mode/#edev [+o prpplague] by ChanServ |
| 19:35.41 | landley | wmat: apple, the bsd guys, and a decent chunk of the Linux community... yeah. |
| 19:35.55 | landley | This is before the gcc developers decided to rewrite 4.5 in C++. |
| 19:37.10 | _abc_ | Ah so 4.2.1 is not gpl3? I did not realize politics were involved |
| 19:37.14 | _abc_ | That explains a lot |
| 19:37.15 | wmat | gcc is still the defacto standard for most dev |
| 19:37.40 | landley | wmat: sure, but MacOS X and the BSDs are stuck on 4.2.1 for the same reason my project is. |
| 19:37.45 | landley | They don't want to get any of GPLv3 on them. |
| 19:37.53 | landley | They're actively sponsoring replacement projects. |
| 19:38.11 | landley | _abc_ I'm not counting Intel's ICC, which is closed source and x86 only, but _is_ only the third compiler to ever build the Linux kernel. |
| 19:38.19 | *** join/#edev miltonm (~miltonm@32.97.110.53) |
| 19:38.30 | landley | (The other two being tinycc, which built a modified subset of an obsolete version back in 2004, and gcc itself.) |
| 19:38.41 | landley | The fourth was, of course, llvm/clang. |
| 19:38.42 | _abc_ | pcc is interesting. The avr-c people seem to be going after clang |
| 19:38.52 | _abc_ | Which given pcc exists is weird |
| 19:39.24 | landley | wmat: Microsoft Office is the defacto standard for word processing. Doesn't mean it's staying there indefintitely or that lots of third parties are working to add to it. |
| 19:39.31 | _abc_ | Anyway makefiles for such a complex tiny bugger set (there are 7 incompatible architectures in the avr line) is pretty much impossible to maintain other than as a full time job |
| 19:39.52 | landley | _abc_ The BSD guys have a cleaner design but Apple has a lot fo money to throw around which gets them faster development. |
| 19:40.04 | _abc_ | Yes but what is the license like? |
| 19:40.18 | _abc_ | pcc must be some BSD licensed thing so clean and no FUD |
| 19:40.26 | _abc_ | clang is how? |
| 19:40.26 | landley | Also, corporations like signing up for Apple's initiative. Qualcomm added Hexagon support to LLVM last year, for example. |
| 19:40.32 | landley | PCC is BSD, yes. |
| 19:40.39 | landley | LLVM/CLANG is also BSD, I believe. |
| 19:40.43 | _abc_ | hmm |
| 19:40.45 | wmat | i agree that the GPLv3 is repulsive but Google is still very much behind GCC |
| 19:40.58 | _abc_ | behind in what way? |
| 19:41.10 | wmat | monetarily |
| 19:41.24 | wmat | they invest money and resources in the GNU Toolchain dev |
| 19:41.33 | wmat | as do many other though |
| 19:41.33 | landley | Google is a large bureacratic entity. Have been for about five years now. They're a technical follower, not a leader. |
| 19:41.42 | _abc_ | Heh I wonder if the G in Google is not borrowed from Gnu... just as a figurehead. Galeon figure whatever. |
| 19:42.01 | wmat | landley: the same could be said for Apple |
| 19:42.06 | landley | Sure. |
| 19:42.23 | landley | wmat: actually, that only kicks in after Steve Jobs dies. |
| 19:42.24 | _abc_ | And I don't think gpl3 is repulsive, it is definitely political. So is the m$ eula. Pick your poison... |
| 19:42.38 | landley | Apple had one guy in charge who could say "do this" and they would. They were following his vision. |
| 19:42.47 | landley | Google is run by committee. Their current big initiative is to clone facebook. |
| 19:43.08 | wmat | perhaps the GPLv3 will get revised before a suitable GCC replacement is complete |
| 19:43.09 | landley | gplv3 was just the last straw. gcc has been bloatware for years. |
| 19:43.21 | wmat | i don't disagree with that |
| 19:43.48 | landley | wmat: gcc is too big and too complicated and _explicitly_ didn't support a modular architecture for years because Richard Stallman was terrified of proprietary plug-ins interoperating with it. |
| 19:43.49 | _abc_ | Well Sony was a good example of how a good company goes to the dogs when the principal founder exits. When Mr. Sony released the reins Sony took a dip it almost did not survive. I worked in the video systems industry and witnessed their quality and good name go into the ground in 2 years or fewer. |
| 19:43.59 | landley | There was a good LWN article on that years ago. |
| 19:44.03 | wmat | but it kind of has to be, given the diversity of technologies using it |
| 19:44.27 | landley | pcc and llvm convinced him to rescind that part, but they've got a decade of cleanup to do and still have a copyright assignment barrier to random drive-by cleanups. |
| 19:44.45 | landley | The kind of developers likely to clean up an existing project are exactly the kind that won't bother with copyright assignment. |
| 19:44.52 | _abc_ | I will take a close look at pcc |
| 19:44.59 | _abc_ | Is pcc available as a package for linux too? |
| 19:45.11 | landley | The corporate developers who fill out copyright assignment paperwork because their boss pays them to don't bother with cleanups. That's not productivity, adding new features is. |
| 19:45.17 | _abc_ | I mean standard ELF and ld.so out ? |
| 19:45.18 | landley | abc: yup. |
| 19:45.27 | _abc_ | goes to look |
| 19:45.40 | landley | wmat: not really. |
| 19:46.00 | landley | The messiness of gcc is because of the FSF, not because of the problem space. |
| 19:46.24 | wmat | oh yeah, i agree with that |
| 19:46.36 | wmat | but it is what it is |
| 19:48.50 | _abc_ | landley: mister you made my day |
| 19:48.52 | _abc_ | with pcc |
| 19:49.03 | landley | abc: Glad I could help. :) |
| 19:49.18 | landley | wmat: the present is reality. The future is malleable. |
| 19:49.26 | wmat | no doubt |
| 19:49.27 | _abc_ | heh I hesitated to make the move on pcc but now... wow 650k sources and builds in a few minutes? |
| 19:49.29 | landley | I'm aware that gcc is dominant today the same way windows is dominant today. |
| 19:49.34 | landley | Doesn't mean it's what I want to work on improving. |
| 19:50.02 | landley | If "it's dominant therefore it's what you must do" had any effect on me, I wouldn't be doing Linux. |
| 19:50.38 | wmat | one can do whatever they so choose, i suppose |
| 19:50.46 | landley | abc: you think that's impressive, tinycc had 100k source and built in a couple seconds. Alas, it had no optimizer whatsoever and still not quite fully c99 (no variable length arrays). |
| 19:51.15 | landley | I keep thinking about poking at tinycc again, but... https://www.codeaurora.org/contribute/projects/hexagon-linux-kernel/ |
| 19:51.16 | _abc_ | landley: you can't scare me. I work with mcus and I used compilers on CP/M yet. |
| 19:51.20 | landley | Ah, wrong paste buffer |
| 19:51.26 | landley | http://landley.net/code/tinycc |
| 19:51.37 | _abc_ | landley: 2 diskettes, one compiler, one linker... |
| 19:51.41 | wmat | what archs does tinycc and pcc support? |
| 19:51.47 | _abc_ | And I mean 150k diskettes |
| 19:51.55 | _abc_ | wmat: I can tell you pcc right now: |
| 19:52.06 | landley | abc: by first computer was a commodore 64. |
| 19:52.30 | _abc_ | amd64 arm CVS hppa i386 m16c mips nova pdp10 pdp11 powerpc sparc64 superh vax |
| 19:52.46 | landley | wmat: the netbsd guys are switching over to pcc so the _plan_ is to eventually support everything in the world. But it's very much a work in progress. |
| 19:52.57 | wmat | no doubt |
| 19:52.59 | _abc_ | landley: hehe my first bought computer was a Sinclair ZX80 second hand. I never bought one after that, always built. Excepting laptops... |
| 19:53.08 | wmat | that's a pretty good amount of archs though |
| 19:53.11 | _abc_ | wmat: ^^ that was for you |
| 19:53.14 | landley | http://lwn.net/Articles/255558 |
| 19:53.17 | _abc_ | wmat: also os list: |
| 19:53.27 | landley | That was coverage of its relaunch announcement, in 2007. |
| 19:53.36 | _abc_ | bsd CVS darwin dragonfly freebsd linux midnightbsd minix mirbsd netbsd nextstep none openbsd sunos symbian win32 |
| 19:53.55 | wmat | is too old for the first computer question |
| 19:53.58 | _abc_ | mirbsd is a Russian one I think, midnightbsd I never heard of |
| 19:54.10 | _abc_ | I will run some tests to see how gcc and pcc outputs compare |
| 19:54.24 | _abc_ | wmat: hehe |
| 19:54.34 | *** join/#edev miltonm (~miltonm@32.97.110.53) |
| 19:54.38 | _abc_ | wmat: say "Babbage V0.9" and you'll impress us :) |
| 19:55.17 | landley | Ah, here we go: http://lwn.net/Articles/441018 |
| 19:55.23 | _abc_ | The avr people concentrate on the 'none' support |
| 19:55.31 | _abc_ | avrs are too small for an os |
| 19:55.56 | landley | I've been regression testing build environments with uClibc and busybox for years. Adding in a nonstandard compiler isn't a big stretch at this point. :) |
| 19:56.02 | _abc_ | Typically 4 to 26kbytes flash and 1kram and 1keeprom, 32 registers, 8bit arch |
| 19:56.15 | landley | (And by regression testing I mean getting it to rebuild itself under itself, then getting it to build Linux From Scratch and Beyond Linux From Scratch packages...) |
| 19:56.39 | _abc_ | you can build full linux under busybox? |
| 19:56.54 | landley | _abc_ Yup. |
| 19:57.10 | landley | Seven packages: linux, uClibc, busybox, binutils 2.17, gcc 4.2.1, make, and bash. |
| 19:57.31 | landley | (Someday busybox shells should be able to replace bash, but last I checked...) |
| 19:57.51 | landley | I'm using an ancient bash 2.05b rather than current, though. it's like 1/3 the size. |
| 20:00.25 | _abc_ | Wait busybox did not update to gplv3 |
| 20:00.26 | _abc_ | ? |
| 20:00.38 | _abc_ | That 4.2.1 again...? |
| 20:01.22 | landley | abc: Bruce Fucking Perens tried to make us do so. |
| 20:01.24 | _abc_ | realizes a political issue around gcc made him waste a week or more on trying to build avr-gcc and feels really p**d off |
| 20:01.26 | landley | I was busybox maintainer at the time. |
| 20:01.36 | _abc_ | So what was the outcome? |
| 20:01.41 | landley | I locked the license down to GPLv2 only, same as the Linux kernel, before retiring wounded from the field. |
| 20:01.57 | _abc_ | And did they move on? |
| 20:03.04 | landley | (Basically after I drove the darn troll off I was so disgusted wtih the project I couldn't stand to LOOK at the code anymore. He could take credit for it even though I proved he had nothign to do with the current project, Erik Anderson did it.) |
| 20:03.08 | landley | And by proved I mean http://busybox.net/~landley/forensics.txt |
| 20:03.28 | landley | Erik said I should take a few months off, but that didn't seem int he best interests of the project, so I handed it to Denys instead. |
| 20:04.16 | landley | http://lwn.net/Articles/202108 |
| 20:05.07 | *** join/#edev bgamari (~ben@physicsnat56.physics.umass.edu) |
| 20:13.08 | *** join/#edev sndcrb (~sndcrb@dsl-hkibrasgw2-fef8de00-126.dhcp.inet.fi) |
| 20:17.01 | prpplague | anyone remember a rant on lwn awhile back about the evils of eMMC? |
| 20:17.15 | alex4nder | yah |
| 20:17.35 | prpplague | alex4nder: you have the url? |
| 20:17.45 | prpplague | alex4nder: i've been trying to find it |
| 20:18.14 | alex4nder | er, no,.. I just remember seeing it. |
| 20:18.21 | landley | What's eMMC? |
| 20:19.06 | prpplague | embedded MMC |
| 20:19.19 | prpplague | basically a low voltage sd card that is soldered to the motherboard |
| 20:23.00 | landley | prpplague: not http://lwn.net/Articles/428584 |
| 20:23.23 | prpplague | naw |
| 20:26.49 | landley | prpplague: a rant in the comments or a rant in an article? |
| 20:27.01 | prpplague | rant in the article |
| 20:28.09 | landley | There were a number about using non-flash filesystems on flash... |
| 20:28.12 | landley | Which "evils" did it cover? |
| 20:30.22 | prpplague | wear leveling, performance, filesystem degradation |
| 20:35.30 | alex4nder | the biggest problem my clients have been seeing with eMMC/MMC/SD/whatever is the variance in performance |
| 20:35.45 | alex4nder | no batches are the same as previous batches, and sometime the read/write performance is laughable. |
| 20:44.58 | prpplague | alex4nder: yea |
| 20:45.25 | landley | Hmmm, not http://lwn.net/Articles/326471 |
| 20:45.59 | landley | Maybe http://lwn.net/Articles/353411 ? |
| 20:46.33 | prpplague | naw, the whole article was dedicated to the ills of eMMC |
| 20:46.44 | prpplague | seems like it was from someone in the MTD group |
| 20:50.10 | landley | Don't think I read it. |
| 20:59.59 | prpplague | yea, i couldn't find it either |
| 21:00.07 | prpplague | landley: maybe it was my mind playing tricks on me |
| 21:11.09 | landley | Maybe it wasn't lwn? |
| 21:14.02 | *** join/#edev bgamari (~ben@gamari.physics.umass.edu) |
| 21:14.24 | *** part/#edev _abc_ (~user@unaffiliated/ccbbaa) |
| 22:08.36 | *** join/#edev dijenerate_ (~dijenerat@204.212.240.159) |
| 22:08.42 | *** join/#edev balister_ (~balister@pool-96-240-182-99.ronkva.east.verizon.net) |
| 22:14.02 | *** join/#edev GPSFan (~kenm@64.92.145.112) |
| 22:14.09 | *** mode/#edev [+o GPSFan] by ChanServ |
| 22:46.20 | *** join/#edev landley1 (~landley@140.242.26.81) |
| 22:50.43 | *** join/#edev GPSFan_ (~kenm@64.92.145.112) |
| 22:50.43 | *** mode/#edev [+o GPSFan_] by ChanServ |
| 22:53.51 | *** join/#edev jayabharath (~jayabhara@nat/ti/x-euocbdxwmopmcbks) |
| 22:59.59 | *** join/#edev SpeedEvil (~user@tor/regular/SpeedEvil) |
| 23:11.03 | *** join/#edev bgamari (~ben@physicsnat56.physics.umass.edu) |
| 23:42.11 | *** join/#edev kergoth__ (~kergoth@ip24-251-173-232.ph.ph.cox.net) |