| 01:13.02 | *** join/#edev sterNiX (~LessIsMor@unaffiliated/nu253r/x-0655220) |
| 01:56.58 | *** join/#edev sterNiX (~LessIsMor@unaffiliated/nu253r/x-0655220) |
| 03:01.52 | *** join/#edev War2 (~war2@unaffiliated/war2) |
| 03:34.34 | *** join/#edev sterNiX (~LessIsMor@unaffiliated/nu253r/x-0655220) |
| 04:26.35 | *** join/#edev War2 (~war2@unaffiliated/war2) |
| 04:26.46 | *** join/#edev bsdfox (~h36sa@c-71-197-82-240.hsd1.ca.comcast.net) |
| 04:26.46 | *** join/#edev bsdfox (~h36sa@unaffiliated/bsdfox) |
| 04:39.07 | *** join/#edev War2_ (~war2@14.98.107.37) |
| 05:30.30 | *** join/#edev War2 (~war2@unaffiliated/war2) |
| 06:32.02 | *** join/#edev clio (~andrej@85.159.109.222) |
| 08:35.05 | *** join/#edev yoann_ (~yoann@archos.rain.fr) |
| 09:16.36 | *** join/#edev bsdfox (~h36sa@unaffiliated/bsdfox) |
| 09:25.43 | *** join/#edev florian (~fuchs@port-217-146-132-69.static.qsc.de) |
| 09:25.43 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 09:36.43 | *** join/#edev mbuf (user@nat/redhat/x-jeudtkzssbnhdfsk) |
| 10:08.00 | *** join/#edev sledges (~sledges@unaffiliated/sledges) |
| 10:08.02 | sledges | hello |
| 10:08.56 | sledges | where could I buy an Atheros 6003 based SDIO WLAN card? seems that the final product manufacturers are simply not referencing Atheros, hence reverse googling fails.. |
| 10:15.39 | *** join/#edev dijenerate (~dijenerat@66.54.118.89) |
| 10:50.17 | *** join/#edev av500 (~av500@p50996ded.dip0.t-ipconnect.de) |
| 11:39.57 | *** join/#edev g1powermac (~g1powerma@rrcs-173-197-150-24.west.biz.rr.com) |
| 11:39.57 | *** join/#edev g1powermac (~g1powerma@unaffiliated/g1powermac) |
| 11:44.35 | mnemoc | prpplague: Hi, have you heard of http://elinux.org/Embedded_Open_Modular_Architecture ? :) It's a project to make standarized user-replaceable SoMs. currently designing in PCMCIA (EOMA-68) and CF (EOMA-50) form factors/connectors. It would be really really awesome if you could stop by :) |
| 11:59.12 | mnemoc | prpplague: there is an spread fear to TI in there and your overall experience and knowledge would help a lot. |
| 13:11.30 | prpplague | mnemoc: can you ping me in a couple hours? |
| 13:19.11 | *** join/#edev War2 (~war2@unaffiliated/war2) |
| 13:24.18 | mnemoc | prpplague: sure :) |
| 14:17.36 | Mirell | Morning. |
| 14:18.23 | Mirell | http://blog.zx2c4.com/749 |
| 14:19.24 | mnemoc | uh |
| 14:21.25 | Mirell | Lovely, eh |
| 14:25.14 | mnemoc | very :-/ |
| 14:26.26 | mnemoc | but good for rooting >= 2.6.39 devices :p |
| 14:40.51 | Mirell | First day on new job.. |
| 14:41.10 | Mirell | No idea what I'm supposed to do right now. Go Go remote working. |
| 14:47.33 | mnemoc | Mirell: who's the priviledged employer? :) |
| 14:48.57 | Mirell | mnemoc: Basho |
| 14:49.12 | Mirell | They make Riak, an open source Dynamo clone |
| 14:49.51 | Mirell | I'm spending my time reading slides and reading "Programming Erlang" until my boss calls me |
| 14:50.08 | mnemoc | nice, didn't know it |
| 14:50.24 | mnemoc | should learn erlang too |
| 14:50.42 | Mirell | That was my thought, hence why I applied for this job. |
| 14:51.07 | mnemoc | :) |
| 15:04.10 | g1powermac | Erlang. . .interesting |
| 15:05.40 | Mirell | It's quite a different language, so far into this book I can tell. |
| 15:10.46 | g1powermac | yea, I'm looking at some code examples on wikipedia |
| 15:11.42 | g1powermac | it almost has a strange sorta resemblance to python, which is what I've been learning lately |
| 15:12.31 | Mirell | Except not even vaguely close. |
| 15:12.39 | g1powermac | yea |
| 15:12.56 | Mirell | Go Go Functional Programming Languages |
| 15:15.07 | g1powermac | there's some interesting projects using erlang |
| 15:19.49 | Mirell | Riak, for one. =D |
| 15:20.02 | Mirell | CouchDB is another well-known one. |
| 15:21.19 | *** join/#edev landley (~landley@140.242.26.2) |
| 15:21.34 | mnemoc | github's git server is also written in erlang iirc |
| 15:21.35 | g1powermac | yea, the CouchDB one surprised me |
| 15:23.37 | g1powermac | is still pondering about messing around with NoSQL stuff. . .its so similar to semantic web triplestores |
| 15:24.17 | mnemoc | there is an unsql standard comming out |
| 15:24.41 | mnemoc | "unstructured" sounds better than "no structured" :p |
| 15:25.19 | landley | mnemoc: doesn't a nosql standard defeat the purpose of not using sql? |
| 15:25.26 | landley | Let's have a standard for not using the computer at all, shall we? |
| 15:25.36 | mnemoc | :D |
| 15:25.40 | g1powermac | heh |
| 15:25.52 | g1powermac | points to paper and pencil ;-) |
| 15:26.20 | mnemoc | http://www.unqlspec.org/display/UnQL/Example+Queries+and+Usage |
| 15:27.14 | Mirell | The strongest mental block with Erlang is that '=' is a pattern-matching operator, rather than strictly assignment. |
| 15:28.33 | g1powermac | yea, I could see that being quite annoying |
| 15:28.43 | g1powermac | mnemoc, that's interesting |
| 15:32.26 | *** join/#edev CosmicPenguin (~nobody@soa.codeaurora.org) |
| 15:32.33 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 15:32.57 | mnemoc | g1powermac: I like standarized interfaces :) |
| 15:34.01 | g1powermac | mnemoc, I've been doing a bit of research into these different database systems, but just not sure which to use |
| 15:34.39 | mnemoc | Mirell: support Mirell, try Riak :p |
| 15:34.43 | mnemoc | err |
| 15:34.51 | mnemoc | g1powermac: support Mirell, try Riak :p |
| 15:34.54 | g1powermac | going to be using a CMS that already uses MySQL for its needs, but would love augment it with other kinds of data |
| 15:35.06 | g1powermac | which is why I've been looking at semantic web triplestores |
| 15:35.25 | g1powermac | but now this nosql stuff looks interesting too, agh, to many options :-) |
| 15:35.51 | Mirell | Hehe |
| 15:36.11 | Mirell | Really depends what you're doing. |
| 15:36.24 | Mirell | Riak aims for the "OMG MUST BE AVAILABLE 24/7 NO MATTER WHAT" crowd. |
| 15:36.39 | g1powermac | ahh |
| 15:37.37 | mnemoc | Mirell: if it supports unql it would be even better, "no lock-in" as a sales point :p |
| 15:38.12 | Mirell | mnemoc: Bleh. Yet another standard. |
| 15:38.22 | mnemoc | :p |
| 15:38.23 | Mirell | mnemoc: I like my ORM wrappers mainly. |
| 15:38.55 | mnemoc | Mirell: don't tell that "herecy" to your new boss :) |
| 15:39.10 | mnemoc | heresy* |
| 15:39.28 | Mirell | mnemoc: He's already been following me on Twitter, so he knows my disdain for YET ANOTHER STANDARD TO UNIFY THE WORLD. |
| 15:41.01 | mnemoc | :) |
| 15:43.26 | Mirell | Wee, yet another Google account. |
| 15:44.17 | *** join/#edev prpplague (~danders@nat/ti/x-fimuwklmvuojqvpf) |
| 15:44.38 | *** mode/#edev [+o prpplague] by ChanServ |
| 15:46.50 | g1powermac | could use a guide explaining the different database paradigms and how they could be used. . . |
| 15:48.35 | landley | g1powermac: the honking big ones were covered in lwn.net article a few months back. |
| 15:48.50 | Mirell | g1powermac: http://kkovacs.eu/cassandra-vs-mongodb-vs-couchdb-vs-redis |
| 15:49.38 | g1powermac | Mirell, hmm, thanks |
| 15:50.03 | landley | g1powermac: http://lwn.net/Articles/452307/ |
| 15:50.28 | landley | And http://lwn.net/Articles/453451/ |
| 15:52.28 | mnemoc | Data Warehousing 101... nice... I missed that one |
| 15:53.50 | g1powermac | oh, data warehousing stuff, I was just digging a bit into that |
| 15:54.59 | g1powermac | been looking at data mining stuff that mentioned it |
| 15:56.46 | *** join/#edev sndcrb (~sndcrb@dsl-hkibrasgw2-fef8de00-126.dhcp.inet.fi) |
| 16:00.45 | Mirell | Hmm...should also brush up on my Javascript |
| 16:05.08 | mnemoc | get o'reilly's "javascript: the good parts" |
| 16:05.44 | mnemoc | there is a companion "google techtalk" very good too |
| 16:05.58 | g1powermac | wants to see Javascript: the naughty parts ;-) |
| 16:06.17 | mnemoc | == is the naughtiest |
| 16:06.27 | g1powermac | hehe |
| 16:14.30 | g1powermac | landley, its somewhat surprising that one article didn't mention semantic web database/triplestore, sesame |
| 16:14.52 | g1powermac | especially with sparql 1.1 now being out |
| 16:15.28 | *** join/#edev GPSFan (~kenm@64.92.145.112) |
| 16:15.28 | *** mode/#edev [+o GPSFan] by ChanServ |
| 16:16.08 | g1powermac | err, actually sparql 1.1 is still a working draft |
| 16:16.33 | landley | g1powermac: all SQL databases descend from the 1970's IBM "R-series" stuff, which was designed around the idea you have big disk and tiny memory, so you only read one record at a time from each table, do your comparison, and write the resulting record back to disk. |
| 16:16.37 | g1powermac | yea |
| 16:16.38 | landley | That was the "joining streams" idea. Central design idea of SQL, with decades of optimization work built on top of that. |
| 16:17.09 | landley | In the late 90's, memory started to get big enough you could fit small database tables entirely in memory, as python dictionaries and such. |
| 16:17.32 | landley | This gives you a three orders of magnitude speedup and random seek capability. |
| 16:18.03 | landley | In 2000 I saw a sourceforge project that was 1000 lines of python, ~800 of which implemented an SQL parser and the rest was just storing the actual tables as python dictionaries. |
| 16:18.19 | landley | And suddenly, all that r-series crap just _goes_away_. You don't need it. |
| 16:19.05 | landley | Start with a gzipped snapshot of the database state, load it into ram and keep it there. Open a log file, and append any transaction that modifies the database state to the log file. (Just writes, reads don't matter.) |
| 16:19.25 | landley | If you lose power or something, re-read your snapshot and replay the commit log. |
| 16:19.55 | g1powermac | interesting |
| 16:20.01 | landley | If your log gets too long and your snapshot too old, fork the database. Have the parent close the old log and start a new one. Have the child write a new snapshot (freeing its copy-on-write memory as it does so). |
| 16:20.34 | landley | Unix process semantics 101: fork gives you copy on write of all the memory, the parent can dirty to its hearts content, the child frees its copies as it streams them to disk. |
| 16:20.42 | landley | If you lose power during _THAT_, read the old snapshot and replay both logs in order. |
| 16:20.57 | landley | Once the new log is complete, you don't need the old snapshot or the old log, the child can move them to archival storage on the way out. |
| 16:21.39 | landley | Taking advantage of SMP brings back locking and threading and such, but it's still entirely in memory and is the same as dealing with any other large data set, not brain surgery to scale that to 8-way or so. |
| 16:22.34 | landley | This is why Oracle has been shitting bricks for the past decade: a disruptive technology has been eating into the low-end of their business and forcing them upmarket, and there's nothing they can do about it because it's based on being so simple you can literally do it in 1000 lines of python. |
| 16:22.48 | landley | (Yeah, python's 10x slower than C so you only get a _2_ order of magnitude speedup. Oh well.) |
| 16:23.42 | landley | Of course if your tables are too big to fit in ram, you still need the old way. But once 64 bit hardware showed up, throwing money at the problem got you tens of gigs of ram. That's about 95% of the market, right there. |
| 16:24.27 | landley | The next obvious step is "so if all this R-series optimization stuff went away because it's optimizing something we don't DO anymore, what do I need SQL for again? It's a hash table, my program can just use those directly." |
| 16:24.36 | landley | And thus, the nosql "movement". |
| 16:24.40 | landley | If _not_ doing something can be considered a movement. |
| 16:25.07 | landley | Up at the high end you've got mapreduce and stuff, and buckets of effort for distributing enormous databases over clusters of machines. |
| 16:25.47 | landley | But again, Oratroll's not driving that. They're being squeezed between nosql at the low end and data warehousing at the high end, and their conventional database business is winding down, and they know it. |
| 16:26.24 | landley | g1powermac: do you now have enough of an idea you could _implement_ a nosql database in under a week? :) |
| 16:26.38 | landley | They're really, really simple. Especially if somebody already wrote the hash table code for you. :) |
| 16:27.00 | g1powermac | huh |
| 16:27.24 | landley | (Linux has fork. Linux has O_APPEND. All the hard bits are part of the OS. :) |
| 16:27.54 | SpeedEvil | Interesting |
| 16:28.18 | *** join/#edev dijenerate (~dijenerat@173.225.251.175) |
| 16:29.53 | g1powermac | http://answers.semanticweb.com/questions/723/rdf-storages-vs-other-nosql-storages |
| 16:30.18 | g1powermac | the second answer has some interesting info discussing triplestores vs nosql |
| 16:31.01 | SpeedEvil | http://tech.slashdot.org/story/12/01/23/0419242/startup-combines-cpu-and-dram |
| 16:31.04 | Mirell | Wee, Skype conferences |
| 16:31.18 | SpeedEvil | may be fun for databses, if they can get them out at a sort-of-comparable cost to DRAM |
| 16:31.29 | SpeedEvil | Or even to DRAM + motherboard |
| 16:31.49 | g1powermac | a real SoC type thing. . . |
| 16:32.05 | SpeedEvil | CPU in a sea of DRAM. |
| 16:32.10 | SpeedEvil | The 1980s is back! |
| 16:32.17 | g1powermac | heh |
| 16:32.22 | landley | Locality of reference never went away. |
| 16:32.27 | landley | The speed of light: still a thing. |
| 16:32.28 | SpeedEvil | True. |
| 16:32.47 | landley | It's all NUMA. |
| 16:33.30 | landley | At what point does "In my SMP system each processor has 4 megabytes of dedicated L2 cache" become NUMA? |
| 16:33.38 | SpeedEvil | I find it kinda mindboggling that capacity has gone up a million times, yet random access time more like 10* from 1980 till now. |
| 16:34.11 | landley | Latency's the same. The round trip speed of light trip is still there. |
| 16:34.36 | SpeedEvil | Well - yes. |
| 16:34.43 | landley | But for sequential accesses you can clock it way fast and parallelize it way high and burst back kilobytes where you used to fetch words. |
| 16:34.44 | SpeedEvil | But DRAM chips are not several feet wide. |
| 16:34.56 | landley | They're 64 bits wide. They used to be 9. |
| 16:35.03 | SpeedEvil | Oh - sure. |
| 16:35.14 | SpeedEvil | Which doesn't help at all if you need a single storage unit. |
| 16:35.16 | landley | And they're DDR, triggering on rising edge _and_ falling edge. |
| 16:35.26 | SpeedEvil | Bandwidth is so much better. |
| 16:35.34 | landley | And they're clocked at a gigahertz, just like the processors. |
| 16:35.45 | landley | But that does nothing for latency of getting a signal three inches across the motherboard. |
| 16:35.56 | landley | Grace Hopper would hand you a nanosecond and a packet of pepper. |
| 16:36.28 | landley | I dunno why I expect more from wonkypedia: http://en.wikipedia.org/wiki/In-memory_database is near useless. |
| 16:36.46 | landley | SpeedEvil: google for "grace hopper david letterman", she was on late night in the 80's, it's a marvelous interview. |
| 16:36.59 | SpeedEvil | I know. Awesome lady. |
| 16:41.12 | *** join/#edev War2 (~war2@unaffiliated/war2) |
| 16:42.30 | SpeedEvil | http://www.ivorcatt.org/icrew2003jun_0003.htm |
| 16:42.35 | SpeedEvil | I hadn't read this before. |
| 16:42.46 | SpeedEvil | On Anamrtic - WSI in the UK in the 80s. |
| 16:43.10 | SpeedEvil | Properly insane managment. |
| 17:17.10 | *** join/#edev florian (~fuchs@sign-4db60287.pool.mediaWays.net) |
| 17:17.11 | *** join/#edev florian (~fuchs@Maemo/community/contributor/florian) |
| 17:17.56 | Mirell | Ahhh, so many new accounts to set up |
| 17:20.52 | landley | SpeedEvil: so this guy is saying "all this Moore's Law doubling we've had for the past 50 years was the result of accountants suppressing engineers". |
| 17:20.54 | landley | Uh-huh. |
| 17:21.08 | landley | I think there's a _reason_ I've never heard of him. |
| 17:21.59 | SpeedEvil | No. |
| 17:22.02 | SpeedEvil | Not quite. |
| 17:22.23 | SpeedEvil | Anamartic was initially setup as a WSI company, with most of the focus on CPUs. |
| 17:22.40 | SpeedEvil | It later got bought and all the focus put on DRAM. |
| 17:22.43 | landley | Computer history is a hobby of mine. I've never heard of anamartic. |
| 17:22.46 | SpeedEvil | RAM |
| 17:23.17 | SpeedEvil | And possible WSI computing applications got dropped. |
| 17:23.31 | SpeedEvil | Indeed - small UK company - not much coverage in the US probably. |
| 17:23.56 | landley | "tried to make a revolutionary CPU" could be transmeta, DEC alpha, the original RISC team that did mips... |
| 17:23.57 | SpeedEvil | They diddn't end up doing much, due to the concentration on RAM, when RAM prices were crashing. |
| 17:24.12 | landley | The itanic was a failed attempt to be revolutionary. |
| 17:24.26 | landley | I think the hexagon's pretty darn neat in a lot of ways, hobbled by bad management. |
| 17:24.50 | landley | SpeedEvil: yeah, I got that from reading the article. (First page and change, anyway.) |
| 17:25.04 | landley | But "companies that went off into the woods and never came back" is not a select group. |
| 17:25.12 | landley | There's three digits of 'em if you wanna go back to the 1960's. |
| 17:25.12 | SpeedEvil | No, I'm not saying it is. |
| 17:25.17 | SpeedEvil | yup |
| 17:26.34 | landley | "The core reason for blocking liquid cooling is that we dare not depart from usage of the Ancients, and the way they designed and built computers in 1950". |
| 17:26.36 | landley | Um: no. |
| 17:26.41 | landley | Liquid cooling leaks. |
| 17:26.44 | landley | Flourinert is expensive. |
| 17:26.53 | landley | This guy has his head so far up his ass he's gone fractal. |
| 17:27.33 | landley | UT had a liquid cooled cluster for a bit. |
| 17:27.42 | landley | There were retail liquid cooled computers at fry's for a while, gaming rigs. |
| 17:27.57 | landley | Heck, IBM tried liquid cooling. |
| 17:27.59 | av500 | once lost a liquid cooled cpu because cooling and power were on 2 different circuit breakers, one tripped, the other did not :) |
| 17:28.17 | landley | This guy is talking about how all the heat these things produce isn't a _real_ problem. |
| 17:28.30 | landley | The Pentium 4 designers thought the same way. |
| 17:28.44 | landley | That particular wall got hit 3 years before he wrote this article. |
| 17:29.16 | landley | "And when they turn the power on, it's sure to dim the lamps. Plus and minus fifteen volts at fourteen hundred amps." |
| 17:29.32 | landley | "And it's, cheer up me lads, let your hearts never fuss, when you're integrating systems, for the S100 bus..." |
| 17:29.35 | landley | Marvelous song. |
| 17:43.06 | mnemoc | prpplague: [reminder ping about EOMA-related talk] :p |
| 17:53.04 | prpplague | mnemoc: yep, thanks, let me ping you in about 30 minutes, finishing up with something now |
| 17:53.44 | mnemoc | prpplague: sure, no rush :) |
| 18:34.08 | prpplague | mnemoc: ok back |
| 18:34.13 | prpplague | mnemoc: i read the url |
| 18:34.28 | prpplague | mnemoc: seems a reasonable idea, but then it has been an item for about 10 years |
| 18:35.04 | prpplague | mnemoc: problem usually happens when companies like LogicPD want to put a bunch CPLD stuff onboard and get pissy about the standard being open |
| 18:35.25 | mnemoc | uh |
| 18:36.16 | g1powermac | well now: http://yro.slashdot.org/story/12/01/23/1645201/senator-rand-paul-detained-by-the-tsa |
| 18:38.46 | *** join/#edev T0mW (~Tom@70.15.162.118.res-cmts.t132.ptd.net) |
| 18:39.25 | T0mW | You know, git works very well, except it falls flat on its behind when it comes to a log. |
| 18:39.50 | g1powermac | hey T0mW |
| 18:39.54 | T0mW | used 'git mv ' to move a file, now the log for that file disappeared. |
| 18:40.05 | T0mW | g1powermac: good afternoon. |
| 18:40.56 | T0mW | I tried 'git log --summary --follow --find-copies-harder -- include/stdFacpDefs.h' and it retuns nothing. |
| 18:41.17 | T0mW | however a 'git diff ' does show the last change I did. |
| 18:41.29 | T0mW | at least svn could do this. |
| 18:41.45 | prpplague | T0mW: hey bud! |
| 18:43.28 | mnemoc | prpplague: so no interest in joining? :( the first alpha eoma-68 boards (using allwinner a10 SoC) are going to be build the first week of february and been hobbiests-budget we really need to get it done right. some want to make a bleaglebone clone in this format but others argue TI won't sell chips or give documentation for anything under 100k units :-/ |
| 18:44.09 | T0mW | prpplague: hey, got a freakin' demo tomorrow been working like a fool for the past two weeks. |
| 18:45.04 | prpplague | mnemoc: i'll read up on it and see what i can suggest |
| 18:45.12 | prpplague | T0mW: hehe, btdt |
| 18:45.38 | av500 | mnemoc: you should be able to buy single bone CPUs easily |
| 18:45.42 | av500 | its a catalog part |
| 18:45.46 | av500 | unlike omap4 |
| 18:46.05 | mnemoc | prpplague: most documentation lives at http://rhombus-tech.net/ |
| 18:46.18 | prpplague | mnemoc: dandy |
| 18:46.46 | T0mW | prpplague: worked for four hours last night trying to get the RS485 communications working only to find out that I had to purge the memory on the other guys equipment before his side started working. |
| 18:46.56 | T0mW | prpplague: gak |
| 18:47.09 | T0mW | prpplague: :-) I need time off. |
| 18:47.44 | Mirell | is getting used to using Skype, Campfire, et cetera... |
| 18:47.55 | prpplague | T0mW: hehe |
| 18:47.57 | mnemoc | prpplague: thank you :) |
| 18:48.04 | prpplague | T0mW: yea, struggle with some DSI stuff myself |
| 18:48.29 | prpplague | mnemoc: sending an email to some folks about it |
| 18:49.05 | T0mW | prpplague: I just broke everything and git cannot recover due to a file rename. |
| 18:49.16 | Mirell | T0mW: What's the issue exactly? |
| 18:49.20 | prpplague | av500: i can't seem to get the data to match up correctly on this DSI port |
| 18:49.44 | prpplague | av500: it's like it is expecting packed RGB666 and i'm sending it RGB888 |
| 18:49.53 | mnemoc | av500: I'm just a software guy. Even when I believe the wisest is to make animalboard compatible stuff I can't change people minds about TI restrictions because my lack of knowledge :( |
| 18:49.54 | prpplague | T0mW: doh |
| 18:50.22 | mnemoc | prpplague: amazing, thank you VERY much |
| 18:50.32 | T0mW | prpplague: at least subversion could do this. |
| 18:50.39 | prpplague | mnemoc: basically there are a number of divisions in TI, the catalog or (ASP) side will generally sell you a single piece if you want |
| 18:50.49 | Mirell | T0mW: You're trying to see the log of a moved/deleted file? |
| 18:51.06 | prpplague | mnemoc: the Wireless Business Unit, or WBU, you generally have to have 100K for them to even burp in your presences |
| 18:51.32 | prpplague | mnemoc: right now, all the OMAP3 stuff is in ASP, all the OMAP4 stuff is in WBU |
| 18:51.52 | T0mW | Mirell: yeah, moved it with 'git mv from tohere' now 'git log --follow tohere' returns nothing. |
| 18:52.12 | Mirell | T0mW: And you commited the move, I assume. |
| 18:52.21 | mnemoc | so we can forget about a panda-alike EOMA-68/50 board but beagle should be doable |
| 18:52.40 | T0mW | Mirell: yes |
| 18:52.42 | av500 | mnemoc: you can forget about it for several reasons |
| 18:52.45 | prpplague | mnemoc: most companies like TI and Samsung will make requirements for large orders of their cutting edge products. once they have generally paid for themselves, they will move it to another division which will sell lower quantities |
| 18:53.13 | av500 | mnemoc: there is also the PMIC that you need and maybe the audio companion |
| 18:53.18 | av500 | and you are limited to POP |
| 18:53.20 | prpplague | mnemoc: the price for panda is subsidized |
| 18:53.22 | av500 | which is expensive |
| 18:54.12 | mnemoc | ic |
| 18:54.35 | Mirell | T0mW: How odd. Just worked for me when I did it in a test repo. |
| 18:54.35 | T0mW | Mirell: it won't show the logs through the rename, only the commits since the move. Tried 'git log --summary --follow --find-copies-harder -- include/stdFacpDefs.h' |
| 18:55.20 | av500 | mnemoc: and 0.4mm BGA which according to some people is impossible to produce |
| 18:55.38 | T0mW | crap |
| 18:55.45 | av500 | mnemoc: sizewise is might work: http://www.phytec.com/products/som/Cortex-A9/phyCORE-OMAP4460-OMAP4430.html |
| 18:56.18 | mnemoc | av500: but it would be much more expensive than the panda, right? |
| 18:56.58 | av500 | mnemoc: that would depend on TI pricing |
| 18:57.11 | av500 | which I cannot predict for such a use case |
| 18:58.45 | T0mW | Mirell: bah, you know, I was pretty tired last night, I probably borked this file's history. did a 'mv' instead of a 'git mv'. |
| 18:58.58 | Mirell | Le Woops |
| 18:59.14 | Mirell | You can still view the history of the old file |
| 19:01.43 | T0mW | naw |
| 19:02.31 | T0mW | ah |
| 19:03.15 | T0mW | Mirell: manually copied the file back into its old location, now git log works on it from the new location (where it had been). |
| 19:03.48 | Mirell | Huh. |
| 19:04.16 | prpplague | av500: what i hate most about DSI is there is no one to ask about it |
| 19:05.14 | T0mW | whew! |
| 19:06.11 | T0mW | prpplague: 17 hours left to resolve 'scheduling while atomic' :( |
| 19:06.17 | prpplague | hehe |
| 19:06.47 | T0mW | prpplague: nothing like a bug that intermittently borks the system. |
| 19:06.47 | av500 | prpplague: tell me about it |
| 19:07.02 | av500 | prpplague: we even had to add video mode |
| 19:07.20 | av500 | prpplague: plus a supplier that provided setup hex codes in a doc file |
| 19:07.37 | av500 | we saw like nothing for 3 weeks |
| 19:07.51 | prpplague | av500: yea i'm having problems with the video mode |
| 19:31.46 | *** join/#edev hieu (~hieu@222.254.184.194) |
| 19:36.53 | T0mW | rule #1 "do not move files with git" |
| 19:50.25 | prpplague | hehe |
| 19:51.05 | *** part/#edev hieu (~hieu@222.254.184.194) |
| 19:51.24 | kergoth | T0mW: git mv is no different from git rm + git add. it has no notion of renaming under the hood, it recognizes it after the fact. if you mv first, just git rm --cached oldname or git rm -f oldname, then git add the new one |
| 19:51.52 | landley | Mercurial has a notion of renaming under the hood. FYI. |
| 19:51.53 | kergoth | alternatively, git add -u will add all files already in the repo, including the removed file, which basically records the removal as though you used rm |
| 19:51.56 | kergoth | yeah |
| 19:52.01 | kergoth | important difference to be aware of |
| 19:52.39 | *** join/#edev hieuletrung (~hieu@222.254.184.194) |
| 19:53.19 | T0mW | kergoth: but, won't 'git rm ; git add' lose the history? |
| 19:53.24 | kergoth | no |
| 19:53.31 | kergoth | again, it doesn't record renames |
| 19:53.56 | kergoth | it recognizes that this commit removed this file with this content and added that file with 100% the same content as this, and sees that it was a rename |
| 19:54.01 | T0mW | kergoth: I assumed that 'git mv' would do the git rm; git add and keep the history intact on the new file? |
| 19:54.05 | kergoth | you just create a commit that does both a removal and an add |
| 19:54.10 | kergoth | again, it doesn't do anything special |
| 19:54.15 | kergoth | its exactly the same as rm + add |
| 19:54.32 | kergoth | you can follow the history either way, the behavior is identical |
| 19:54.36 | kergoth | use git log with --follow |
| 19:54.46 | T0mW | yea, but all trace of the old file was gone from the logs. |
| 19:54.52 | kergoth | hm? |
| 19:54.57 | kergoth | no, it doesn't lose any history at all |
| 19:55.09 | T0mW | sorry, meant revision. |
| 19:55.12 | kergoth | you're just not using --follow or -M/-C. log doesn't traverse renames by default |
| 19:55.36 | kergoth | again, mv is add + rm. add and rm both record changes in the index, and the commit will get the removal and the add |
| 19:56.09 | T0mW | I could find it in the log using 'git log --summary' (no filename used), but to try to revert (git revert xxxxxxxxxxx filename) would fail |
| 19:56.42 | kergoth | eh, revert reverts a commit, not af ile |
| 19:56.46 | kergoth | you're using the wrong command to revert |
| 19:56.46 | T0mW | I thought that I could resurrect the old file. |
| 19:56.57 | kergoth | git revert HEAD will do so, yes |
| 19:57.02 | kergoth | assuming you did the commit correctly |
| 19:57.04 | T0mW | I tried checkout |
| 19:57.12 | alex4nder- | put down the bunny |
| 19:57.23 | kergoth | http://progit.org/book/ch2-2.html - see the section on Moving Files |
| 19:57.42 | kergoth | further, see http://progit.org/book/ch2-4.html - "Undoing Things" |
| 19:57.50 | kergoth | etc |
| 19:57.51 | T0mW | kergoth: like I told Mirell, it was pretty late and I don't recall how I moved that file. thanks for your comments. |
| 19:57.55 | kergoth | np |
| 19:58.18 | T0mW | for the most part, I'm finding git more useful than svn. |
| 19:58.28 | kergoth | its possible you did an mv and committed just the removal of the old file or just the add of the new one, not both. |
| 19:58.33 | alex4nder- | T0mW: understatement of 2012 |
| 19:58.38 | kergoth | either would be seen as something other than a rename |
| 19:58.49 | T0mW | alex4nder-: heh |
| 19:59.17 | T0mW | alex4nder-: yes, you could really screw things up with svn. |
| 19:59.47 | T0mW | alex4nder-: git may be a pita right now to learn it, but it is consistent. |
| 20:00.43 | kergoth | I really like both mercurial and git for different reasons. they're so similar, but the differences are important |
| 20:00.50 | kergoth | (also: <3 hg-git) |
| 20:01.22 | alex4nder- | T0mW: subversion is the psycho suburban cheerleader girlfriend, who seems normal but when she flips out, she burns down your car |
| 20:01.47 | alex4nder- | git is the quirky urban art chick girlfriend, in dating her, you have to actually think. |
| 20:02.05 | alex4nder- | but once you get used to it, it's smooth sailing |
| 20:03.48 | alex4nder- | even if you can survive the cheerleader's ups and downs, she has no personality and eventually she gets fat, and you hate every minute you spend with her. |
| 20:06.34 | T0mW | alex4nder-: At one time, i was married to that woman. |
| 20:06.49 | alex4nder- | yah, she sucks |
| 20:07.00 | alex4nder- | urban art chick wins every time |
| 20:34.01 | Mirell | Heh |
| 20:34.05 | Mirell | Nice description. |
| 20:34.32 | Mirell | alex4nder-: Now make anime representations of all the SCM systems |
| 20:35.35 | alex4nder- | Mirell: yah, we should find someone with some actual talent to make that. |
| 20:41.11 | roxfan | post a request on 2ch |
| 20:54.50 | roxfan | 20:52:22 <@Chroche_Latel_Pastalie> i can't think of any designs other than prostitutes |
| 20:54.51 | roxfan | 20:52:37 <@Chroche_Latel_Pastalie> git repos are open for everyone and anyone who wants to use them after all |
| 20:54.53 | roxfan | 20:52:50 <+marcan> you can only pull from them, not push |
| 20:54.53 | roxfan | 20:52:58 <&hrm> so strippers then |
| 21:05.45 | Mirell | Hee |
| 21:08.25 | alex4nder- | roxfan: excellent |
| 21:34.15 | *** join/#edev dijenerate (~dijenerat@173.225.251.175) |
| 22:14.42 | *** join/#edev dijenerate (~dijenerat@173.225.251.175) |
| 22:49.30 | Mirell | Man, Erlang is weird. |
| 22:53.28 | g1powermac | Mirell, this might provoke a bit of a bias answer, but what database solution would you suggest if you want to track user usage on a site? |
| 22:53.52 | g1powermac | specifically for data mining usage later on |
| 22:54.33 | Mirell | Hmm. |
| 22:54.37 | Mirell | What sort of usage data? |
| 22:55.15 | g1powermac | basically how they interact with the site I'm designing |
| 22:56.15 | Mirell | If you're not concerned about high-availablity, CouchDB is rather nice. |
| 22:56.15 | g1powermac | of course the stuff they generate will go into a standard sql database, like comments and articles users write |
| 22:56.34 | g1powermac | but I'd like to track what is read, what is clicked, etc as well |
| 22:57.00 | Mirell | CouchDB is also good for storing document data |
| 22:57.10 | Mirell | Actually a lot better than a SQL database for things like that. |
| 22:57.16 | g1powermac | hmm |
| 22:57.53 | g1powermac | may mess with that later, rather not have to modify the CMS I chose that much at first :-) |
| 22:58.03 | Mirell | Hehe |
| 22:58.34 | Mirell | Don't think too much about it, as long as you don't put too much logic on the database side of things, like stored queries and whatnot |
| 22:58.49 | Mirell | You can always change types later |
| 22:58.55 | g1powermac | true |
| 22:59.38 | Mirell | The main power of a lot of the NoSQL databases, is that they use MapReduce |
| 23:01.55 | g1powermac | takes a closer look at CouchDB |
| 23:02.21 | g1powermac | at least it does HTTP/Rest. . .makes it simple since I want to use python :-) |
| 23:02.41 | Mirell | Aye |
| 23:11.34 | g1powermac | been eyeing MongoDB as well, any opinions on that one? |
| 23:13.20 | g1powermac | it seems to have massively more documentation available for it |
| 23:13.48 | g1powermac | there's a whole o'reilly book just one mongodb and python |
| 23:13.51 | g1powermac | *on |
| 23:14.37 | Mirell | I have no real opinion on it, other than it's in C++, and was boring to me because of that. |
| 23:14.43 | *** join/#edev grummund (~user@unaffiliated/grummund) |
| 23:15.14 | g1powermac | heh |
| 23:18.36 | g1powermac | hmm. . .think my firefox is about ready to blow up. . .gotta have 200+ tabs open. . . |
| 23:29.56 | kergoth | thats easy enough to resolve. how you do it depends ont he firefox version. easy to get it to not keep them all in ram at all times. load on demand |
| 23:30.55 | kergoth | originally it was done by the bartab extension. then in ff4 there was a max_concurrent_tabs in about:config you could set to 0 to at least get the bit where it doesn't load all tabs when restoring a session. recent ff does that bit automatically, and theres' an extension to unload the tabs if idle for a given period of time (bartab did this in the past also) |
| 23:31.44 | g1powermac | kergoth, oh, interesting |
| 23:32.39 | kergoth | if you're on ff8 or whatever, just restart, it won't re-load them all into ram when restoring the session, so that works when it's getting up there, if you don't want to install the unload when idle extension |
| 23:33.14 | kergoth | aside: ff is much, much better at handling large amounts of tabs than chrome, sadly. i wanted to switch to chrome, but was blocked by the tab thing, at least until I started heavily leveraging instapaper to avoid keeping them open all the time |
| 23:33.23 | kergoth | heh |
| 23:33.51 | g1powermac | yea, ff is getting better at large tab numbers |
| 23:34.12 | kergoth | at least when they aren't all in ram :) |
| 23:34.13 | g1powermac | is currently using 9.0.1 |
| 23:34.17 | g1powermac | yea |
| 23:34.37 | kergoth | i can't remember the name of the extension that unloads after a period of time. I'm sure you could find it by searching. my fiance has it installed in hers, but she's at work :) |
| 23:47.59 | *** join/#edev gustavoz (~gustavoz@host234.190-136-155.telecom.net.ar) |
| 23:55.22 | *** join/#edev uwe_ (~uwe_@dslb-088-066-186-052.pools.arcor-ip.net) |