2008-09-29 00:09 -!- RazvanM(~RazvanM@pool-151-196-118-156.balt.east.verizon.net) has joined #tux3 2008-09-29 00:12 flips, going through the code 2008-09-29 00:12 trying to figure out the unit tests 2008-09-29 00:12 some are obvious, others not so obvious 2008-09-29 00:17 file_bwrite: block write <0:0> 2008-09-29 00:17 ---- extent 0x0/7 ---- 2008-09-29 00:17 existing extents: 0x0 => 2/4; 0x5 => 9/2; 2008-09-29 00:17 ---- rewind to 0x0 => 2/4 ---- 2008-09-29 00:17 balloc extent -> [4/1] 2008-09-29 00:17 segs: 0x2/4 0x4/1 0x9/2 (3) 2008-09-29 00:17 dwalk_chop_after: 1 groups, 0 entries in last 2008-09-29 00:17 1 entry groups: 2008-09-29 00:17 0/0: 2008-09-29 00:17 group 0/1 at entry -1/0 2008-09-29 00:17 group 0/1 at entry 1/1 2008-09-29 00:17 group 0/1 at entry 3/2 2008-09-29 00:17 1 entry groups: 2008-09-29 00:17 0/3: 0 => 2/4; 4 => 4/1; 5 => 9/2; 2008-09-29 00:17 flush... Success 2008-09-29 00:17 woohoo, first time this ever worked 2008-09-29 00:17 rewrite a region of file containing two discontiguous extents 2008-09-29 00:18 it properly fills in the 1 block gap between them 2008-09-29 00:18 nice 2008-09-29 00:41 hmm, a bug in set_bits, that has to hurt 2008-09-29 00:54 ah, no it was a bug in balloc_extent_from_range 2008-09-29 01:42 ACTION resets his firefox home page from google.com to tux3.org 2008-09-29 01:45 -!- pgquiles(~pgquiles@166.Red-88-16-39.dynamicIP.rima-tde.net) has joined #tux3 2008-09-29 01:45 morning pgquiles 2008-09-29 01:46 hey flips 2008-09-29 01:47 I was thinking of you yesterday 2008-09-29 01:47 I'm touched ;) 2008-09-29 01:48 mkfs reserves 5% of the blocks for root, what's the use case for that? system crashes or is full and root comes to the rescue? 2008-09-29 01:48 I'm asking because that accounts for a whopping 50GB in my 1TB disk, which is a lot of seemingly wasted space :-/ 2008-09-29 01:48 attempt to avoid DoS of root 2008-09-29 01:48 by ordinary user, in absense of quotas 2008-09-29 01:49 but 5% is excessive 2008-09-29 01:49 on the other hand, it does need to scale with time as everything gets more bloated 2008-09-29 01:49 I guess 1GB would more more than enough 2008-09-29 01:49 50 MB would be enough 2008-09-29 01:49 probably 2008-09-29 01:50 I looked at the source of mkfs and was surprised to discover it accepts -mDOUBLE, I thought it only accepted an int 2008-09-29 01:50 ( I used -m1 and it still hurts!) 2008-09-29 01:51 you never know how many blocks sombody might want to reserve 2008-09-29 01:51 takes a double... 2008-09-29 01:52 yes, that's dumb but does no harm 2008-09-29 01:52 email ted and ask for a decimal point 2008-09-29 01:53 "avoids fragmnentation"... I doubt that 2008-09-29 01:54 I think what the man page means there is, performance gets so pathetically bad when the filesystem is 95%+ full that we just don't allow it 2008-09-29 01:54 think of it as a lameness tax ;-) 2008-09-29 01:54 we will try to make tux3 perform reasonably well even at 99% full 2008-09-29 01:54 going to be lots of work, hope you are ready to help 2008-09-29 01:55 send us some of your spanish dev buddies 2008-09-29 02:00 :-D 2008-09-29 02:00 I don't know anybody in Spain developing filesystems 2008-09-29 02:01 althought I must say the tux3 university is really useful 2008-09-29 02:01 I hope I'll have some spare time to catch up 2008-09-29 02:01 eventually :-) 2008-09-29 02:02 i think we need a video version of tux3 U someday 2008-09-29 02:02 hi pgguiles 2008-09-29 02:13 -!- pgquiles_(~pgquiles@62.43.226.52.static.user.ono.com) has joined #tux3 2008-09-29 03:07 -!- kbingham(~kbingham@193.132.141.186) has joined #tux3 2008-09-29 03:25 hello ! 2008-09-29 03:32 i finally understood what "versioning filesystem" really means :P 2008-09-29 03:33 how about saying it in your own words 2008-09-29 03:34 flips : its like when it keeps old copies of file or folder? 2008-09-29 03:35 partly 2008-09-29 03:38 ACTION keeps searching about "versioning filesystems" 2008-09-29 03:38 flips : today, Im gonna buy some books about that xD 2008-09-29 03:38 haha 2008-09-29 03:38 i have to learn about these things.. sounds interesting :P 2008-09-29 03:38 we're writing the books ;) 2008-09-29 03:39 source codes? 2008-09-29 03:41 orgthingy, did u understand the version pointer part? 2008-09-29 03:41 pranith : not yet 2008-09-29 03:41 about how maintaining versions this way is useful? 2008-09-29 03:41 hmm 2008-09-29 03:41 because once i want to learn something, i find out that i need to learn another thing 2008-09-29 03:41 complicated :P 2008-09-29 03:42 hmm, yeah 2008-09-29 03:42 kind of 2008-09-29 03:42 i had to read the document thrice 2008-09-29 03:42 before i got a semblence of understanding 2008-09-29 03:43 orgthingy, u from germany? 2008-09-29 03:43 peanitth : no, but im interested in german language 2008-09-29 03:43 :P 2008-09-29 03:44 ohk 2008-09-29 03:44 ACTION stares at http://www.ext3cow.com/Welcome_files/example1.jpg 2008-09-29 03:44 ACTION smiles 2008-09-29 04:06 -!- pgquiles__(~pgquiles@166.Red-88-16-39.dynamicIP.rima-tde.net) has joined #tux3 2008-09-29 04:30 flips, can u explain the fields in struct dwalk? 2008-09-29 04:31 pranith, sure 2008-09-29 04:31 specially about the gdict and edict fields 2008-09-29 04:31 they are just what is necessary for dwalk_next to work efficiently 2008-09-29 04:31 ok, what are they used for? 2008-09-29 04:31 needs to have a pointer to an extent and a group and an entry, and a limit for each to know when to go to the next 2008-09-29 04:32 exbase is a little different, the ->limit field of entry is relative to that 2008-09-29 04:32 and the "mock" fields are for calculating the finished size of a packed leaf, without actually writing the leaf 2008-09-29 04:32 i suppose exbase is the base of the extent 2008-09-29 04:33 yes 2008-09-29 04:33 i.e., the current extent's pointer 2008-09-29 04:33 no 2008-09-29 04:33 it's the lowest extent for an entire group of entries 2008-09-29 04:33 there is a description of the dleaf format somewhere 2008-09-29 04:34 ping shapor about it maybe 2008-09-29 04:34 he's the expert 2008-09-29 04:34 ohk, will do 2008-09-29 04:34 writing a mail.. that will better document it 2008-09-29 04:34 notices that the "limit" field of the entry and the "count" field of the group are only one byte, that goes a long way towards explaining some of the apparent complexity 2008-09-29 04:35 a dleaf index is highly compressed and therefore a little tricky to edit 2008-09-29 04:35 hmm 2008-09-29 04:35 yeah, u were trying to simplify that 2008-09-29 04:36 with an api... 2008-09-29 04:36 successfully, see the latest checkins 2008-09-29 04:36 filemap.c is now pretty obvious I think 2008-09-29 04:37 ok, how do i update using hg? i usually remove the tux3 folder and pull it again :( 2008-09-29 04:37 that works 2008-09-29 04:37 just hg pull will do it 2008-09-29 04:37 ok 2008-09-29 04:37 then hg update I think 2008-09-29 04:37 ok 2008-09-29 04:39 night 2008-09-29 04:40 gn 2008-09-29 04:42 flips : have you compared ext3cow to tux3 ? 2008-09-29 06:08 -!- pgquiles(~pgquiles@166.Red-88-16-39.dynamicIP.rima-tde.net) has joined #tux3 2008-09-29 07:16 -!- orgthingy_(~orgthingy@62.150.55.188) has joined #tux3 2008-09-29 07:28 -!- pgquiles(~pgquiles@166.Red-88-16-39.dynamicIP.rima-tde.net) has joined #tux3 2008-09-29 07:54 -!- smitht(~chatzilla@ool-182f94db.dyn.optonline.net) has joined #tux3 2008-09-29 07:57 -!- smitht(~chatzilla@ool-182f94db.dyn.optonline.net) has left #tux3 2008-09-29 10:08 -!- RazvanM(~RazvanM@dazzler.isi.jhu.edu) has joined #tux3 2008-09-29 10:18 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #tux3 2008-09-29 11:21 -!- SEJeff(~jeff__@66.151.59.138) has joined #tux3 2008-09-29 12:13 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #tux3 2008-09-29 13:06 -!- hirofumi(~hirofumi@210.171.168.39) has joined #tux3 2008-09-29 13:10 -!- orgthingy(~orgthingy@62.150.55.188) has joined #tux3 2008-09-29 14:16 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #tux3 2008-09-29 14:22 -!- kbingham(~kbingham@92.21.238.93) has joined #tux3 2008-09-29 15:15 bah 2008-09-29 15:15 ACTION is kind of annoyed by how sloppy the lockdep code is 2008-09-29 15:16 like folks never figured out in the Linux community to use a lot of small abstraction so that function bodies express things simply and are readable in that way as well 2008-09-29 15:16 it's just extra brain wankery that I could do without 2008-09-29 15:30 -!- orgthingy(~orgthingy@62.150.55.188) has left #tux3 2008-09-29 15:30 -!- orgthingy(~orgthingy@62.150.55.188) has joined #tux3 2008-09-29 15:30 ops 2008-09-29 16:01 -!- orgthingy(~orgthingy@62.150.55.188) has joined #tux3 2008-09-29 16:06 so, are you all developers in tux3? 2008-09-29 16:56 -!- orgthingy(~orgthingy@62.150.55.188) has joined #tux3 2008-09-29 18:02 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #tux3 2008-09-29 18:43 -!- tim_dimm(~timothyhu@cpe-76-90-98-247.socal.res.rr.com) has joined #tux3 2008-09-29 19:04 -!- tim_dimm(~timothyhu@cpe-76-90-98-247.socal.res.rr.com) has joined #tux3 2008-09-29 19:42 bh, ingo isn't big on abstraction, more a hack it now kinda guy 2008-09-29 19:42 has its uses 2008-09-29 19:43 bh, but you can always abstract it as it should be 2008-09-29 19:43 beauty of open source 2008-09-29 19:44 orgthingy, ext3cow looks like a fine project 2008-09-29 19:44 but it's missing a few essential things 2008-09-29 19:45 like: writable snapshots, snapshots of snapshots, deletion of snapshots 2008-09-29 19:46 it's very clever as far as it goes 2008-09-29 19:46 oh and limited to 2 TB 2008-09-29 19:47 like ext3 2008-09-29 19:47 well 2008-09-29 19:47 maybe they have increased that to 16TB 2008-09-29 19:47 still too small by today's standards 2008-09-29 19:48 -!- RazvanM(~RazvanM@pool-151-196-118-156.balt.east.verizon.net) has joined #tux3 2008-09-29 19:53 flips: trying to get my patches integrate, I fear, would be a pain 2008-09-29 19:53 if they're good developers will use them integrated or not 2008-09-29 19:53 they're for devs anyway 2008-09-29 19:54 yeah, looking at his code drives me up the fucking wall at times 2008-09-29 19:54 I mean, I'm use to it now after all of these years, but, man, it's massive mess for clean up for other developers 2008-09-29 19:54 like with the scheduler 2008-09-29 19:54 or rtmutex, et.c.. 2008-09-29 19:55 there's a certain point where you can't really do much other than hack it in a more limited way without a mass refactorng 2008-09-29 20:50 -!- Kirantpatil(~kiran@122.167.178.24) has joined #tux3 2008-09-29 20:50 -!- Kirantpatil(~kiran@122.167.178.24) has left #tux3 2008-09-29 21:02 -!- ajonat(~ajonat@190.48.120.169) has joined #tux3 2008-09-29 22:29 -!- tim_dimm(~timothyhu@cpe-76-90-98-247.socal.res.rr.com) has joined #tux3 2008-09-29 22:59 -!- tim_dimm(~timothyhu@cpe-76-90-98-247.socal.res.rr.com) has joined #tux3 2008-09-29 23:57 -!- bobby(~bobby@nat-inn.mentorg.com) has joined #tux3 2008-09-29 23:57 heya 2008-09-29 23:59 hello all, anyone here?