| 00:09.10 | starseek1r | ``Erik: looks like Notify is AWOL again |
| 00:22.20 | *** join/#brlcad vkoyjyndunudbbmx (~armin@dslb-092-074-250-073.092.074.pools.vodafone-ip.de) |
| 01:01.17 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 01:19.40 | *** join/#brlcad archivist (~archivist@host81-149-189-98.in-addr.btopenworld.com) |
| 02:51.59 | brlcad | maths22: sudoers reloaded |
| 02:52.11 | maths22 | brlcad: Thanks! |
| 02:52.30 | maths22 | Having permissions to edit files makes life a lot easier :) |
| 02:53.22 | brlcad | maths22: ``Erik: .bz's etc files are in /etc.bz/etc/ and /usr/local/etc.bz/etc for reference -- intentionally did not overwrite the new machine's files so we can weed out some unnecessary customization |
| 02:53.38 | brlcad | but leaving them there for reference for a while is a good thing |
| 02:54.23 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 02:57.53 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 03:02.33 | maths22 | brlcad: Installed nano, htop |
| 03:02.44 | maths22 | htop is beautiful: 16 cores on top |
| 03:02.48 | *** join/#brlcad kintel_ (~kintel@unaffiliated/kintel) |
| 03:02.49 | brlcad | maths22: cool, thx |
| 03:03.07 | brlcad | yeah, this new machine is really sweet compared to the old |
| 03:03.29 | maths22 | What is the reason for making everything one partition? |
| 03:03.32 | brlcad | we're paying a pretty penny for it, but got a pretty good deal |
| 03:03.40 | maths22 | brlcad: Cool |
| 03:03.40 | brlcad | that's just how it came |
| 03:04.00 | maths22 | brlcad: Makes sense |
| 03:04.17 | brlcad | fyi, keeping a log of major sysadmin actions in /etc/motd until everything is done and set up |
| 03:04.26 | brlcad | feel free to update it as you do stuff |
| 03:04.47 | brlcad | so ``Erik and myself will know, don't yet have any other admins to worry about |
| 03:06.28 | brlcad | maths22: you installed those via ports or pkg? |
| 03:06.39 | maths22 | Ports |
| 03:06.47 | maths22 | I thought that is what we normally use |
| 03:07.46 | maths22 | Should I note any dependencies that got automatically installed? |
| 03:08.07 | brlcad | nah |
| 03:08.26 | brlcad | just make sure you are looking at the lastest version when editing before saving |
| 03:08.33 | maths22 | OK |
| 03:08.35 | brlcad | so you don't clobber my notes that I just added ;) |
| 03:09.21 | maths22 | brlcad: Could you transfer the password for travis-upload? |
| 03:09.37 | brlcad | it's there |
| 03:10.27 | maths22 | Great! I didn't know if the service account passwords got transfered? |
| 03:10.34 | maths22 | s/?// |
| 03:10.35 | brlcad | copy it from etc.bz/etc/master.passwd to /etc/master.passwd, then run: pwd_mkdb -p /etc/master.passwd |
| 03:11.21 | brlcad | the entry is there, just not the password hashes -- want to make sure that we only readd as needed (like for this, service accounts) |
| 03:11.34 | brlcad | do you know the format of /etc/master.passwd? |
| 03:11.53 | maths22 | brlcad: I don't |
| 03:12.04 | maths22 | Could you explain so I don't break things |
| 03:12.58 | brlcad | colon delimited, second column (after the first : up to the second : colon) is a long password hash encoded |
| 03:13.35 | brlcad | that's how passwords are stored, so keep them private, but you can cut and paste them from the old to the new for one account |
| 03:13.55 | maths22 | brlcad: OK. |
| 03:14.08 | brlcad | I can do it if you feel uncertain, but it's good to learn how traditional unix accounts are managed (it's pretty simple really) |
| 03:15.06 | maths22 | brlcad: Done! |
| 03:16.00 | maths22 | travis-upload won't work until I get more stuff set up, but it allows me to upload builds from travis/appveyor without any risk to the rest of the system |
| 03:16.21 | brlcad | checks |
| 03:16.56 | brlcad | looks good! |
| 03:17.20 | brlcad | undoubtedly a lot of web stuff is broken ... LOTS of ports things missing |
| 03:17.51 | brlcad | even the main config isn't yet getting read for some reason, didn't investigate |
| 03:20.38 | maths22 | brlcad: How do we go about testing the web stuff? |
| 03:24.01 | maths22 | BRL-CAD Release 7.25.0, Build 20160415 |
| 03:24.01 | maths22 | Elapsed compilation time: 2 minutes 15 seconds |
| 03:24.04 | brlcad | I update dns and we see if it works ;) |
| 03:24.15 | brlcad | usually on a subdomain first |
| 03:25.51 | brlcad | in brlcad.org's case, we have the other tld's we can switch over first |
| 03:25.57 | brlcad | (brlcad.net and .com) |
| 03:26.06 | brlcad | does some of those now |
| 03:39.44 | maths22 | brlcad: Web configs don't work because we hardcoded the source IP |
| 03:40.05 | brlcad | ah, right -- that's necessary |
| 03:40.29 | brlcad | that's an easy fix |
| 03:40.52 | brlcad | okay, .net and com point to new server |
| 03:41.23 | brlcad | .com also needed nameserver updates, so it'll take a bit longer to propagate |
| 03:41.46 | maths22 | OK |
| 03:42.12 | maths22 | Should the ip be fixed with sed in all the configs or manually? |
| 03:42.31 | brlcad | perl will do a better job for this |
| 03:42.48 | brlcad | perl -pi -e 's/OLDIP/NEWIP/g' FILE |
| 03:43.05 | maths22 | Just curious-why is that better than sed? |
| 03:43.13 | brlcad | perl will do the edit in place |
| 03:43.21 | brlcad | sed requires redirecting then copying over |
| 03:43.39 | maths22 | Isn't that just 'sed -i' |
| 03:44.16 | brlcad | ah yeah, I suppose .. :) |
| 03:45.08 | brlcad | learned sed a decade or two before they finally added that option, perl's always had in place edits (at least as long as I can remember) |
| 03:46.03 | maths22 | Makes sense |
| 03:46.08 | maths22 | I never knew a sed without it |
| 03:46.23 | maths22 | Should I go ahead and apply it to all the configs? |
| 03:46.29 | brlcad | one of those things that got hardwired as "doesn't do that" .. I'll eventually remember, maybe :) |
| 03:46.37 | brlcad | sure |
| 03:47.24 | brlcad | then the only advantage is that perl's regex features are generally quite a bit more advanced, and it's a different regex engine |
| 03:47.55 | brlcad | irrelevant for simple search and replace of course |
| 04:03.02 | maths22 | brlcad: I guess we need to set up mysql |
| 04:06.50 | brlcad | yeah, that reminds me ... I kicked off db dumps yesterday |
| 04:07.10 | brlcad | the db files themselves were copied straight up |
| 04:07.22 | brlcad | so may just need to install mysql and start it |
| 04:08.01 | brlcad | unless some config or the mysql db gets overwritten on install |
| 04:08.40 | maths22 | I'm installing right now, so we will see soon |
| 04:09.53 | brlcad | same version? |
| 04:10.22 | brlcad | looks like we're on 14.14 |
| 04:10.57 | brlcad | moves the web root etc files into place |
| 04:11.43 | brlcad | huh, already done |
| 04:12.00 | brlcad | ah, you did it :) |
| 04:16.49 | brlcad | starseeker: since saying "full size" would require the caller to consult documentation to know what full means, is there another way to phrase it that would be self-documenting? |
| 04:20.15 | brlcad | maybe rt_db_internal_size() and rt_db_external_size() |
| 04:22.37 | brlcad | with that, it'd clearly cover the two potential sizes unambiguously -- the function you added would be the internal |
| 04:29.24 | brlcad | I get that the intent of 'full' is the expansion to include referenced objects and you have an option that is not expanded too, so that I would suggest be handled with a flag to the internal size function |
| 04:32.18 | brlcad | how to expose the two to the user ... looks like find already has a convention, just -size [type] |
| 04:37.04 | brlcad | how about "-size n[ckrMGTP]" with that 'r' (or perhaps R) flag meaning the recursive size similar to c, becomes a compatible extension |
| 04:38.02 | brlcad | you're just trying to make release harder, aren't you :) |
| 04:48.38 | *** join/#brlcad shubham (a5e1683c@gateway/web/freenode/ip.165.225.104.60) |
| 05:26.00 | *** join/#brlcad ``Erik_ (~erik@pool-100-16-40-20.bltmmd.fios.verizon.net) |
| 05:36.52 | *** join/#brlcad Mandeep_Singh (~Mandeep@117.207.152.230) |
| 05:40.38 | *** join/#brlcad archivist (~archivist@host81-149-189-98.in-addr.btopenworld.com) |
| 05:41.14 | *** join/#brlcad KimK (~Kim__@ip68-102-66-31.ks.ok.cox.net) |
| 05:58.34 | maths22 | brlcad: mysql is up and running |
| 05:58.42 | maths22 | I had to retransfer it from the old server |
| 06:13.34 | maths22 | brlcad: https://paste.ee/p/ZWFjV |
| 06:16.32 | maths22 | For reference: https://paste.ee/p/wxMqD |
| 06:32.40 | *** join/#brlcad Mandeep_Singh (~Mandeep@117.220.151.195) |
| 06:48.45 | *** join/#brlcad teepee_ (~teepee@unaffiliated/teepee) |
| 07:12.24 | *** join/#brlcad KimK (~Kim__@ip68-102-66-31.ks.ok.cox.net) |
| 08:31.57 | *** join/#brlcad Mandeep_Singh (~Mandeep@117.199.96.10) |
| 09:00.47 | *** join/#brlcad Mandeep_Singh (~Mandeep@59.96.143.78) |
| 09:45.09 | *** join/#brlcad Mandeep_Singh (~Mandeep@59.96.143.78) |
| 09:53.27 | Notify | 03BRL-CAD:brlcad * 67631 brlcad/trunk/doc/STRATEGY: clean up the overview documentation to more carefully frame our strategic work in a positive light |
| 09:55.19 | Notify | 03BRL-CAD:starseeker * 67632 brlcad/trunk/TODO: Eeek - something wrong with rtcheck/gdiff -R |
| 09:55.23 | Notify | 03BRL-CAD:starseeker * 67633 brlcad/trunk/TODO: Cancel that (provisionally) - working with very small objects, overlaps may be below threshold |
| 09:56.43 | Notify | 03BRL-CAD:starseeker * 67634 brlcad/trunk/doc/docbook/system/mann/draw.xml: Document the -L option to draw. Note in passing that there seem to be a whole lot of draw options that aren't documented... |
| 09:58.24 | Notify | 03BRL-CAD:starseeker * 67635 (brlcad/trunk/include/ged/objects.h brlcad/trunk/src/libged/CMakeLists.txt brlcad/trunk/src/mged/setup.c): Start playing around with the notion of a stat command for objects. May or may not make sense... |
| 09:59.04 | Notify | 03BRL-CAD:starseeker * 67636 (brlcad/trunk/include/rt/db_io.h brlcad/trunk/src/libged/stat.c brlcad/trunk/src/librt/db5_io.c): Move the size calculator to a librt function - will have more than one use for this calculation. |
| 09:59.06 | Notify | 03BRL-CAD:starseeker * 67637 (brlcad/trunk/NEWS brlcad/trunk/doc/docbook/system/mann/search.xml and 2 others): Add -size and -fsize filters to the search command. The -size option will filter based on individual object sizes. -fsize, by comparison, filters based on the 'full' defintion of the object based on its hierarchy - i.e., for a comb, it will filter based on that combs size summed with the sizes of all the |
| 09:59.08 | Notify | objects below the comb in the hierarchy. These options are intended to help users quickly narrow down what the 'big' objects are in a .g file and what hierarchy definitions might be combining to produce large objects. |
| 09:59.10 | Notify | ... |
| 09:59.57 | Notify | 03BRL-CAD Wiki:Eddysalazar * 0 /wiki/User:Eddysalazar: |
| 10:16.40 | *** join/#brlcad gaganjyot (~gaganjyot@122.173.147.28) |
| 10:37.08 | *** join/#brlcad Mandeep_Singh (~Mandeep@117.214.207.228) |
| 11:53.20 | starseeker | brlcad: if it's a concern for release, by all means yank it - I just tossed that and the beginnings of stat in to be able to find big things in a .g more quickly/easily - my only other known options were to draw and see how long draws took or do a bazillion keeps and look at the .g files on the filesystem |
| 11:53.56 | starseeker | generally speaking, search options don't themselves take options - it's like name and iname. What about size and rsize for "recursive size"? |
| 11:55.00 | starseeker | I'm fine with those function names - I'll switch it later this morning |
| 11:57.02 | starseeker | or if we're doing a flag to the internal function, rt_db_size(int type) maybe? |
| 11:58.13 | starseeker | i guess that'd actually be rt_db_size(struct db_i *dbip, struct directory *dp, int type) |
| 12:37.33 | brlcad | starseeker: many find options take options, so was that stuff yanked when it was adapted? |
| 12:37.56 | brlcad | cursory glance looked like all the code for handling options is still in there |
| 12:39.28 | brlcad | I got no problem with size+rsize options, just one of the points of -size on find isn't to report the size (-ls does that), it's to filter on size |
| 12:40.05 | brlcad | so in the name of compatibility and expected behavior, I would expect it to at least take a number... :) |
| 12:41.04 | brlcad | yeah .. -depth options take args.. |
| 12:42.22 | brlcad | common use case: search . -name *.bot -size +1M -ls |
| 12:43.12 | brlcad | i.e., find all bots greater than 1MB, print a table listing of those objects |
| 12:50.43 | brlcad | search . -size -4000c #<== find objects less than 4000 bytes |
| 13:27.25 | *** join/#brlcad teepee` (bc5c2133@gateway/web/freenode/ip.188.92.33.51) |
| 13:28.25 | *** join/#brlcad jasleen (~jasleen@117.255.208.72) |
| 14:18.18 | Notify | 03BRL-CAD:starseeker * 67638 brlcad/trunk/src/other/CMakeLists.txt: mark as advanced |
| 14:18.54 | starseeker | brlcad: the options take args, but they usually don't have options of their own... |
| 14:20.03 | starseeker | isn't familiar with the -ls option |
| 14:20.04 | starseeker | hmm |
| 14:20.50 | starseeker | brlcad: maybe I'm missing something - -size is a filter |
| 14:21.51 | starseeker | search . -name *.bot -size >1000000 should do your use case, modulo -ls |
| 14:22.16 | starseeker | I don't have anything to take the "friendly" size specifications like 1M, although I'd like to... |
| 14:22.46 | starseeker | -size and -rsize are filters, not reporting options |
| 14:23.46 | starseeker | tries the find -ls option... |
| 14:23.49 | starseeker | oh, I see |
| 14:23.58 | starseeker | yeah, we don't have that reporting option currently |
| 14:24.26 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 14:25.34 | starseeker | brlcad: I'm not 100% sure if dp->d_len is what I think it is - keeping things results in .g files with sizes not at all close to what a tree-walk summation of the d_len values reports |
| 14:26.26 | starseeker | the other problem is that the tree walk size will report the xpushed size currently, since I don't recognize uniqueness of objects... |
| 14:26.30 | starseeker | hmm |
| 14:26.38 | starseeker | that's a real problem, actually... |
| 14:26.51 | starseeker | and could explain the size difference |
| 14:35.13 | maths22 | brlcad: I updated the homepage to use relative URLs |
| 14:37.29 | maths22 | Also, brlcad.net works (I temporarily disabled the redirect so we could test it on the new server) |
| 15:19.22 | *** join/#brlcad ishwerad1in (~ishweradm@servers.ishwerdas.com) |
| 15:22.16 | *** join/#brlcad kintel_ (~kintel@unaffiliated/kintel) |
| 15:24.38 | *** join/#brlcad ``Erik (~erik@pool-100-16-40-20.bltmmd.fios.verizon.net) |
| 15:29.27 | *** join/#brlcad KimK (~Kim__@ip68-102-66-31.ks.ok.cox.net) |
| 15:54.47 | *** join/#brlcad poxip (~poxip@unaffiliated/mrpoxipol) |
| 16:41.28 | Notify | 03BRL-CAD:starseeker * 67639 (brlcad/trunk/include/rt/db_io.h brlcad/trunk/src/libged/stat.c and 2 others): go with db5_size and a flag for recursive |
| 16:41.58 | *** join/#brlcad Mathnerd314 (~quassel@supertux/Mathnerd314) |
| 16:46.34 | Notify | 03BRL-CAD:starseeker * 67640 (brlcad/trunk/doc/docbook/system/mann/search.xml brlcad/trunk/src/librt/search.c brlcad/trunk/src/librt/search.h): Go with rsize rather than fsize for the hierarchy aware size filter. |
| 16:47.38 | Notify | 03BRL-CAD:starseeker * 67641 brlcad/trunk/NEWS: Added recursive size calculating filter to search |
| 16:49.18 | Notify | 03BRL-CAD:starseeker * 67642 brlcad/trunk/NEWS: Note that the new -size filter for search is a filter that will select based on indiviual object size, not a printing option to report size. |
| 16:52.30 | *** join/#brlcad kintel (~kintel@unaffiliated/kintel) |
| 17:09.33 | *** join/#brlcad Mathnerd314 (~quassel@supertux/Mathnerd314) |
| 17:30.31 | brlcad | starseeker: ah, hm maybe just a feature missing from the version of find you started with ... |
| 17:30.52 | brlcad | I see that posix only dictates [-+]#[c] |
| 17:31.17 | brlcad | so http://pubs.opengroup.org/onlinepubs/009695399/utilities/find.html |
| 17:31.33 | *** join/#brlcad Mathnerd314 (~quassel@supertux/Mathnerd314) |
| 17:31.43 | *** join/#brlcad ``Erik (~erik@pool-100-16-40-20.bltmmd.fios.verizon.net) |
| 17:31.54 | brlcad | curiously doesn't say anything about <=> |
| 17:33.06 | brlcad | SUSv2 is the same< http://pubs.opengroup.org/onlinepubs/007908799/xcu/find.html |
| 17:34.57 | brlcad | looks like the other suffixes are specific to bsd and linux find, not the standard |
| 17:35.32 | brlcad | still no mention of <> though -- is that something you introduced or did the find you used have that? |
| 17:42.24 | brlcad | maths22: I see, awesome! did mysql just come up (so all db's are available) or did you just restore the one needed? |
| 17:43.15 | *** join/#brlcad archivist (~archivist@host81-149-189-98.in-addr.btopenworld.com) |
| 17:43.24 | maths22 | All of them (but see my note in motd) |
| 17:43.48 | brlcad | starseeker: gut feeling is that it'll be at least a week of tweaking and bug fixing to get -size/-rsize working right ... I think it will make sense to hold it from release |
| 17:46.50 | brlcad | great feature, very useful, but it'll take some investment to make sure it's correct |
| 17:47.11 | brlcad | especially for all cases like search is supposed to handle (e.g., recursive size of an extrude, instances in a comb, etc) |
| 17:47.55 | brlcad | there's also how to report disk size vs in-memory size (if they're drastically different) |
| 17:48.14 | *** join/#brlcad Mandeep_Singh (~Mandeep@117.214.207.228) |
| 17:52.12 | *** join/#brlcad witness_ (uid10044@gateway/web/irccloud.com/x-xsxmfsvyrlohdrlh) |
| 17:57.40 | starseeker | brlcad: I introduced ><= - I wasn't aware of the [-+]#[c] bit |
| 17:58.16 | starseeker | brlcad: ok, I'll back it out in a single commit so it can be re-introduced after release |
| 18:00.05 | starseeker | yeah, I think db5_size needs special reporting for extrude, dsp, submodel and possibly vol |
| 18:01.41 | starseeker | we'll probably want some variation on an -ls option to get the size info reported in a search result - we've never needed it previously, but we will once we add this |
| 18:05.09 | *** join/#brlcad gaganjyot (~gaganjyot@122.173.244.25) |
| 18:07.46 | Notify | 03BRL-CAD:starseeker * 67643 (brlcad/trunk/include/ged/objects.h brlcad/trunk/src/libged/CMakeLists.txt brlcad/trunk/src/mged/setup.c): Back out stat command |
| 18:08.21 | Notify | 03BRL-CAD:starseeker * 67644 (brlcad/trunk/NEWS brlcad/trunk/include/rt/db_io.h and 3 others): Back out size/rsize search filters. |
| 18:10.42 | starseeker | brlcad: I wasn't following the posix find docs all that closely for search - if you want to tack more closely to the standard it's probably worth a systematic audit of the commands. I find the <=> notation more intuitive than +-, but that may just be me... |
| 18:11.12 | starseeker | s/commands/filters and options |
| 18:12.27 | Notify | 03BRL-CAD:starseeker * 67645 (brlcad/trunk/NEWS brlcad/trunk/include/rt/db_io.h and 3 others): shoot, forgot doc changes |
| 18:14.21 | Notify | 03BRL-CAD:starseeker * 67646 (brlcad/trunk/NEWS brlcad/trunk/doc/docbook/system/mann/search.xml and 4 others): Back out size/rsize search filters (this time including docs). |
| 18:19.52 | starseeker | brlcad: ok, we can reverse merge r67646 once we're done with the release |
| 18:20.06 | *** join/#brlcad teepee (~teepee@unaffiliated/teepee) |
| 18:52.57 | *** join/#brlcad tflendrich (~tflendric@eac26.neoplus.adsl.tpnet.pl) |
| 19:21.45 | *** join/#brlcad Mathnerd314_ (~quassel@supertux/Mathnerd314) |
| 19:22.41 | Notify | 03BRL-CAD:brlcad * 67647 brlcad/trunk/TODO: note to reinstate stat and search -size next release. very useful. started in commits r67636 through r67646. |
| 19:29.22 | *** join/#brlcad witness_ (uid10044@gateway/web/irccloud.com/x-uqsutahexjiazved) |
| 19:53.45 | *** join/#brlcad Zitara (~vsvs@188.236.201.98) |
| 19:58.15 | *** join/#brlcad Zitar (bcecc962@gateway/web/freenode/ip.188.236.201.98) |
| 19:59.57 | Zitar | Hello |
| 20:00.45 | Zitar | Using xcode as an IDE, should I use build all? Like the defult thing? |
| 20:10.54 | *** join/#brlcad Zitara (~vsvs@188.236.201.98) |
| 20:25.03 | Notify | 03BRL-CAD:starseeker * 67648 (brlcad/trunk/src/tclscripts/rtwizard/lib/FbPage.itk brlcad/trunk/src/tclscripts/rtwizard/lib/PictureTypeA.itcl and 5 others): Spaces in output file paths were causing trouble. Also, if the render failed, we were ending up with zombie fbserv processes. Note: at some point refactor this file output logic into a common proc... |
| 20:41.55 | starseeker | brlcad: I wonder if there's any way, in addition to the on-disk vs. in-mem size questions, to also assess and report the memory impact of the draw and prep of various objects. I doubt it's workable as a search option, but maybe in the stat command we could do something to figure that out of a user specifically requests it? |
| 20:43.14 | starseeker | would be handy, for example, to be able to get a feel for how much memory a NURBS tessellation is taking up at the current settings... maybe we could report it only for things that are currently drawn/prepped... |
| 20:46.27 | starseeker | s/that out of/that out if/ |
| 21:12.20 | *** join/#brlcad teepee_ (~teepee@unaffiliated/teepee) |
| 21:21.50 | *** join/#brlcad shubham_ (71c189bf@gateway/web/freenode/ip.113.193.137.191) |
| 23:18.34 | Notify | 03BRL-CAD:starseeker * 67649 (brlcad/trunk/NEWS brlcad/trunk/doc/docbook/system/mann/draw.xml and 3 others): This implements a user request for an option (-M) to the draw command that skips adding subtraction solids to the wireframe. Useful for objects that have lots of large subtractions that contribute in a minimal fashion to the final shape. |