| 00:22.28 | *** join/#brlcad KimK (~Kim__@wsip-184-176-200-171.ks.ks.cox.net) | |
| 00:34.41 | *** join/#brlcad dtidrow (~dtidrow@c-68-84-167-135.hsd1.mi.comcast.net) | |
| 00:52.02 | *** join/#brlcad KimK (~Kim__@wsip-184-176-200-171.ks.ks.cox.net) | |
| 01:14.56 | *** join/#brlcad Notify (~notify@66-118-151-70.static.sagonet.net) | |
| 01:23.03 | *** join/#brlcad KimK (~Kim__@wsip-184-176-200-171.ks.ks.cox.net) | |
| 01:52.39 | *** join/#brlcad KimK (~Kim__@wsip-184-176-200-171.ks.ks.cox.net) | |
| 02:23.05 | *** join/#brlcad KimK (~Kim__@wsip-184-176-200-171.ks.ks.cox.net) | |
| 03:15.06 | Notify | 03BRL-CAD Wiki:Carpinteyrorcz * 0 /wiki/User:Carpinteyrorcz: |
| 03:40.48 | *** join/#brlcad agkphysics (~aaron@206.10.252.27.dyn.cust.vf.net.nz) | |
| 03:54.31 | *** join/#brlcad aaron (~aaron@206.10.252.27.dyn.cust.vf.net.nz) | |
| 04:11.30 | *** join/#brlcad EricPoehlsen (~Eric@dslb-084-056-082-235.pools.arcor-ip.net) | |
| 04:13.46 | *** join/#brlcad Guest24238 (~aaron@206.10.252.27.dyn.cust.vf.net.nz) | |
| 05:43.41 | Notify | 03BRL-CAD:brlcad * 54210 (brlcad/trunk/src/conv/step/DerivedUnitElement.cpp brlcad/trunk/src/conv/step/DerivedUnitElement.h and 2 others): fixed bad header. looks like 'United' regex gone mad. |
| 05:44.30 | brlcad | starseeker: your src/librt/test_* files seem to be lacking their header (sh/header.sh will fix it for you) |
| 06:12.29 | Notify | 03BRL-CAD:brlcad * 54211 brlcad/trunk/src/tclscripts/mged/botedit.tcl: add missing header and footer |
| 06:14.29 | Notify | 03BRL-CAD:brlcad * 54212 brlcad/trunk/sh/template.sh: if /bin/sh isn't a bash shell, footer.sh will fail (due to the array usages). make sure we're using bash. clean up backup recovery outputs too. |
| 06:19.29 | Notify | 03BRL-CAD:brlcad * 54213 (brlcad/trunk/src/tclscripts/boteditor/botEditor.tcl brlcad/trunk/src/tclscripts/boteditor/botPropertyBox.tcl brlcad/trunk/src/tclscripts/boteditor/botTools.tcl): add missing headers and footers |
| 06:23.44 | Notify | 03BRL-CAD:brlcad * 54214 brlcad/trunk/src/tclscripts/archer/itk_redefines.tcl: missing header/footer added |
| 06:27.11 | Notify | 03BRL-CAD:brlcad * 54215 brlcad/trunk/src/libsysv/tcl.c: add missing header, remove per-file authorship per dev guide |
| 06:29.40 | Notify | 03BRL-CAD:brlcad * 54216 brlcad/trunk/sh/copyright.sh: ignore cmake build directories and svn dirs. |
| 06:43.08 | *** join/#brlcad EricPoehlsen (~Eric@ip-109-84-173-38.web.vodafone.de) | |
| 07:22.22 | Notify | 03BRL-CAD:brlcad * 54217 (brlcad/trunk/CMakeLists.txt brlcad/trunk/COPYING and 2735 others): update copyright for 2013 welcoming our files to the year of slithering slathering serpent subjects |
| 07:48.45 | Notify | 03BRL-CAD Wiki:033-compare-online-loans-no-credit-check-approval * 0 /wiki/User:033-compare-online-loans-no-credit-check-approval: |
| 07:56.14 | *** join/#brlcad d_rossberg (~rossberg@66-118-151-70.static.sagonet.net) | |
| 13:15.35 | *** join/#brlcad EricPoehlsen (~Eric@ip-109-85-12-50.web.vodafone.de) | |
| 15:29.36 | brlcad | tightens up mediawiki to not allow dashes in username and usernames > 12char |
| 15:30.11 | brlcad | still don't know why the MinimumNameLength extension isn't catching those |
| 15:30.27 | brlcad | modified the core User.php file to impose harder limits |
| 15:31.01 | Notify | 03BRL-CAD Wiki:Jacksixb * 4942 /wiki/User:Jacksixb/Independent_Study_2012-13_Log: |
| 15:33.54 | brlcad | starseeker: any hints on how I can fix this? http://pastebin.ca/2308596 |
| 15:34.44 | brlcad | at least what the right fix is, clearly caused by not finding fontconfig package, but is the fix to empty out the var so it's not NOTFOUND? |
| 15:36.42 | maths22 | brlcad: why did you want to do that? |
| 15:38.02 | maths22 | lastlog - |
| 15:38.06 | maths22 | sorry |
| 15:42.46 | maths22 | it ist to prevent spam users? |
| 15:56.57 | brlcad | yep |
| 15:58.03 | brlcad | maths22: I installed an extension to limit usernames, but it doesn't seem to prevent them fully |
| 15:58.13 | brlcad | note the recent.. |
| 15:58.13 | brlcad | 02:48 < Notify> BRL-CAD Wiki:033-compare-online-loans-no-credit-check-approval * 0 /wiki/User:033-compare-online-loans-no-credit-check-approval |
| 15:59.29 | brlcad | we're also still getting 1-2 spams a day, so I'm looking for more measures that don't impede fly-by-editing |
| 15:59.44 | brlcad | our regex is getting pretty huge |
| 16:04.48 | maths22 | true. |
| 16:05.42 | brlcad | we had it down to about 1-2 a month, which was quite acceptable |
| 16:06.00 | brlcad | but then hundreds of spam accounts were getting created, nice to limit that too |
| 16:06.35 | brlcad | an option to double-verify new accounts would be nice.. |
| 16:08.32 | ``Erik | disable new page creation for non-verified users? |
| 16:09.38 | maths22 | what do you mean double-verify |
| 16:11.02 | ``Erik | a common approach is to require an email address and send an email there with a validation link to 'enable' an account, would that be a useful approach? |
| 16:11.23 | maths22 | Let me look into that. |
| 16:11.26 | brlcad | users must already provide an e-mail address I "think" |
| 16:11.38 | brlcad | haven't tried to create a new account in a while.. ;) |
| 16:12.13 | brlcad | requiring an account for new pages would be good if that's not already enforced |
| 16:12.26 | brlcad | still want to allow anonymous edits to any page |
| 16:13.15 | maths22 | by they half to confim it. |
| 16:13.19 | maths22 | Let me look |
| 16:13.50 | ``Erik | I'd imagine some of the spam is generated mechanical turk style :/ |
| 16:14.43 | brlcad | the majority are rather complex but at least half-automated |
| 16:15.09 | brlcad | there was one that clearly got posted in error a while back that showed the regexes the spam script was using |
| 16:15.25 | brlcad | it had it's own internal regexes to get past filtering |
| 16:16.05 | ``Erik | cool, didja save it anywhere? be interesting to see the red teams secret technologies in this arms race |
| 16:16.18 | brlcad | (buy|purchse|sell) (car|boat|house) (loans|investments|estate) on the best (property|deal|goal) ... |
| 16:16.42 | brlcad | i'm sure it's in the wiki history somewhere, but no not specifically |
| 16:16.55 | brlcad | just gave me a couple dozen patterns to add to our regex ;) |
| 16:17.13 | maths22 | I have deactivated page creation for unregistered users |
| 16:17.21 | brlcad | maths22: awesome |
| 16:18.06 | brlcad | what was REALLY effective was disabling all URLs ... |
| 16:18.16 | maths22 | what do you mean. |
| 16:18.22 | brlcad | but of course then we couldn't even edit our own existing pages |
| 16:18.37 | brlcad | the regex pattern match extension |
| 16:18.44 | brlcad | I modified it to match any url |
| 16:18.50 | ``Erik | if text contains "http://" reject |
| 16:18.58 | brlcad | lil more complex, but yeah |
| 16:19.16 | brlcad | "http:\/\/.*\.com|http:\/\/.*\.net|http:\/\/.*[^cC][^aA][^dD]\.org|". # any link to an external site |
| 16:19.46 | maths22 | that would make using our own site hard. |
| 16:19.48 | brlcad | but that was a no-go, way too restrictive |
| 16:20.19 | brlcad | what WOULD be useful is if a submit contains a url, that it prompts an additional captcha |
| 16:20.33 | brlcad | even better, if it contains a *new* url :) |
| 16:20.38 | ``Erik | 'cept it was fine for the mountains of spam to http://mrastercard.com.cn *cough* :) |
| 16:20.42 | brlcad | but I'm not aware of such an extension |
| 16:21.19 | ``Erik | .cn is china |
| 16:21.25 | brlcad | ``Erik: that'd still get caught by http:\/\/.*\.com |
| 16:21.31 | ``Erik | ah, hm |
| 16:21.48 | brlcad | but yeah, some isoteric matches *could* get through, just unlikely :) |
| 16:22.06 | ``Erik | registers brlcad.xxx before it becomes a hot property O.o |
| 16:22.09 | brlcad | myawesomespamcad.org would get in |
| 16:22.48 | brlcad | and any other tld actually not containing ".com", ".net", or "cad.org" |
| 16:23.12 | brlcad | they're just currently not even in the noise |
| 16:24.24 | maths22 | look at this: http://www.mediawiki.org/wiki/Extension:QuestyCaptcha |
| 16:26.42 | brlcad | maths22: we already have that installed |
| 16:26.50 | brlcad | via ConfirmEdit |
| 16:27.18 | brlcad | I think that captcha is only enabled during an edit |
| 16:27.29 | brlcad | pretty effective at edit spam |
| 16:28.25 | brlcad | hell, most fly by users would have to do some research to answer the questions I put in there... :) |
| 16:28.30 | brlcad | e.g., What is the dot product of [2, 12, 1] with [7, 2, 4] |
| 16:30.40 | maths22 | Why are we allowing all registered users to skip the captcha |
| 16:30.54 | maths22 | We should only let autoconfirmed users skip it. |
| 16:31.00 | brlcad | because I don't want to burden a real user with a real edit |
| 16:31.17 | brlcad | there's that distinction? |
| 16:31.20 | maths22 | what about this: $wgGroupPermissions['emailconfirmed']['skipcaptcha'] = true; |
| 16:31.39 | brlcad | s/there/where/ |
| 16:32.07 | brlcad | that sounds good to me, I assumed you couldn't be a registered user without confirming your e-mail |
| 16:32.10 | maths22 | Sorry. Autoconfirmed means: |
| 16:32.31 | maths22 | never mind. |
| 16:32.40 | maths22 | What we should do is set http://www.mediawiki.org/wiki/Manual:$wgAutoConfirmAge |
| 16:32.53 | maths22 | and/or http://www.mediawiki.org/wiki/Manual:$wgAutoConfirmCount |
| 16:34.36 | maths22 | or rather http://www.mediawiki.org/wiki/Manual:$wgAutoPromote |
| 16:34.59 | brlcad | autoconfirmage doesn't look very useful |
| 16:35.19 | brlcad | much of the spam accounts are registered several days before they start spamming |
| 16:35.30 | brlcad | and they are then only used to spam once or twice |
| 16:35.52 | brlcad | probably to combat exactly that |
| 16:36.13 | maths22 | then use autoconfirmcount |
| 16:36.32 | brlcad | yeah, that sounds useful |
| 16:36.34 | maths22 | ignore my last link |
| 16:36.38 | brlcad | what's a good value? |
| 16:36.40 | brlcad | 3? |
| 16:37.03 | brlcad | effectively 0 now right? |
| 16:37.23 | maths22 | yes. |
| 16:37.29 | maths22 | 3 to five sounds good |
| 16:37.44 | brlcad | yeah, a few captchas when they're just getting started sounds reasonable |
| 16:37.52 | maths22 | ok. |
| 16:38.32 | brlcad | more than 3 is probably not preventing anything -- if they get past even one, it's a real human and they'll be able to get past N |
| 16:40.34 | maths22 | true |
| 16:40.47 | maths22 | I have set that |
| 16:42.00 | brlcad | what about captcha during account creation? |
| 16:42.10 | brlcad | does it currently prompt? |
| 16:42.50 | maths22 | It already does |
| 16:44.44 | brlcad | but not the questy one, I imagine? |
| 16:45.56 | maths22 | yes. it is the questy one |
| 16:46.04 | brlcad | huh |
| 16:46.34 | maths22 | I remember I had to think about the sum of elements in a zero matrix. |
| 16:46.41 | brlcad | unrelated, any way to return users that log in from drupal to the drupal page they logged in from? |
| 16:46.54 | maths22 | Let me look at that |
| 16:47.53 | brlcad | if you all have other good CADy mathy matrixy questions to include, lemme know ;) |
| 16:58.11 | ``Erik | <-- kinda wonders if a dot product isn't too much, plenty of people do modelling and purdee pikchurz without doing that much math regularly O.o |
| 16:58.58 | ``Erik | basic 4 operations might be good, I've seen those used as captchas before |
| 17:02.06 | brlcad | ``Erik: yet a simple search reveals how to compute a dot product with trivial ease |
| 17:02.13 | brlcad | even online calculators |
| 17:03.03 | brlcad | I don't think we have a single user that couldn't figure it out if they had to |
| 17:03.14 | ``Erik | yeah, it's easy to do, but I'd guess that a lot of non-geeks wouldn't think to google it *shrug* |
| 17:03.20 | brlcad | and they get a different captcha if they fail anyways |
| 17:03.59 | ``Erik | (not that the contribution rate is high among new/driveby users *shrug*) |
| 17:04.38 | brlcad | it's non-zero, so that's in part why I still want to support anonymous editing at all |
| 17:04.42 | brlcad | I think it works |
| 17:04.58 | brlcad | we've had a half dozens anon edits in the past two months |
| 17:05.25 | ``Erik | aight, be interesting to collect the data from the old machine on contribution rates and compare it to that kind of captcha |
| 17:05.43 | ``Erik | d'no if there's enough to do a statistically significant a/b test :/ |
| 17:06.01 | brlcad | questions in our domain that require thought are okay by me, just can't be "hard" questions (as in not easily searched or understood) |
| 17:07.01 | ``Erik | sooo, not "provide a ponynomial time algorithm to the traveling salesman problem in haskell" |
| 17:09.20 | brlcad | yeah |
| 17:09.26 | brlcad | haskell would be too much ;) |
| 17:13.58 | maths22 | let's see if autoconfirm + no anonymous pages helps |
| 17:14.27 | brlcad | that's no anonymous NEW pages right? |
| 17:14.35 | brlcad | still want to allow anonymous edits on existing pages |
| 17:14.45 | ``Erik | how's this for mental sickness: http://paste.lisp.org/display/134995 (sh script for cron) |
| 17:16.15 | brlcad | what for? |
| 17:16.57 | ``Erik | in case data accidently gets stomped |
| 17:17.14 | brlcad | okie |
| 17:18.05 | brlcad | I like to keep a local and offsite backup of dumps, have it scripted up for the mysql dbs |
| 17:18.17 | brlcad | was what was filling up .bz's hard disks a couple years ago if you remember |
| 17:19.08 | ``Erik | yeah, I had a nightly script on forge that did db dumps before an fs dump, good to have a replayable script |
| 17:51.12 | *** join/#brlcad EricPoehlsen (~Eric@dslb-084-056-082-235.pools.arcor-ip.net) | |
| 18:32.34 | brlcad | pokes Notify |
| 18:59.38 | brlcad | http://pastebin.ca/2308674 |
| 19:02.50 | ``Erik | last email notify got was at 11:48:15 and was posted O.o |
| 19:03.12 | ``Erik | var is fun |
| 19:03.14 | ``Erik | full |
| 19:05.34 | Notify | 03BRL-CAD:tbrowder2 * 54221 (brlcad/trunk/doc/docbook/resources/brlcad/images/logo-vm-gears.png brlcad/trunk/doc/docbook/resources/brlcad/images/logo-vm-gears.svg): change to 128x128 per web site |
| 19:11.07 | brlcad | ahh, /var/log is growing now that we're all on it |
| 19:11.24 | brlcad | heh, web logs |
| 19:11.36 | brlcad | already 3GB |
| 19:12.26 | ``Erik | mebbe move 'em to /backup or /usr |
| 19:12.49 | ``Erik | d'no what the plans are for /backup, but it's a different spindle |
| 19:13.22 | brlcad | yeah. no plans |
| 19:13.45 | brlcad | it came that way, we can make it be anything but a level 0 backup of the main disk is actually kinda useful |
| 19:14.09 | brlcad | of course, that's half a tb, which we could use if space gets tight |
| 19:14.41 | brlcad | probably just move var/log to usr/var/log like the other dirs for now |
| 19:14.50 | ``Erik | ghetto raid-1 done with dd or something? |
| 19:14.50 | brlcad | s/move/link/ |
| 19:14.55 | brlcad | yeah |
| 19:15.03 | brlcad | rsync it |
| 19:15.06 | ``Erik | or cpio once it's got the first image |
| 19:15.36 | brlcad | sync it weekly with no deletes or similar |
| 19:18.12 | ``Erik | the notify posts should show up over the next few hours in random order as smtp retries come in |
| 19:19.00 | ``Erik | assuming the 47 free megs don't get used up first O.o |
| 19:22.17 | ``Erik | turns off the hourly tarball regeneration of the git mirror.. should slow down the logs a bit |
| 19:32.15 | Notify | 03BRL-CAD:r_weiss * 54222 brlcad/trunk/src/tclscripts/mged/lc.tcl: Update to the mged "lc" command adding the "-r" option to skip listing regions which their parent is a region and the child is subtracted within the parent. Also added the 'los' test to the "-s" option. |
| 19:32.29 | brlcad | lets see how hard this hits the fan |
| 19:34.15 | brlcad | waits on apachectl stop && mv /var/log /usr/var/log && ln -s /usr/var/log && apachectl start |
| 19:36.58 | brlcad | done, now hopefully the mail spool will be a lil more resiliant |
| 19:36.58 | ``Erik | all of /var/log, not just /var/log/httpd ? |
| 19:37.08 | brlcad | whole deal |
| 19:37.16 | brlcad | user logs are huge once they get re-enabled |
| 19:37.50 | ``Erik | could cause boot issues :/ |
| 19:38.07 | brlcad | hrm, how so? shouldn't care.. |
| 19:38.42 | brlcad | I've had /var/log redirected elsewhere without a problem iirc |
| 19:39.27 | ``Erik | ah, syslogd isn't started until thhe FILESYSTEMS and criticalremotes |
| 19:39.31 | ``Erik | aight, cool |
| 19:40.06 | brlcad | hups syslogd just for good measure |
| 19:41.27 | starseeker | brlcad: I think the fontconfig fix looks OK |
| 19:41.39 | ``Erik | lsof seems happy |
| 19:42.11 | starseeker | mutters under his breath some more about Tcl/Tk... |
| 19:43.53 | brlcad | maths22: do you recall when you submitted your first task (or first popped on irc)? |
| 19:45.43 | brlcad | i'm seeing nov I think |
| 19:47.15 | ``Erik | heh, var is now 3% used |
| 19:48.31 | Notify | 03BRL-CAD:brlcad * 54218 brlcad/trunk/src/other/tk/CMakeLists.txt: also need to account for a non-existent fontconfig |
| 19:48.51 | Notify | 03BRL-CAD Wiki:Lajosgodly * 0 /wiki/User:Lajosgodly: |
| 19:55.26 | Notify | 03BRL-CAD:brlcad * 54223 brlcad/trunk/AUTHORS: credit burroughs (maths22) for his awesome work getting most of our website infrastructure migrated to newer hardware and fixing/improving numerous website features (upgrades, search, unified login, spam filtering, more). came in under GCI but interested in taking over website ops. |
| 20:03.31 | Notify | 03BRL-CAD:n_reed * 54224 (brlcad/trunk/src/librt/librt_private.h brlcad/trunk/src/librt/primitives/pipe/pipe.c brlcad/trunk/src/librt/primitives/primitive_util.c): replace pipe draw routine struct parameters & returns with pointer parameters to avoid struct copies |
| 20:14.00 | Notify | 03BRL-CAD:brlcad * 54225 brlcad/trunk/src/other/openNURBS/opennurbs_array.h: gcc 4.4.6 warns about being unable to inline the (implicit) destructor. declaring the destructor implies to not inline, making the warnings/errors go away. performance should be moot, destructor is ON_SimpleArray's virtual. |
| 20:25.08 | Notify | 03BRL-CAD:tbrowder2 * 54220 (brlcad/trunk/doc/docbook/resources/brlcad/images/logo-vm.svg =================================================================== and 12 others): rename file to match that used on the site |
| 20:25.17 | *** join/#brlcad luca79 (~luca@net-2-35-211-123.cust.dsl.teletu.it) | |
| 20:28.06 | Notify | 03BRL-CAD:brlcad * 54219 (brlcad/trunk/src/other/boost/boost/detail/container_fwd.hpp brlcad/trunk/src/other/boost/boost/mpl/has_xxx.hpp): apply boost patch regarding error: "BOOST_MPL_HAS_XXX_NO_EXPLICIT_TEST_FUNCTION" is not defined along with a handful of other defines not defined. see https://svn.boost.org/trac/boost/ticket/6142 for some relevance. |
| 20:35.27 | Notify | 03BRL-CAD:brlcad * 54226 brlcad/trunk/src/other/openNURBS/opennurbs_layer.h: more inline warnings/errors due to the compiler attempting to inline default functions. declaring them quells via intent. |
| 20:40.13 | maths22 | brlcad: November 27 |
| 20:40.35 | maths22 | day after GCI started |
| 20:41.05 | maths22 | yes. No autonomous new pages |
| 20:50.24 | Notify | 03BRL-CAD:carlmoore * 54227 brlcad/trunk/src/shapes/coil.c: simplify 2 more derivations of coil_radius |
| 21:07.06 | Notify | 03BRL-CAD:brlcad * 54228 (brlcad/trunk/src/other/openNURBS/opennurbs_array.h brlcad/trunk/src/other/openNURBS/opennurbs_layer.h): revert 54226 and 54225 since they cause vtable symbol failures when linking against the lib (declaring them apparently also implied don't define the default. |
| 21:11.53 | *** join/#brlcad luca79 (~luca@net-2-35-211-123.cust.dsl.teletu.it) | |
| 21:19.53 | Notify | 03BRL-CAD:carlmoore * 54229 brlcad/trunk/src/shapes/coil.c: remove commented-off code because of the simplicity of the replacement |
| 21:30.33 | Notify | 03BRL-CAD:carlmoore * 54230 brlcad/trunk/src/shapes/coil.c: add decimal points although this should already compute in floating point mode |
| 21:36.54 | Notify | 03BRL-CAD:carlmoore * 54231 brlcad/trunk/src/conv/iges/extrudcon.c: simplify computation of q1 |
| 22:39.29 | Notify | 03BRL-CAD:carlmoore * 54232 brlcad/trunk/src/conv/shp/shapelib/shpopen.c: simplify what I take to be integer-mode calculations |