*** dyki has joined #ingres | 00:28 | |
*** Deyan has joined #ingres | 02:00 | |
Deyan | hi everybody | 02:00 |
---|---|---|
*** atrofast has quit IRC | 02:24 | |
* Deyan is reading INGRES docs again... | 02:43 | |
*** thiagomz has quit IRC | 03:03 | |
*** grantc has joined #ingres | 03:36 | |
*** ChanServ sets mode: +o grantc | 03:36 | |
*** rossand has joined #ingres | 03:45 | |
*** ChanServ sets mode: +o rossand | 03:45 | |
grantc | http://www.youtube.com/watch?v=ViUL769ilro .... | 03:49 |
grantc | morning all (btw) | 03:50 |
Deyan | meep | 03:52 |
*** grantc has quit IRC | 03:53 | |
*** grantc has joined #ingres | 03:53 | |
*** ChanServ sets mode: +o grantc | 03:53 | |
*** thiagomz has joined #ingres | 03:57 | |
pboro | ah, damn... a bug | 05:08 |
pboro | Or not. If I do "WHERE int NOT IN (SELECT col FROM foo)", does int match nulls? | 05:09 |
pboro | since it looks like it | 05:09 |
pboro | yea, if I change it to "WHERE int NOT IN (SELECT col FROM foo WHERE col IS NOT NULL)" it works as I expected | 05:10 |
grantc | nulls ... there be dragons | 05:15 |
grantc | or nothing :) | 05:15 |
pboro | checking the semantics right now :) | 05:15 |
pboro | at least with joins, NULL will never match any other value | 05:18 |
pboro | ah, Ingres works correctly and I work incorrectly :) | 05:19 |
*** Gerhard has joined #ingres | 05:22 | |
*** atrofast has joined #ingres | 05:34 | |
Deyan | guys, i have a question about ingres_query() (PHP)... | 06:02 |
Deyan | what is the rationale for raising an error if i give more arguments than it expects? | 06:02 |
Deyan | ie. if i have a query "INSERT ... (?,?,?)" | 06:02 |
Deyan | and i give an array with 10 elements to it | 06:02 |
Deyan | why make that an error? | 06:03 |
Deyan | use the first 3 elements, and discard the rest | 06:03 |
Deyan | right? | 06:03 |
pboro | that's not good | 06:04 |
pboro | it's always good to raise an error if the input does not meet the expectations | 06:04 |
pboro | it prevents programming errors | 06:04 |
Deyan | well, i often test | 06:04 |
Deyan | if i can add first few columns | 06:04 |
Deyan | now i must make a new array | 06:05 |
pboro | you can write a wrapper function to do it as you used to do | 06:05 |
Deyan | instead of giving one that is already correct for the full-blown insert statement... | 06:05 |
Deyan | pboro: yep | 06:05 |
Deyan | imho it should be just a warning... | 06:05 |
Deyan | if i give an array with more elements than expected | 06:05 |
Deyan | and error if i give less | 06:06 |
pboro | considering how badly PHP programmers take action on notices, I think it's good to raise an error :) | 06:06 |
pboro | but it's up to grantc | 06:06 |
pboro | if you talk nicely to him he might do something about it... or not ;D | 06:07 |
grantc | i am not sure allowing a 10 element array to be used when the driver is expecting 3 is a good thing | 06:41 |
Deyan | in fact, it is more in my case | 06:42 |
grantc | one could argue that perhaps I should only accept the last 3 elements of the array.. | 06:42 |
Deyan | i have a table with around 30 columns | 06:42 |
Deyan | because migration is kida complex | 06:42 |
Deyan | i do column by column | 06:42 |
grantc | if you know the names of the columns it's a no brainer to create a new array based on the key | 06:43 |
Deyan | that is what i do now | 06:43 |
Deyan | i need to make a new array | 06:43 |
Deyan | from the original one | 06:43 |
grantc | alternatively patches are welcome :) | 06:51 |
*** atrofast has quit IRC | 07:03 | |
*** atrofast has joined #ingres | 07:06 | |
thiagomz | guys,,, doubt... | 07:26 |
thiagomz | need I collect statistcs from my indexes ? | 07:26 |
Deyan | wow | 07:33 |
Deyan | this VectorWise looks promising | 07:34 |
magnusg| | Absolutely | 07:37 |
*** magnusg| has left #ingres | 07:37 | |
*** Alex| is now known as Alex|off | 07:51 | |
*** Gerhard has quit IRC | 07:52 | |
dyki | it looks like Ingres will kick Oracle's ass yet again! | 07:54 |
dyki | good stuff :) | 07:54 |
*** FrankW has joined #ingres | 07:55 | |
grantc | thiagomz, you need to generate proper statistics so your indexes get used | 07:56 |
Deyan | is VectorWise in SVN? | 08:10 |
grantc | Deyan, i've not seen any vw branches | 08:14 |
grantc | i know our people are working with their people but where the code is being managed is another issue | 08:15 |
grantc | Deyan, see http://community.ingres.com/forum/26959-post6.html | 08:19 |
rossand | Deyan: I can confirm you will not find VectorWise code in the publicly available svn for now. | 08:19 |
rossand | Marcin's post shows the intention. | 08:20 |
rossand | Deyan: one more thing. I wanted to share. I was not able to reproduce the scenario you talked about the other day. I did a createdb, created a table, destroyed it. Re-ran this many times with no sleep 2 as you had done. No issues found. Perhaps you might share the logs with us or ideally your script so we can reproduce the issue? | 08:22 |
pboro | wow... vectorwise really promises quite a bit :) | 08:25 |
*** rossand1 has joined #ingres | 08:27 | |
Deyan | grantc: i am reading forum all the time | 08:30 |
*** rossand has quit IRC | 08:32 | |
Deyan | rossand1: sorry for late reply | 08:51 |
Deyan | rossand1: i do not have time for that :( | 08:51 |
rossand1 | Deyan: Sorry, I can't help you then. We'll have to hope someone else comes across the issue. | 08:51 |
*** rossand1 is now known as rossand | 08:52 | |
rossand | I'll keep my eye out for it of course. | 08:52 |
Deyan | rossand: thank you - i did not say i expect any help with it | 08:55 |
rossand | Deyan: You're very welcome. Helping you or someone else be successful with Ingres helps us so it's comfortable pragmatism. | 09:00 |
FrankW | The need to build lots of stuff in 32bit mode on 64bit linux has been pulling my hair out! | 09:01 |
pboro | FrankW, you only need 32-bit gcc + gcclibs and 32-bit xerces :) depending on your distro, it may be difficult or not... | 09:02 |
FrankW | apparently I also need the 32bit pam-dev package, though I'm not positive about that. | 09:03 |
pboro | FrankW, hmm, that's possible, since latest versions include pam support... | 09:04 |
pboro | FrankW, which distro are you using? | 09:04 |
FrankW | I installed the 32bit xerces but it did not get picked up in the build - now I'm trying to substitute it the low level build directory. | 09:04 |
FrankW | I'm on gutzy gibbon (7.10) | 09:04 |
pboro | FrankW, 7.10 is pretty old... which ingres version, latest from svn? | 09:04 |
FrankW | Part of the xerces problem may be that I'm using my own build of xerces27 64bit in /usr/local. | 09:04 |
FrankW | latest from svn, yes. | 09:04 |
pboro | it compiles dead simple on a up to date ubuntu :) | 09:05 |
FrankW | I'm following the guides at http://community.ingres.com/wiki/Building_ingres_on_ubuntu and https://help.ubuntu.com/community/Servers/Ingres but it is still a bit hair raising. | 09:05 |
* rossand feels FrankW's pain. | 09:06 | |
rossand | We're chipping away at it. Build is one of those things that's hard to get prioritized as no customer will pay money and it's not very sexy. :-) | 09:06 |
FrankW | I can understand that! | 09:07 |
FrankW | Is there a reason parts are built in 32bit mode? I'm getting the impression it is a bunch of utility tools built this way? | 09:07 |
rossand | It's one of those things that affects everyone every day so it gets done one way or another. | 09:07 |
FrankW | Link /wrk/home/warmerda/ingres/server/build/utility/lp32/unimapcompile | 09:08 |
FrankW | /usr/bin/ld: cannot find -lstdc++ | 09:08 |
FrankW | Hmm. | 09:08 |
atrofast | Did you get gcc-multilib AND g++-multlib, FrankW? | 09:08 |
atrofast | Both are required for Ingres | 09:09 |
FrankW | I imagine I'm missing g++-multilib. | 09:09 |
FrankW | back to dig around on launchpad. | 09:09 |
rossand | To spare your hair, we could give you a VMWare VM? | 09:09 |
FrankW | I'd much prefer to get to where I can build it locally. | 09:10 |
FrankW | What I ought to do is download a modern ubuntu dvd when I'm on the fast internet some day. | 09:10 |
pboro | FrankW, sounds like a plan :) | 09:10 |
rossand | I can give you one if you'd like? Already downloaded. | 09:10 |
rossand | I run Jaunty (9.04) on a few of my machines. FC10 on the others. | 09:11 |
atrofast | I run Fedora11... It's actually easier to build Ingres on as compare dto F10 due to it having xerces 2.7 in yum | 09:11 |
FrankW | unimapcompile built! Now on to the next problem. | 09:13 |
Deyan | i run fedora 11 at home | 09:16 |
Deyan | it rocks | 09:16 |
FrankW | cp: cannot create regular file `/wrk/home/warmerda/ingres/server/build/ice/icetool/access.gif': No such file or directory | 09:22 |
FrankW | It seems like some subdirectories are not getting created. | 09:22 |
FrankW | I had to manually create build/ice/DTD to get past the previous step. | 09:22 |
FrankW | Is there a way of running a clean build? | 09:22 |
grantc | jam -a clean | 09:23 |
grantc | from $ING_SRC | 09:23 |
rossand | OR, runbuild -c | 09:23 |
rossand | runbuild is a wrapper around the steps to build by hand. | 09:23 |
FrankW | thanks, done. I haven't set any environment variables yet so I'll do it from runbuild. | 09:24 |
rossand | runbuild protects you from environment variables just in case. It'll reset any variables necessary. | 09:26 |
rossand | Some folks switch back and forth often and it could get messy otherwise. | 09:26 |
FrankW | Are the geospatial components built as part of the standard build from svn? | 09:27 |
atrofast | If you checked it out from http://code.ingres.com/ingres/branches/geospatial | 09:28 |
FrankW | I assume that branch holds a complete ingres tree then? | 09:29 |
rossand | FrankW: yes | 09:30 |
rossand | It's a branch off main | 09:30 |
* FrankW pulls down a new tree. | 09:32 | |
* FrankW wonders if he has an ingres community userid, and what it might be. | 09:38 | |
rossand | hmm, it may be possible for one of us with LDAP access to grep for you. Of course we'd need a bit of a clue of what it might be to match against. | 09:39 |
*** Deyan has quit IRC | 09:47 | |
FrankW | warmerda? It would be nice if the "forgot my password" tool let you give an email address. | 09:47 |
*** DerMeister has joined #ingres | 09:52 | |
rossand | brb, powering down for a bit to let the storm blow over. | 09:57 |
FrankW | np | 09:58 |
*** rossand1 has joined #ingres | 09:58 | |
atrofast | Crazy thunder here, might lose power at any point :P | 10:06 |
FrankW | grantc has provided my userid, so I'm good now. | 10:06 |
atrofast | Great! Thanks grantc | 10:06 |
grantc | no problem - i am knocking up a dirty application to do queries against the back end. i know we have an ldap tool but still | 10:07 |
FrankW | I ended up writing something like that for the OSGeo LDAP (http://www.osgeo.org/cgi-bin/ldap_web_search.py) | 10:08 |
atrofast | Yeah sometimes something small and simple is way better than a complete and fully featured app :) | 10:09 |
*** Gerhard has joined #ingres | 10:15 | |
*** rossand has quit IRC | 10:15 | |
*** rossand has joined #ingres | 10:18 | |
*** ChanServ sets mode: +o rossand | 10:18 | |
*** Dejan has joined #ingres | 10:24 | |
Dejan | dyki, dude | 10:24 |
Dejan | :) | 10:25 |
Dejan | he guys | 10:28 |
* Dejan is home | 10:28 | |
*** dyki has quit IRC | 10:46 | |
*** Gerhard has quit IRC | 10:46 | |
*** Alex|off is now known as Alex| | 11:16 | |
*** DerMeister has quit IRC | 11:36 | |
*** DerMeister has joined #ingres | 11:36 | |
*** mull has joined #ingres | 11:46 | |
*** rossand1 has joined #ingres | 12:06 | |
*** rossand has quit IRC | 12:24 | |
*** clach04 has quit IRC | 13:22 | |
*** clach04 has joined #ingres | 13:23 | |
*** atrofast has quit IRC | 13:49 | |
*** DerMeister has quit IRC | 14:14 | |
*** Alex| is now known as Alex|off | 14:17 | |
*** Dejan has quit IRC | 16:22 | |
*** rossand1 has quit IRC | 17:32 | |
* FrankW is defeated by geos link issues building the geospatial branch. | 18:45 | |
FrankW | I'll revisit this tomorrow. | 18:45 |
*** rossand has joined #ingres | 18:47 | |
*** ChanServ sets mode: +o rossand | 18:47 | |
*** rossand has quit IRC | 19:32 | |
*** mull has quit IRC | 21:14 | |
*** Alex|off is now known as Alex| | 22:37 | |
*** FrankW has quit IRC | 22:45 | |
*** Alex| is now known as Alex|off | 23:11 | |
*** magnusg| has joined #ingres | 23:18 | |
*** Alex|off is now known as Alex| | 23:26 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!