*** dyki has joined #ingres | 00:29 | |
*** Deyan has joined #ingres | 01:06 | |
Deyan | meepo | 01:06 |
---|---|---|
pboro | Deyan, I was able to compile 9.2 on openSUSE 11.1 64-bit, will be publishing instructions today | 01:09 |
Deyan | perfect! | 01:09 |
Deyan | thank you in advance | 01:09 |
*** Gerhard has joined #ingres | 01:21 | |
magnusg| | Hi all, I'm trying to buld Ingres on Fedora 8 but it doesn't find the command/program buildrel. Anyone having an idea? | 01:58 |
*** dyki has quit IRC | 01:58 | |
grantc | hi magnusg|, i am guessing it failed to get built | 02:02 |
grantc | did you log the jam build process? | 02:02 |
magnusg| | grantc, I just realized that it could be built during the process and not being there in advance. I'll look in the logs and see what I can find. | 02:03 |
grantc | src/front/st/install_unix - run jam there and you might see why | 02:03 |
magnusg| | Probably something wrong with the environmental setup. I complains about not finding buildrel.c but it's the so I guess something is broken. I use a virtual machine I got from Andy in Punta Cana and it worked fine to begin with but I've had problems with it later. | 02:10 |
magnusg| | I'll take it from the beginning with the risk that I might learn something in the process :-) | 02:11 |
grantc | :) | 02:22 |
grantc | the process has improved a bit since then | 02:22 |
grantc | if you checkout http://code.ingres.com/ingres/main then source buildtools/set_env.sh you should be good to go | 02:23 |
magnusg| | The problem seems to having to do with obsolete VERS file. I've changed it will test again. | 02:23 |
grantc | alternatively there is a script that does the build and install | 02:23 |
*** dyki has joined #ingres | 02:35 | |
Deyan | hey dyki ! \o/ | 02:44 |
Deyan | :D | 02:45 |
Deyan | long time no see | 02:45 |
Deyan | grantc: \o/ | 02:45 |
magnusg| | grantc, there was an old VERS file which caused the first problem but even with that fixed the build fail. Seems like problems with libs (...skipped /home/ingres/ingres/server/build/lib/libcuf.a for lack of /home/ingres/ingres/server/build/lib/libcuf.a(ccmsrc.o)...). I'll investigate further after lunch. | 02:57 |
pboro | gnaah, I have hard time figuring how Jam works, I have both 32-bit and 64-bit xerces installed but still getting "don't know how to make lp32/libxerces-c.so.28.0" | 02:58 |
Deyan | same here | 03:03 |
Deyan | guys, i think i will start doing CMake building for Ingres | 03:04 |
pboro | I can fix that by copying the file manually, but it's not a proper solutions | 03:04 |
pboro | -s | 03:04 |
pboro | oh well, may I'll write an extra jam rule to do that if I can't figure out it... | 03:04 |
pboro | wow, got if fixed | 03:05 |
pboro | I had a typo in my fix... oh well :) | 03:05 |
pboro | ah, instructions are ready, I'll try that they works from the beginning :) | 03:06 |
grantc | fwiw this seems to work for me on ubuntu vis a vis xerces-c | 03:10 |
grantc | http://ingres.pastebin.com/f7dd44be0 | 03:10 |
pboro | grantc, that helps for figuring out the version correctly, but there is a problem in 9.2's Jamrules for hybrid builds | 03:11 |
grantc | pboro, ah | 03:11 |
pboro | it doesn't include both 32-bit and 64-bit xerces lib, so I took changeset 22 from svn which addresses it | 03:11 |
grantc | 22 ? from the beginning? | 03:13 |
pboro | yeah | 03:13 |
grantc | mmm | 03:13 |
pboro | r22 | hanje04 | 2008-04-23 17:31:39 +0300 (ke, 23 huhti 2008) | 1 line | 03:13 |
pboro | #109 Fix up build for x86_64 Linux. Add in-line assembler macros to csnormal.h to replace those previously defined in asm/bitops.h | 03:14 |
pboro | that one | 03:14 |
pboro | linking of 32-bit unimapcompile still fails on my 64-bit hybrid build since it's looking for 32-bit in wrong place (/usr/lib64), after that's fixed it should build fine | 03:21 |
* Deyan still thinks it wrong to do the hybrid build | 03:22 | |
grantc | Deyan, why is it wrong? | 03:22 |
Deyan | people should be able to do the 64bit build only or 32bit | 03:22 |
Deyan | and do two separate builds | 03:23 |
grantc | it allows people who have built 32bit apps to run against a 64bit server | 03:23 |
Deyan | if they want both | 03:23 |
Deyan | if i want 32bit version, i will run the build process again | 03:23 |
Deyan | with different parameters | 03:23 |
grantc | it's most likely possible since certain platforms only have a 64 bit release | 03:27 |
grantc | but it's easier for Ingres to have a single build which has both since it cuts down the time needed to release | 03:28 |
grantc | also we use to have a single byte build and a double byte build now they too are combined | 03:28 |
pboro | ahh, the problem was in openSUSE's libraries, not in jam :) | 03:28 |
grantc | partly for the same reason | 03:28 |
pboro | Hmm... Xerces is compiled using GCC 4.2 and libstdc++.so.6, and I'm building Ingres with GCC 3.3 and libstdc++.so.5... I wonder if this is going to be a problem :/ | 03:31 |
Deyan | why with so old gcc? | 03:34 |
pboro | the INSTALL file told to use 3.3 | 03:34 |
Deyan | aha... | 03:35 |
Deyan | 3.3 is just too old :) | 03:35 |
Deyan | (IMHO) | 03:35 |
pboro | why? it works fine :) | 03:35 |
pboro | anyway, compiler can easily be switched (if the source code works with newer gcc...) | 03:37 |
grantc | pboro, it should build with the newer gcc compiler | 03:38 |
grantc | i use 4.3.3 | 03:39 |
pboro | grantc, ok, I'll try that once I get it compiled with 3.3 | 03:39 |
pboro | this is 9.2 build 134 anyway | 03:39 |
pboro | not the trunk... | 03:39 |
grantc | not sure it matters :) | 03:39 |
pboro | yeah, we'll see :) | 03:39 |
grantc | i am not sure i have built from the internal tree for 9.2 on ubuntu but i have done it with suse 11.0 and it built ok | 03:43 |
grantc | using 4.2 iirc | 03:43 |
Deyan | guys, is there possibility to tell INGRES to release all locks, and cancel all transactions ? | 03:49 |
Deyan | i am trying to shutdown the server, but it fails because there are some locks still there | 03:49 |
grantc | it's possible in IMA i believe | 03:50 |
Deyan | ima? | 03:50 |
Deyan | ipm? | 03:50 |
grantc | IMA is Ingres Management Arch | 03:51 |
grantc | allows you query the internals using SQL | 03:51 |
grantc | http://community.ingres.com/wiki/IMA_Remove_Sessions_Script | 03:51 |
Deyan | cheers | 03:52 |
Deyan | it does not work | 04:00 |
Deyan | i get an error | 04:00 |
Deyan | ingres is actually executing that '\p\g' line | 04:02 |
Deyan | in the SELECT statement | 04:02 |
Vroomfondle | what're you running it through? isql, or sql? | 04:05 |
Vroomfondle | via isql you need to use ";commit;" instead of "\p\g" (at least, you do in Ingres 2.6 - not sure about later versions) | 04:06 |
grantc | the instructions say to use "sql" not "isql" | 04:06 |
Deyan | ok | 04:08 |
Deyan | it was too many apostrophes | 04:08 |
Deyan | in that SELECT statement | 04:08 |
grantc | can you paste your corrected version to the pastebin | 04:12 |
Deyan | looks to me like it actually does not do anything | 04:16 |
Deyan | i still have sessions open | 04:17 |
Deyan | grantc: one line there has ''''' | 04:18 |
Deyan | 5 single quotes | 04:18 |
Deyan | instead of 4 | 04:18 |
Deyan | that is the only mistake | 04:18 |
grantc | ok i c it | 04:18 |
Deyan | however, i still have those sessions... | 04:19 |
Deyan | so ima_session_remove() does not work... | 04:20 |
Deyan | yep | 04:22 |
Deyan | even if i do manually ima_remove_session it does not work | 04:22 |
*** dyki_ has joined #ingres | 04:22 | |
grantc | are you removing the session that is waiting or has the lock | 04:23 |
Deyan | yes | 04:23 |
Deyan | that was the reason why i need it | 04:23 |
Deyan | because server does not want to stop | 04:23 |
Deyan | for some strange reasons i got around 10 locks | 04:23 |
Deyan | that i am examining atm | 04:23 |
grantc | if you connect to the server the using iimonitor and paste the output from "show sessions formatted" there might be a reason why you cannot remove them | 04:24 |
Deyan | ty, i will try that | 04:24 |
* grantc goes to prepare lunch | 04:25 | |
* grantc is away: Away | 04:25 | |
magnusg| | Hi all, I'm trying to find the root of my build problems. The error message I have is apparently a syntax error i iiapi.h but that is hardly the true problem: | 04:36 |
magnusg| | Cc /home/ingres/ingres/server/src/common/cuf/ccm/ccmsrc.o | 04:37 |
magnusg| | In file included from /home/ingres/ingres/server/src/common/cuf/ccm/ccmsrc.c:9: | 04:39 |
magnusg| | iiapi.h:530 error:expected "=" , etc or attribute before '‘IIAPI_DT_ID’ | 04:41 |
magnusg| | Last line is inexact because for unkown reason I couldn't copy paste it here. | 04:44 |
Deyan | well, check iiapi.h | 05:00 |
Deyan | and see that line | 05:00 |
pboro | Deyan, http://pauli.borodulin.fi/blog/?p=50 | 05:01 |
pboro | there's still one issue: if you run buildrel -a after the build, it tells that there's two missing files... I'm yet trying to figure out why | 05:02 |
pboro | I have not yet tried GCC 4.x, but it should work per grantc | 05:03 |
Deyan | i have already commented | 05:03 |
Deyan | :D | 05:03 |
pboro | just install gcc4something instead of gcc33* and so on | 05:03 |
Deyan | gcc 4.3.0 is already here | 05:04 |
pboro | you may want to wait a bit until I figure out the missing files problem, it has got something to do with old xerces | 05:06 |
pboro | I just have to figure where it is still referenced | 05:06 |
magnusg| | Deyan, that line is part of a typedef which I don't think has been changed for years. I would expect the problem being something in the environment or with the compiler. | 05:06 |
magnusg| | That wasn't quite correct. | 05:10 |
pboro | Deyan, ah, found the final problem, have to modify couple ccpp-files | 05:18 |
pboro | way too many fixed values in the build systems :( fortunately it's better today... | 05:21 |
pboro | damn, why is ccpp not evaluating $XERCVERS | 05:26 |
*** atrofast has joined #ingres | 05:54 | |
pboro | ah, it must be exported elsewhere :) shinfolibs.sh... | 05:55 |
*** rossand has joined #ingres | 05:57 | |
*** ChanServ sets mode: +o rossand | 05:57 | |
*** Gerhard has quit IRC | 06:03 | |
*** Gerhard has joined #ingres | 06:04 | |
Deyan | grantc: i see you now use your code to do syntax highlighting | 06:09 |
Deyan | :) | 06:09 |
Deyan | (wiki) | 06:09 |
grantc | my code? | 06:12 |
Deyan | yeah, that syntax highlighting you talked about in your log | 06:33 |
Deyan | guys, can an AUTO_INCREMENT field be SMALLINT? | 06:46 |
Deyan | or it has to be an INTEGER ? | 06:46 |
magnusg| | Hi, the build succeeded when I did runbuild.sh -a and now I have Ingres 10.0.0 up and running. | 06:48 |
*** toumi01 has joined #ingres | 06:50 | |
Deyan | 10? | 06:52 |
Deyan | dude, you are my idol | 06:52 |
magnusg| | I hadn't expected 10 but I did a svn update and the build and createdb gave me "Ingres Release II 10.0.0 (int.lnx/00)NPTL Server -- Normal Startup." | 07:01 |
magnusg| | Sorry, should be createdbms of course (it is Monday after all) | 07:02 |
grantc | we had to fix a problem in ingres where by the version number could not go beyond 9.9 | 07:05 |
grantc | 9.4 was always going to be 10 :) | 07:05 |
magnusg| | Really, I thought there would be some more 9.x releases in before 10. | 07:08 |
*** mull has joined #ingres | 07:08 | |
grantc | there was supposed to be a 9.4 but that changed | 07:09 |
magnusg| | Because of need for system catalog changes or because it was time for a 10? | 07:10 |
grantc | i don't believe it was a technical decision to go to 10.0 | 07:15 |
grantc | the decision to do a 10.0 release required that we overcome some internal limitations | 07:16 |
magnusg| | OK | 07:17 |
magnusg| | Thanks, grantx | 07:17 |
grantc | np | 07:17 |
magnusg| | Me and my keyboard isn't on friendly terms today. | 07:17 |
*** magnusg| has quit IRC | 07:31 | |
*** Gerhard has quit IRC | 07:37 | |
*** rossand has quit IRC | 07:37 | |
*** dyki has quit IRC | 07:37 | |
*** toumi01 has quit IRC | 07:37 | |
*** Vroomfondle has quit IRC | 07:37 | |
*** bonro01 has quit IRC | 07:37 | |
*** clach04 has quit IRC | 07:37 | |
*** mull has quit IRC | 07:37 | |
*** atrofast has quit IRC | 07:37 | |
*** Nornagest has quit IRC | 07:37 | |
*** pboro has quit IRC | 07:37 | |
*** Deyan has quit IRC | 07:37 | |
*** Alex|off has quit IRC | 07:37 | |
*** rossand has joined #ingres | 07:48 | |
*** mull has joined #ingres | 07:48 | |
*** toumi01 has joined #ingres | 07:48 | |
*** Gerhard has joined #ingres | 07:48 | |
*** atrofast has joined #ingres | 07:48 | |
*** bonro01 has joined #ingres | 07:48 | |
*** clach04 has joined #ingres | 07:48 | |
*** Alex|off has joined #ingres | 07:48 | |
*** Nornagest has joined #ingres | 07:48 | |
*** Vroomfondle has joined #ingres | 07:48 | |
*** pboro has joined #ingres | 07:48 | |
*** irc.freenode.net sets mode: +oo rossand Alex|off | 07:48 | |
*** Deyan has joined #ingres | 07:54 | |
*** Gerhard has quit IRC | 08:33 | |
Deyan | grantc: are you there, buddy? | 09:04 |
Deyan | :) | 09:04 |
Deyan | grantc: how can i examine why ingres_close() does not return? | 09:04 |
Deyan | there must be some reason - i would like to find it | 09:04 |
grantc | debugger? | 09:05 |
Deyan | it just does not return | 09:07 |
Deyan | it waits for something | 09:07 |
grantc | building the driver in debug and attaching to the stalled process should show you where it's blocked | 09:10 |
grantc | CFLAGS=-g ./configure | 09:11 |
grantc | assuming you are on linux/unix | 09:11 |
grantc | or an API trace would give you and idea as to which function it is in | 09:11 |
*** Alex|off is now known as Alex| | 09:41 | |
Deyan | ingres locking issues are making my head boil | 09:45 |
*** dyki_ has quit IRC | 10:00 | |
Deyan | grantc: in order to do simple client connections all i need is to follow the DAMP specification, right? | 10:24 |
Deyan | i am thinking of writing a client library for D | 10:24 |
Deyan | using DAMP | 10:24 |
Deyan | but only if that is all i need | 10:24 |
Deyan | does anyone know? | 10:25 |
pboro | wha'ts DAMP? | 10:40 |
pboro | what's even... | 10:40 |
Deyan | Data Access Messaging Protocol | 10:40 |
pboro | ahh, yes | 10:45 |
*** DerMeister has joined #ingres | 10:49 | |
*** Deyan has quit IRC | 10:55 | |
clach04 | Deyan: I know you've left but just in case you check the IRC logs. Yes following the DAMP protocol would be what you need to do. However your client will need to handle things like Unicode conversion and timezone (date arithmetic) conversion. If you are lucky the standard libs with your language will support this otherwise you need to use 3rd party libs (or write your own which I would not recomend). | 10:59 |
clach04 | I'm assuming you found the spec in the wiki | 10:59 |
clach04 | one tip I would have would be to get (say) the JDBC driver working and do packet tracing on it. So you can compare a working system with the spec | 11:00 |
clach04 | just put a simple tcp proxy in the middle | 11:00 |
*** Args has joined #ingres | 13:09 | |
Args | is anyone around who know why I might be getting the following error form iilink: | 13:19 |
Args | /opt/Ingres/IngresII/ingres/lib/libscf.1.so: undefined reference to `IIudadt_register' | 13:19 |
*** Args is now known as siman01 | 13:21 | |
pboro | siman01, sounds like user defined type? or something... | 13:32 |
siman01 | yep I think I need to reread the manual - I tried following and example on the wiki - I think it might be missing a step | 13:33 |
*** Alex| is now known as Alex|off | 13:34 | |
siman01 | turns out your supposed to provide IIudadt_register yourself :) | 13:46 |
pboro | hehe :) | 13:48 |
*** DerMeister has quit IRC | 14:31 | |
*** mull has quit IRC | 14:47 | |
*** siman01 has quit IRC | 15:59 | |
*** mull has joined #ingres | 16:17 | |
*** mull has quit IRC | 20:28 | |
*** rossand has quit IRC | 20:29 | |
*** Alex|off is now known as Alex| | 22:36 | |
*** Alex| is now known as Alex|off | 23:01 | |
*** Alex|off is now known as Alex| | 23:18 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!