| *** sbishop411 has quit IRC | 00:00 | |
| *** sbishop411 has joined #ingres | 00:06 | |
| *** sbishop411 has quit IRC | 00:08 | |
| *** sbishop411__ has joined #ingres | 00:15 | |
| Gerhard | On an Ingres 9.2 / Win installation I ran "usermod dbname" and "usermod dbname -online" against a rather small database... | 00:21 |
|---|---|---|
| Gerhard | ... the first command took < 5 minutes, the second one took over 1 hour. Normal behaviour or bug? | 00:21 |
| *** Mud has joined #ingres | 00:30 | |
| *** sbishop411_ has quit IRC | 00:33 | |
| grantc | Gerhard, could it be that in the second case you had a lock on one or more tables? | 00:42 |
| Gerhard | grantc, definitely no locked tables, it's a rather small database just for testing purposes, a single user environment. No applications, no ISQLs running against the db. | 00:55 |
| grantc | can you run it again with printqry set to get timings? | 00:56 |
| grantc | perhaps it's just one table or all of them? | 00:57 |
| Gerhard | It is very slow for *each* table in the db. Each modify command appearing on the screen takes about 5 seconds... | 01:16 |
| Gerhard | ...even for tables that cotain 0 rows. | 01:16 |
| Gerhard | ...contain... | 01:16 |
| grantc | and if you run the modify command manually? | 01:17 |
| Gerhard | Without -online flag, it will modify several tables in *one* second, the database is quite small... | 01:18 |
| Gerhard | I may test this also with the demodb provided with Ingres, but I did not check this option during install time... | 01:19 |
| Gerhard | ...any chance to setup the demodb later? | 01:19 |
| Gerhard | grantc: manually = specifying one table explicitely? | 01:20 |
| grantc | i meant running the "modify" SQL command on the table, which is the same as doing it offline (i guess) | 01:21 |
| grantc | you can manually create the demodb using the copy.in script in ingres/demo/data | 01:23 |
| *** Alex| has joined #ingres | 01:25 | |
| *** ChanServ sets mode: +o Alex| | 01:25 | |
| *** paulm05 has joined #ingres | 01:28 | |
| paulm05 | morning all | 01:29 |
| grantc | morning | 01:31 |
| paulm05 | hi grant - did you get your rollforward working ok? | 01:31 |
| *** [1]Gerhard has joined #ingres | 01:31 | |
| grantc | yup - i now need to understand the difference between -rollback and -norollback | 01:31 |
| *** thoda05 has joined #ingres | 01:32 | |
| paulm05 | other than that one doesn't rollback uncommitted transactions? | 01:32 |
| *** bonro011 has joined #ingres | 01:32 | |
| *** Alex| has quit IRC | 01:33 | |
| *** Gerhard has quit IRC | 01:33 | |
| *** bonro01 has quit IRC | 01:33 | |
| *** bonsaikitten has quit IRC | 01:33 | |
| *** [1]Gerhard is now known as Gerhard | 01:33 | |
| grantc | also i think -norollback leave the database in a readonly state | 01:33 |
| paulm05 | that's true | 01:33 |
| paulm05 | I've got an outline of incremental rollforward that I did for my chapter in the Ingres book - want me to send it to you? | 01:34 |
| grantc | yes please | 01:34 |
| paulm05 | sent you an email - brief but may help | 01:37 |
| grantc | ta - so copying an open journal will be problematic? | 01:37 |
| *** grantc has quit IRC | 01:39 | |
| *** grantc has joined #ingres | 01:39 | |
| *** ChanServ sets mode: +o grantc | 01:39 | |
| paulm05 | not so much the copying but the rollforward | 01:39 |
| paulm05 | it will try to rollforward all the journals it sees | 01:40 |
| *** Alex| has joined #ingres | 01:41 | |
| *** ChanServ sets mode: +o Alex| | 01:41 | |
| paulm05 | hmm - actually I'm not so sure now I think about it | 01:42 |
| grantc | perhaps i can get rsync to ignore open files | 01:44 |
| grantc | or rather not copy open files | 01:45 |
| paulm05 | I've still never tried this - I'd be interested in your progress, any snags you hit etc | 01:46 |
| *** bonsaikitten has joined #ingres | 01:49 | |
| *** grantc has quit IRC | 02:16 | |
| *** grantc has joined #ingres | 02:45 | |
| *** ChanServ sets mode: +o grantc | 02:45 | |
| *** grantc has quit IRC | 02:55 | |
| *** grantc has joined #ingres | 02:55 | |
| *** ChanServ sets mode: +o grantc | 02:55 | |
| *** grantc has quit IRC | 03:04 | |
| *** grantc has joined #ingres | 03:15 | |
| *** ChanServ sets mode: +o grantc | 03:15 | |
| grantc | paulm05, from reading teh intrawebs rsync will copy a file even if it's being written to | 03:31 |
| paulm05 | so the question is whether that would matter | 03:33 |
| paulm05 | the rollforward -incremental must either know where it got up to last time or re-apply and not care previous journals | 03:33 |
| grantc | when i get the chance to reduce the archiver_interval/refresh on lxr i'll let you kknow | 03:33 |
| paulm05 | so it may not matter that you have an incomplete journal file | 03:34 |
| grantc | since we only append to a journal file would assume so too | 03:34 |
| grantc | if we don't just append... things could get interesting | 03:35 |
| paulm05 | my only worry would be if you got an incomplete journal record at the end of the file and rollforward didn't handle it gracefully | 03:35 |
| grantc | http://ingres.pastebin.com/d76ac7cc8 - it seems to be graceful so far | 03:37 |
| paulm05 | ok cool | 03:39 |
| grantc | although it appears to apply the same transactions until there's a new journal file | 03:41 |
| paulm05 | are you re-running it? | 03:43 |
| grantc | yup | 03:43 |
| grantc | the output from rollfwddb implies that | 03:44 |
| paulm05 | so what - you just rsync+rollforwarddb at regular intervals? | 03:44 |
| grantc | at the moment it's manual and will remain that way until i better understand what's going on | 03:45 |
| paulm05 | sure | 03:46 |
| paulm05 | I guess I was wondering if you need to check/wait for a new journal file before re-running rollforwarddb | 03:47 |
| paulm05 | looks like it only does one journal file at a time | 04:01 |
| grantc | i.e. when it's complete/closed? | 04:13 |
| paulm05 | no I mean that if you have 2 new journal files since the last incremental rfwd, you'd need to run it twice - I think! I'm getting this from reading the code | 04:14 |
| grantc | eh? | 04:18 |
| grantc | if that's what the code says then fine but I would have assumed it will process all files it's not seen if not the transactions within the files | 04:20 |
| paulm05 | it is possible I'm reading it wrong | 04:23 |
| paulm05 | I mentioned incremental rollforward to John Smedley as he was passing a moment ago and he was going to dig out an IUA presentation on it | 04:25 |
| grantc | thanks - searching my email there was a thread in june on this topic | 04:26 |
| grantc | @google ingres closed journals and incremental rollforwarddb | 04:27 |
| ii_log` | grantc: Problems with c.d.i. integration - Ingres Community Forums: <http://community.ingres.com/forum/newsgroups-mailing-lists/10852-problems-c-d-i-integration.html>; Incremental Rollforward Ingres 9.2 Michael Flower: <http://downloads.ingres.com/online/media/pdf/Ingres-DBMS/IncrementalRollforward.pdf>; Ingres Backup and Recovery: <http://downloads.ingres.com/online/media/pdf/Education- (2 more messages) | 04:27 |
| paulm05 | aha | 04:27 |
| grantc | @google ingres "closed journals and incremental rollforwarddb" | 04:27 |
| ii_log` | grantc: Problems with c.d.i. integration - Ingres Community Forums: <http://community.ingres.com/forum/newsgroups-mailing-lists/10852-problems-c-d-i-integration.html>; Incremental Rollforward Ingres 9.2 Michael Flower: <http://downloads.ingres.com/online/media/pdf/Ingres-DBMS/IncrementalRollforward.pdf>; Ingres Backup and Recovery: <http://downloads.ingres.com/online/media/pdf/Education- (2 more messages) | 04:27 |
| grantc | meh damn search | 04:28 |
| grantc | http://downloads.ingres.com/online/media/pdf/Ingres-DBMS/IncrementalRollforward.pdf | 04:28 |
| paulm05 | KB doc 417507 is also useful - though it stresses the need to only copy closed journal files without really saying why | 04:33 |
| *** rossand has joined #ingres | 04:35 | |
| *** ChanServ sets mode: +o rossand | 04:35 | |
| paulm05 | ah looks like there's now a way of checking if the journal file is closed so it should ignore open journal files unless you're on the final rollforward (i.e. the -rollback one) | 04:37 |
| paulm05 | (from KB 417424) | 04:38 |
| grantc | so until you have the disaster you don't issue the -rollforward | 04:43 |
| grantc | -rollback even | 04:43 |
| Vroomfondle | grantc: I have ingres.describe set to 1 in php.ini, but ingres_query errors whenever I try to omit the $types parameter. Do I need to do something else to make it autodetect parameter types (on 9.2) | 04:43 |
| grantc | Vroomfondle what error do you see? | 04:44 |
| Vroomfondle | Incorrect number of parameter types received, expected 10 got 0 | 04:45 |
| paulm05 | yes grant - that's my understanding | 04:45 |
| Vroomfondle | wait, sorry | 04:46 |
| * Vroomfondle pasted the wrong bit | 04:46 | |
| Vroomfondle | Unsupported type, 0 | 04:46 |
| Vroomfondle | that's the one | 04:46 |
| grantc | hmm - pls hold | 04:47 |
| Vroomfondle | I'm calling it with ingres_query($handle, $sql, $params); | 04:47 |
| grantc | paulm05, and i guess you need to restart the process for each new ckp? | 04:47 |
| Vroomfondle | I'm wondering if I should be using ingres_prepare and ingres_execute actually, but I'm not sure what advantage that brings | 04:47 |
| grantc | if you use cache_dynamic it will save the QEP generation | 04:49 |
| grantc | Vroomfondle, how simple is the query? | 04:50 |
| pboro | grantc, does cache_dynamic work nowadays? | 04:51 |
| Vroomfondle | simple. Just a SELECT with a few params in the WHERE. | 04:51 |
| grantc | pboro, i believe it's better behaved now | 04:51 |
| * Vroomfondle confirms that if he uses prepare/execute he still gets the error regarding types | 04:51 | |
| grantc | it's the same underlying code :) | 04:51 |
| pboro | grantc, ok... I'll enable it next night and if it breaks, I'll blame you ;) | 04:52 |
| Vroomfondle | ah, hang on | 04:52 |
| grantc | feel free | 04:52 |
| grantc | :D | 04:52 |
| Vroomfondle | oh buggerism | 04:52 |
| pboro | grantc, oh well, I will blame Ingres Corp. anyway ;) | 04:52 |
| grantc | ^^ | 04:52 |
| Vroomfondle | I forgot that although our main DB is on a 9.2 server, our menu DB (which gets accessed first) is on 2.6. My local instalation is 9.2. Does the query type detection rely on the local server or the remote one? | 04:53 |
| Vroomfondle | erm, param type | 04:53 |
| grantc | the describe input feature should only work when the server is capable | 04:54 |
| grantc | i.e. 9.1 or newer | 04:54 |
| Vroomfondle | yeah, but which server? | 04:54 |
| pboro | It's enabled now and will activate during next night's automated restart :) I guess I will have to keep eye on it tomorrow... | 04:54 |
| Vroomfondle | the local installation, or the remote installation which actually has the DB? | 04:54 |
| Vroomfondle | 'cos I'm pointing a 9.2 install at 2.6 | 04:54 |
| grantc | your 9.2 install is just a client in that case, | 04:55 |
| grantc | for the purposes of PHP | 04:55 |
| Vroomfondle | I see. | 04:55 |
| Vroomfondle | okay, so I need to wait until we've migrated our other servers | 04:55 |
| grantc | unless you have s separate thread/connection to a 9.2 db | 04:55 |
| Vroomfondle | I guess I'll just stop being impatient. | 04:55 |
| grantc | there are 3 conditions that must be satisfied | 04:56 |
| grantc | the apiLevel / ingres.describe and "canPrepare" | 04:56 |
| grantc | the one you should fail on is the first | 04:56 |
| pboro | ahh, there's again new patch for 9.2 on 64-bit Linux... grantc, can you quickly list me bug numbers fixed between 13599 and 13606?-) | 04:56 |
| grantc | 1..2..3..4... | 04:57 |
| pboro | ahaha :) I'm still dreaming of that service on the ESD that could be used to easily check which bugs has been fixed between patches X and Y :) | 04:57 |
| grantc | what are the patch numbers? | 04:58 |
| pboro | 13599 and 13606 | 04:58 |
| grantc | 2 bugs - 122755 122851 | 05:00 |
| pboro | thanks | 05:00 |
| * grantc does not no why the patch notes do not show this info | 05:00 | |
| grantc | s/no/know/ | 05:00 |
| pboro | 122755 "Segmentation fault on queries involving tables with default values in a high concurrency environment", I guess that's good enough reason to update, thanks | 05:01 |
| grantc | no problem | 05:01 |
| paulm05 | off to lunch - see you later | 05:10 |
| grantc | cya | 05:11 |
| *** zxiiro has quit IRC | 05:14 | |
| *** atrofast has joined #ingres | 05:20 | |
| * grantc is away: Away | 05:45 | |
| *** grantc is now known as grantc|afk | 05:46 | |
| *** mull has joined #ingres | 05:56 | |
| *** SaraDanaher has joined #ingres | 05:58 | |
| SaraDanaher | Hi Grant | 05:58 |
| SaraDanaher | are you here | 05:58 |
| SaraDanaher | ? | 05:58 |
| SaraDanaher | good morning Andrew | 05:59 |
| SaraDanaher | do you know what nick name is using Grant croket | 06:00 |
| SaraDanaher | croket | 06:00 |
| SaraDanaher | croker | 06:00 |
| SaraDanaher | sorry | 06:00 |
| SaraDanaher | he sent me some instructions to add "www-data" user to Ingres | 06:02 |
| SaraDanaher | Here are his instructions: | 06:03 |
| SaraDanaher | You cannot create a database owned by a user that has a "-". It should | 06:03 |
| SaraDanaher | be enough to add the "www-data" user to Ingres as per | 06:03 |
| SaraDanaher | http://community.ingres.com/wiki/Ingres_with_Apache_on_Debian_Etch. | 06:04 |
| SaraDanaher | Then you would create a drupal OS and Ingres user and then the drupal | 06:04 |
| SaraDanaher | databse | 06:04 |
| SaraDanaher | createdb -udrupal drupal | 06:04 |
| SaraDanaher | but u still have some problems | 06:04 |
| SaraDanaher | when i add the command createdb -udrupal drupal | 06:07 |
| SaraDanaher | i got erro: you must install at least one postgresql-client-version package | 06:09 |
| SaraDanaher | can some one help me | 06:09 |
| *** thoda05 has left #ingres | 06:14 | |
| *** zxiiro has joined #ingres | 06:14 | |
| *** ChanServ sets mode: +v zxiiro | 06:14 | |
| * grantc|afk is back (gone 00:38:21) | 06:24 | |
| *** grantc|afk is now known as grantc | 06:24 | |
| grantc | hi SaraDanaher | 06:24 |
| grantc | you were looking for me? | 06:24 |
| grantc | if you are getting that error/message then it means you don't have ingres's createdb command in your path | 06:25 |
| SaraDanaher | i believe i did it | 06:26 |
| grantc | do you have the environment variable II_SYSTEM defined? | 06:26 |
| SaraDanaher | is that process during the building ingres | 06:28 |
| SaraDanaher | ? | 06:28 |
| *** cthibert has joined #ingres | 06:28 | |
| *** toumi01 has joined #ingres | 06:38 | |
| grantc | I believe II_SYSTEM is used as part of the build process but it's likely to be different for a final install | 06:40 |
| grantc | i have a script that I use for setting up my ingres environment | 06:42 |
| grantc | http://ingres.pastebin.com/f6aebf996 - change the value set for II_SYSTEM to the value you have for your ingres installation | 06:43 |
| *** rossand has quit IRC | 06:43 | |
| SaraDanaher | Thanks Grant | 06:45 |
| grantc | once you save that file you will need to source it using ". file" or "source file" | 06:46 |
| Gerhard | grantc: this morning, we talked about usermod's -online flag... | 07:06 |
| Gerhard | I fiddled out the difference between the modify commands | 07:06 |
| grantc | hi Gerhard, sure | 07:06 |
| Gerhard | online.sql: http://ingres.pastebin.com/m579e13c3 | 07:07 |
| Gerhard | offline.sql: http://ingres.pastebin.com/d3c8f23db | 07:08 |
| grantc | the concurrent_updates flag appears to be the only difference | 07:08 |
| *** Alex| has quit IRC | 07:14 | |
| Gerhard | and it makes modify significantly slower. | 07:16 |
| *** Gerhard has quit IRC | 07:16 | |
| *** ccsidiot has joined #ingres | 07:20 | |
| *** SaraDanaher has quit IRC | 07:22 | |
| *** hanje04 has joined #ingres | 07:28 | |
| *** hanje04 has left #ingres | 07:28 | |
| *** hanje04 has joined #ingres | 07:28 | |
| *** zxiiro has quit IRC | 07:50 | |
| ccsidiot | Hey | 07:51 |
| ccsidiot | I double checked with my sysadmin and he said that the server that I'm on is on 32 bit, but cat /proc/cpuinfo is giving me 64, why is that? | 07:52 |
| *** zxiiro has joined #ingres | 07:52 | |
| *** ChanServ sets mode: +v zxiiro | 07:52 | |
| hanje04 | what does uname -a return? | 07:53 |
| hanje04 | It's possible you're running a 32bit OS on a 64bit CPU | 07:54 |
| ccsidiot | Linux dbsrv3 2.6.26-2-686 #1 SMP Wed Nov 4 20:45:37 UTC 2009 i686 GNU/Linux | 07:54 |
| hanje04 | Yup, that's 32bit x86 | 07:54 |
| hanje04 | Nearly all modern x86 processor are x86_64 but they can run the x86 instruction set natively | 07:55 |
| hanje04 | So you can install a 32bit OR 64bit OS | 07:55 |
| ccsidiot | Oh I see, cool, which means runbuild.sh failed to pick up GEOS when the path is correctly set has nothing to do with the hardware part | 07:55 |
| hanje04 | This also means you can run the 32bit Linux version of Ingres on the 64bit x86_64 Linux unmodified | 07:56 |
| hanje04 | Correct | 07:56 |
| ccsidiot | I have this line in the end of the runbuild.sh | 07:57 |
| ccsidiot | echo $LD_LIBRARY_PATH | 07:57 |
| ccsidiot | and it prints | 07:57 |
| ccsidiot | where my GEO libgeso_c.so stuffs are | 07:57 |
| ccsidiot | but it's trying to look into /usr/local/bin | 07:57 |
| ccsidiot | while it's running runbuild.sh | 07:57 |
| hanje04 | What's the exact error you're getting? | 07:58 |
| ccsidiot | http://ingres.pastebin.com/d584d3ad8 | 07:58 |
| hanje04 | LD_LIBRARY_PATH is a runtime variable, it doesn't influence the linker | 08:00 |
| ccsidiot | The path is different from linker? | 08:00 |
| hanje04 | You need to add -L /path/to/osgeo/libs to the link line | 08:00 |
| hanje04 | LD_LIBRARY_PATH tell executables where to find libraries when then are running | 08:01 |
| hanje04 | What you need to do is tell the linker where to find the osgeo library when it's linking | 08:01 |
| hanje04 | you do this with the '-L' flag | 08:01 |
| ccsidiot | So like I run the runbuild.sh with the -L, is that what you mean? | 08:02 |
| hanje04 | No... | 08:02 |
| hanje04 | Did you change anything to add -lgeos_c to the link flags? | 08:03 |
| hanje04 | it's not included in a regular build | 08:03 |
| atrofast | hanje04: This is the geospatial branch and we use GEOS... I was under the impression that LD_LIBRARY_PATH was picked up by ld upon linking, is this not the case? | 08:06 |
| hanje04 | No, it's a runtime variable | 08:06 |
| hanje04 | It's picked up by the link loader, not the linker | 08:07 |
| hanje04 | You need to use -L /path to tell the linker where to look | 08:07 |
| ccsidiot | I found this line echo "LDLIBMACH32 = $prefix32 $libs $suffix" in mkdefault.sh | 08:10 |
| ccsidiot | Not sure if that's the part that I'm modifying to | 08:11 |
| atrofast | Thanks hanje04, ccsidiot I'm sorry about giving that tip, I was not aware taht LD_LIBRARY_PATH worked in that way | 08:11 |
| ccsidiot | atrofast: No problem :) | 08:11 |
| hanje04 | ccsidiot: Well found! Is that in tools/port/jam/Jamdefs.x86hybrid | 08:12 |
| hanje04 | (or whatever it's called) | 08:12 |
| hanje04 | oh, no. you said mkdefault.sh | 08:12 |
| hanje04 | you shouldn't need to change mkdefault | 08:12 |
| ccsidiot | Oh, hanje04: thanks for letting me know :) | 08:13 |
| hanje04 | edit the Jamdefs file I mentioned above | 08:13 |
| hanje04 | atrofast: Strictly speaking we probably need to add a variable to tell the build environment where the osgeo libraries are | 08:14 |
| hanje04 | like we do for Xerces | 08:14 |
| hanje04 | That way we can handle them being in a non standard location | 08:15 |
| *** Mud has quit IRC | 08:21 | |
| cthibert | hanje04 - Can we make this pull from an env var in the Jam file? | 08:22 |
| cthibert | Something like -L $(GEOS_LOC) | 08:22 |
| hanje04 | cthibert: Yeah, that's what I had in mind | 08:23 |
| ccsidiot | Great! Progress, thanks so much hanje04 + atrofast + FrankW + anyone helped me on this | 08:23 |
| hanje04 | In the Jamdefs file, just define it with a default value (var ?= val) so that it can be overwritten | 08:24 |
| hanje04 | We do that in a few other places | 08:24 |
| ccsidiot | Humm... now maybe I changed the jamdefs.hyb86_lnx incorrectly | 08:29 |
| ccsidiot | syntax error | 08:29 |
| hanje04 | ccsidiot: show me svn diff | 08:31 |
| ccsidiot | hanje04: here you go http://ingres.pastebin.com/d21a58eee | 08:33 |
| hanje04 | Space is significant in Jamfiles | 08:34 |
| hanje04 | you need a space before the final ; | 08:34 |
| hanje04 | and you don't need the trailing / either although that shouldn't have any ill effect | 08:34 |
| ccsidiot | But when I check the other lines, they also don't have a space after the final ; | 08:36 |
| hanje04 | _before_ no after | 08:36 |
| hanje04 | *not | 08:36 |
| ccsidiot | Oops, sorry, I read that wrong | 08:36 |
| ccsidiot | :) | 08:36 |
| cthibert | Hmm, putting the -rpath in there might be helpful too. | 08:39 |
| cthibert | I see there's a LDORIGIN variable that has an -rpath arg. Would adding $GEOS_LOC there be the way to approach this, or in the LDLIBMACH32? | 08:41 |
| ccsidiot | Hum, how come the same error is coming back again :( | 08:41 |
| hanje04 | ccsidiot: Sorry, I miss it the first time I looked. The -L /dir has to come _before_ the library you are including | 08:43 |
| hanje04 | cthibert: -rpath is a bit of a contentious issue | 08:44 |
| cthibert | ? | 08:44 |
| hanje04 | cthibert: I'm trying to make the Ingres RPM packaging LSB compliant in an attempt to get the RPMs into the Fedora distribution. RPATH linked binaries are one of there "no nos" | 08:45 |
| cthibert | Ah, okay. | 08:45 |
| cthibert | So, just have to remember to set LD_LIBRARY_PATH before running Ingres. Not that big a deal. | 08:46 |
| hanje04 | cthibert: I had a discussion with Karl about this and I agree with him that using the $ORIGIN still is probably a safe soln. but until I've had a chance to discuss it with Fedora I'd rather not add any more RPATH linkages | 08:47 |
| cthibert | Sounds good. | 08:47 |
| hanje04 | also, $ORIGIN is only really useful if you can guarantee the library is _always_ going to be in the same relative position on every machine you run it on. I'm guessing this isn't the case for osgeo | 08:51 |
| cthibert | Not without root access to install. | 08:51 |
| ccsidiot | Hum, I tried LDLIBMACH32 = -lpthread -lrt -lm -lc -lcrypt -ldl -lgcc_s -L /data/g7iwwong/geos-3.2.0rc1 -lgeos_c ; | 08:57 |
| ccsidiot | same error | 08:57 |
| ccsidiot | I moved the LDLIBMACH32 before LDLIBPATH32 and same error | 08:57 |
| ccsidiot | I tried moving -L before other flags 1 at a time, same error | 08:58 |
| ccsidiot | :( | 08:58 |
| hanje04 | Is the library actually in /data/g7iwwong/geos-3.2.0rc1? or is it in a 'lib' or some other sub directory? | 09:01 |
| cthibert | ccsidiot - if you didn't do make install in geos (which I imagine you didnt't) then the libraries are in two locations actually. In your geos-3.2.0rc1/source/libs and geos-3.2.0rc1/capi/libs | 09:05 |
| ccsidiot | Hum, trying to put the lib after that path | 09:06 |
| cthibert | Sorry, make those libs == .libs So source/.libs | 09:06 |
| ccsidiot | It was there, but I decided to do a clean build this morning and forgot to make install | 09:07 |
| ccsidiot | ^^" my bad | 09:07 |
| cthibert | hanje04 - There's another link missing later on. iimerge fails | 09:09 |
| *** Alex| has joined #ingres | 09:09 | |
| *** ChanServ sets mode: +o Alex| | 09:09 | |
| cthibert | There likely is another place to add the -L to. | 09:10 |
| ccsidiot | Yay, progress! But it's failing in some other case, another -L as suggested by cthibert? | 09:11 |
| hanje04 | Crap. Sorry, that will require changing mkdefault.sh because it needs to go into iisysdep | 09:11 |
| hanje04 | I'd forgotten about the server, wasn't thinking | 09:12 |
| cthibert | No worries. We're getting there. :) | 09:12 |
| ccsidiot | Now the build failed on this "failed Cc /data/g7iwwong/geospatial/src/common/adf/adu/adupoint.o" | 09:13 |
| hanje04 | changing mkdefault.sh will require a full rebuild | 09:13 |
| *** Mud has joined #ingres | 09:14 | |
| hanje04 | ccsidiot: That won't be a linking problem, what's the error? | 09:14 |
| cthibert | ccsidiot - in parallel I'm changing these files to use environment variables. So hopefully we'll have a flexible way for everyone to make use of this fix. | 09:14 |
| ccsidiot | cthibert: sounds sweet :) hanje04: The error is http://ingres.pastebin.com/d6ec5129e | 09:15 |
| hanje04 | cthibert: Good man. The hybrid stuff may be a little more fiddly, I know Karl changed some stuff around | 09:15 |
| FrankW | ccsidiot: btw my claim last night that a clean can be accomplished with runbuild.sh clean was wrong - it is "runbuild.sh -c" | 09:15 |
| hanje04 | cssidiot: Line 742, the compiler can't find the geos_c.h header. Looks like -I /geos/include/dir needs adding to the compile line | 09:17 |
| cthibert | Right, should have thought of that too. | 09:17 |
| hanje04 | how long has ingres.pastbin.com been in operation. That's _really_ useful, not seen it before today | 09:19 |
| pboro | for more than a year :) | 09:20 |
| *** atrofast has quit IRC | 09:22 | |
| ccsidiot | FrankW: No problem :) We are getting there. hanje04: Oh right! Hum, let me try | 09:24 |
| hanje04 | pboro: lol :-/ | 09:26 |
| hanje04 | it amazing what you learn when you finally find an IRC client you like ;-) | 09:27 |
| *** atrofast has joined #ingres | 09:28 | |
| *** paulm05 has quit IRC | 09:34 | |
| *** Alex| has quit IRC | 09:41 | |
| ccsidiot | Hum, is there anything I need to watch out for when I put the -I? I tried the path where the geos_c.h is located before + after, both gave me error | 09:41 |
| ccsidiot | saying geos_c.h not found | 09:41 |
| cthibert | The -I will have to be added before the linker. It's added in the cc line usually. I'm trying to find the appropriate place for adding header file locations. | 09:44 |
| cthibert | hanje04 - Is Jamdefs the file for adding header locations? It has a lot of HDR* type settings. I've tried adding a HDRGEOS= $(GEOS_INC) ; type line but that's no go. | 09:46 |
| *** atrofast has quit IRC | 09:52 | |
| ccsidiot | Yay, I tried to put the -I /path/to/geos_c.h in CCMACH32, and so far it's still building, on its way to 1000th target | 10:15 |
| ccsidiot | lol | 10:15 |
| ccsidiot | :) Looks good so far | 10:16 |
| *** grantc has quit IRC | 10:21 | |
| ccsidiot | Hum, it failed saying the following about IILINK | 10:25 |
| cthibert | I've got a diff that will compile and run. | 10:27 |
| cthibert | It uses $GEOS_LOC for libs, $GEOS_INC for headers and LD_LIBRARY_PATH should include the geos library path at compile and runtime. | 10:27 |
| ccsidiot | Oh, great :) | 10:28 |
| cthibert | I'll email you the diff file and you can apply it as a patch, thereby getting the 'apply a patch' requirement Andrew asked for. | 10:28 |
| cthibert | ccsidiot - You are Eva right? Just want to make sure I send the email to the right person. :) | 10:32 |
| ccsidiot | Yes, I'm | 10:32 |
| ccsidiot | :) | 10:32 |
| ccsidiot | Thanks cthibert | 10:32 |
| *** atrofast has joined #ingres | 10:34 | |
| cthibert | ccsidiot - On it's way. I'm going to grab some lunch though. If you have questions about the patch, I'll be back in a bit. | 10:41 |
| ccsidiot | I got the patch :) It's so sad every day I need to go before 2pm for classes :( | 10:43 |
| ccsidiot | I'll follow the instructions that comes with the email and update you guys as I go | 10:44 |
| ccsidiot | Thanks cthibert + hanje04 :) | 10:45 |
| *** ccsidiot has left #ingres | 10:50 | |
| hanje04 | cthibert: Jamdefs is the place for HDRGEOS | 11:01 |
| hanje04 | but you will also need to add it too the IISUBSYS for the directories where it's needed | 11:03 |
| hanje04 | The include path is context sensitive | 11:04 |
| *** zxiiro has quit IRC | 11:23 | |
| *** zxiiro has joined #ingres | 11:32 | |
| *** ChanServ sets mode: +v zxiiro | 11:32 | |
| cthibert | Thanks hanje04, I stumbled upon it with some liberal use of grep. Good to confirm that it's the right place and not just a place that works. ;) | 11:35 |
| *** Mud has quit IRC | 11:48 | |
| atrofast | Fedora 12 now uses Empathy instead of Pidgin, it has crashed twice on me :P | 12:01 |
| bonsaikitten | atrofast: quality product as usual ;) | 12:02 |
| * bonsaikitten is still not going to talk with yum. unmerging glibc is rude. | 12:02 | |
| atrofast | Heh! | 12:02 |
| atrofast | I still like Fedora | 12:02 |
| atrofast | Not sure if Empathy was mature enough to move to just yet | 12:02 |
| bonsaikitten | it's a funny game | 12:02 |
| bonsaikitten | but usually I don't want to play | 12:02 |
| atrofast | What's the biggest issue you have with yum? | 12:03 |
| bonsaikitten | it doesn't work | 12:04 |
| bonsaikitten | maybe I'm just expecting too much | 12:05 |
| atrofast | BRB | 12:06 |
| *** atrofast has quit IRC | 12:06 | |
| *** atrofast has joined #ingres | 12:15 | |
| *** PaulM05 has joined #ingres | 12:20 | |
| *** Alex| has joined #ingres | 12:24 | |
| *** ChanServ sets mode: +o Alex| | 12:24 | |
| *** atrofast has quit IRC | 12:50 | |
| *** cthibert has left #ingres | 13:00 | |
| *** atrofast has joined #ingres | 13:11 | |
| *** rossand has joined #ingres | 13:13 | |
| *** ChanServ sets mode: +o rossand | 13:13 | |
| *** Alex| has quit IRC | 13:22 | |
| *** PaulM05 has left #ingres | 14:17 | |
| *** toumi01 has left #ingres | 14:33 | |
| *** zxiiro has quit IRC | 14:34 | |
| *** mull has quit IRC | 14:36 | |
| *** PaulM05 has joined #ingres | 14:57 | |
| *** PaulM05 has left #ingres | 14:58 | |
| *** rossand has quit IRC | 15:26 | |
| *** atrofast has quit IRC | 15:29 | |
| *** atrofast has joined #ingres | 15:29 | |
| *** rossand has joined #ingres | 15:41 | |
| *** ChanServ sets mode: +o rossand | 15:41 | |
| *** clach04 has quit IRC | 17:00 | |
| *** rossand has quit IRC | 17:12 | |
| *** hanje04 has quit IRC | 18:47 | |
| *** evanBowling has quit IRC | 18:48 | |
| *** hanje04 has joined #ingres | 19:19 | |
| *** cytrinox_ has joined #ingres | 19:20 | |
| *** cytrinox has quit IRC | 19:36 | |
| *** cytrinox_ is now known as cytrinox | 19:36 | |
| *** sbishop411__ has left #ingres | 20:13 | |
| *** rossand has joined #ingres | 20:24 | |
| *** ChanServ sets mode: +o rossand | 20:24 | |
| *** rossand has quit IRC | 20:47 | |
| *** hanje04 has quit IRC | 22:01 | |
| *** FrankW has quit IRC | 22:16 | |
| pboro | doh... cache_dynamic is still broken | 23:06 |
| *** grantc has joined #ingres | 23:26 | |
| *** ChanServ sets mode: +o grantc | 23:26 | |
| *** thoda04 has joined #ingres | 23:53 | |
| *** bonsaikitten has quit IRC | 23:55 | |
| *** Vroomfondle has quit IRC | 23:55 | |
| *** Vroomfondle has joined #ingres | 23:57 | |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!