*** grantc has joined #ingres | 00:37 | |
*** ChanServ sets mode: +o grantc | 00:37 | |
*** toumi01 has quit IRC | 01:01 | |
*** toumi01 has joined #ingres | 01:07 | |
*** pboro has quit IRC | 01:12 | |
*** davi1 has joined #ingres | 01:12 | |
*** pboro has joined #ingres | 01:12 | |
*** dyki has joined #ingres | 01:15 | |
*** Deyan has joined #ingres | 01:40 | |
Deyan | hello guys | 01:41 |
---|---|---|
*** Alex|off is now known as Alex| | 02:11 | |
*** magnusg| has joined #ingres | 02:19 | |
magnusg| | Hi all, | 03:51 |
magnusg| | One of my customers has a bit of a problem and I wonder if someone has any brgiht idea. What they have done isn't quite according to the book but then | 03:52 |
magnusg| | They upgraded their development environment (ESQL C) to Ingres 9.2 with the production environment still is 9.1.x | 03:53 |
magnusg| | Then they needed to change the application and made a new version | 03:53 |
magnusg| | But there is a new symbol added in 9.2, IIcsRetScroll so the runtime environment screems and cries because it can't find this symbol. | 03:54 |
magnusg| | Anyone having some idea about how to get through this mess without either upgrading production or downgradin development which of course would be the supported way | 03:55 |
Deyan | can you link statically against Ingres libraries? | 03:57 |
Deyan | if not, deploy both libraries | 03:58 |
Deyan | with your application | 03:58 |
Deyan | both new (9.2) and old 9.1.x | 03:58 |
Deyan | actaully no | 03:59 |
Deyan | i will not fix the problem | 03:59 |
Deyan | try to link statically | 04:00 |
Deyan | that is the best solution | 04:00 |
Deyan | (IMHO) | 04:00 |
pboro | magnusg|, nice, that's a mess :( | 04:00 |
magnusg| | Deyan, thanks. My customers are developers where I am DBA etc so I think they can use your advice. | 04:01 |
magnusg| | Pboro, Yes. | 04:01 |
magnusg| | But then if everyone was doing things the most easy and smoothest way thing would be more boring :-) | 04:02 |
pboro | hehe :) | 04:02 |
*** grantc has quit IRC | 04:31 | |
*** grantc has joined #ingres | 04:32 | |
*** ChanServ sets mode: +o grantc | 04:32 | |
*** mull has joined #ingres | 05:54 | |
*** rossand1 has joined #ingres | 06:02 | |
*** magnusg| has quit IRC | 06:53 | |
*** DarylM has joined #ingres | 07:02 | |
Deyan | auto_increment in ingres is soooo buggy | 07:20 |
dyki | how one can become part of secret team of ingres betatesters? | 07:21 |
Deyan | you must embrace the dark force | 07:21 |
pboro | well... since you're using the community builds... you already are :D | 07:22 |
DarylM | it isn't released yet so bug reports are valuable | 07:22 |
dyki | well, I thought there is special hidden society of Ingres bug testers :P | 07:23 |
Deyan | DarylM: what amazes me that people did not notice these problems so far | 07:23 |
DarylM | Well, I have to agree with that assessment | 07:24 |
Deyan | when was the first 9.3 version released? | 07:24 |
Vroomfondle | just out of interest, do recent versions of Ingres do secure ODBC (i.e. SSL or similar)? | 07:24 |
Deyan | Vroomfondle: i haven't noticed that | 07:25 |
dyki | tunnel it through SSH :P Simplest.. | 07:26 |
Deyan | Vroomfondle: http://community.ingres.com/wiki/SSL_Integration_Project | 07:27 |
Deyan | there is the project | 07:27 |
Deyan | but i did not see any progress | 07:27 |
Vroomfondle | that'll be tricky as Ingres->Ingres connections use multiple and varied ports (or so I'm told by our systems manager) | 07:27 |
Vroomfondle | hmm, k. | 07:28 |
dyki | true.. | 07:28 |
Vroomfondle | I'm just wondering really. It'd be nice to have. Ingres' security is somewhat lacking. | 07:28 |
Vroomfondle | If someone can packet-sniff the right username & password to connect their Ingres instance to ours, they can masquerade as any user (including DBA accounts) just by creating a local OS user with the same name | 07:29 |
Vroomfondle | which seems somewhat dodgy to me | 07:29 |
Vroomfondle | though it's possible that our DBAs have just got it configured in some manky old way that isn't suitable any more (that does happen a lot) | 07:29 |
*** Alex| is now known as Alex|off | 07:32 | |
grantc | Vroomfondle, if your auth mech is not system then you a probably prone to packet sniffing | 07:32 |
Vroomfondle | what does "system" mean in this context? | 07:33 |
grantc | it's a name given for one of the authentication mechanisms the others being ingres and null | 07:33 |
grantc | the latter two are there for different compat reasons | 07:34 |
grantc | i am not sure how strong "system" is but it's not plain-text | 07:34 |
Vroomfondle | Hmm. Thanks. I'll check that tomorrow when I'm back in the office. | 07:35 |
Vroomfondle | though no doubt I'll get frowned at for harping on about security again. Ah well :) | 07:35 |
grantc | i believe you can do ingres/net via ssh | 07:35 |
DarylM | Security: Everyone wants it, no one wants to spend the money to get it. | 07:35 |
Deyan | that is why web technologies are so popular | 07:36 |
Deyan | - because it is the easiest way to have secure communication | 07:36 |
Deyan | (over https) | 07:36 |
grantc | like hosting your credit cards in the clouds :) | 07:37 |
dyki | pboro: quick one. does isAutoIncrement() function in ResultSetMetaData works for you? | 07:38 |
DarylM | EEK. Like web technologies are secure. http://news.bbc.co.uk/2/hi/business/8206305.stm | 07:38 |
dyki | in Java^^ | 07:38 |
dyki | latest JDBC ingres driver | 07:38 |
dyki | it has worked for me once when we had 9.3/140 or something like that. | 07:39 |
Deyan | DarylM: i did not say that | 07:40 |
Deyan | i said that is why so many technologies are based on https | 07:40 |
Deyan | - because it offers security in easiest way | 07:40 |
Deyan | sure those technologies can be insecure ;) | 07:40 |
DarylM | Understood. Simply pointing out that securing the pipe is useless if the end points are not secure. | 07:41 |
DarylM | And since Windows is frequently on one end of the pipe............ :D | 07:41 |
pboro | dyki, haven't tried it... we use sql sequences for generating ids | 07:41 |
Deyan | pboro: we are about to do the same... | 07:42 |
Deyan | :D | 07:42 |
Deyan | we learned the lesson | 07:42 |
pboro | There's no auto_increment in 9.2 so I could not even use it :) | 07:42 |
Deyan | trust me | 07:42 |
Deyan | you do not want to use it, yet | 07:42 |
dyki | well, we are reverting back from AUTO_INCREMENT to old sequences :) | 07:42 |
pboro | I will prolly use it once commercial 9.3 is released | 07:43 |
Deyan | yesterday, TM duplicated last 10 columns | 07:43 |
Deyan | for no reason | 07:43 |
pboro | dyki, our system before sequences was using a table for tracking the ids used :) | 07:43 |
Deyan | we ran SQL script through TM | 07:43 |
dyki | wow :) | 07:44 |
Deyan | and it simply duplicated ~10 columns | 07:44 |
dyki | we had sort of similar situation with few tables :) | 07:44 |
pboro | dyki, there simply was no other way :D | 07:44 |
pboro | but this dates way back to eighties or something... | 07:44 |
dyki | its good fun copying and merging ids.. and having those weird tables just for conversions | 07:44 |
*** rossand has joined #ingres | 08:29 | |
*** ChanServ sets mode: +o rossand | 08:29 | |
*** rossand has quit IRC | 08:29 | |
*** rossand1 has quit IRC | 08:47 | |
*** dyki has quit IRC | 09:52 | |
*** Deyan has quit IRC | 09:55 | |
*** rossand has joined #ingres | 09:57 | |
*** ChanServ sets mode: +o rossand | 09:57 | |
*** Dejan has joined #ingres | 10:36 | |
*** grantc has quit IRC | 11:38 | |
CIA-56 | Ingres: loera01 * r1864 /main/src/common/odbc/driver/options.c: | 12:01 |
CIA-56 | Ingres: The ODBC specs allow for a number of attributes to be implicitly set "as | 12:01 |
CIA-56 | Ingres: defined by the driver" when SQLSetStmtAttr() defines items such as the | 12:01 |
CIA-56 | Ingres: cursor type and scrollability. I had not specifically addressed this bit | 12:01 |
CIA-56 | Ingres: of the spec in previous submissions, so addressed this topic generally, and | 12:01 |
CIA-56 | Ingres: added a few changes to silence compiler warnings. | 12:01 |
pboro | ahh, this bot prints changes made to vcs? | 12:03 |
*** Alex|off is now known as Alex| | 12:16 | |
*** Alex| is now known as Alex|off | 13:12 | |
*** rossand has quit IRC | 13:48 | |
Dejan | pboro, yes | 15:25 |
Dejan | erm, if you think it is annoying i will order it to leave channel | 15:26 |
Dejan | personally, i think it is very helpful info | 15:26 |
*** mull has quit IRC | 15:33 | |
*** Dejan has quit IRC | 15:34 | |
*** DarylM has quit IRC | 15:38 | |
*** rossand has joined #ingres | 15:43 | |
*** ChanServ sets mode: +o rossand | 15:43 | |
*** davi1 has quit IRC | 16:08 | |
*** toumi01 has left #ingres | 17:42 | |
*** toumi01 has joined #ingres | 17:43 | |
*** rossand has quit IRC | 18:08 | |
*** mull has joined #ingres | 19:12 | |
*** rossand has joined #ingres | 19:45 | |
*** ChanServ sets mode: +o rossand | 19:45 | |
*** mull has quit IRC | 20:25 | |
*** rossand has quit IRC | 20:30 | |
*** Alex|off is now known as Alex| | 21:04 | |
*** Alex| is now known as Alex|off | 21:41 | |
*** Gerhard has joined #ingres | 23:26 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!