| 00:41.44 | *** join/#storm niemeyer (~niemeyer@201-25-35-133.pltce701.dsl.brasiltelecom.net.br) |
| 01:10.55 | *** join/#storm fcorrea (~fcorrea@201-62-110-55-ma.static.vivax.com.br) |
| 01:43.43 | *** join/#storm skermajo (~j@ppp118-208-17-34.lns20.bne1.internode.on.net) |
| 02:06.21 | *** join/#storm jdobrien (~john@97.103.22.146) |
| 02:18.11 | *** join/#storm jdobrien (~john@97.103.22.146) |
| 02:30.47 | *** join/#storm stub (~stub@canonical/launchpad/stub) |
| 07:48.56 | *** join/#storm zmijunkie (~chatzilla@i59F5DF04.versanet.de) |
| 07:57.13 | *** join/#storm bigjools (~quassel@82-71-93-254.dsl.in-addr.zen.co.uk) |
| 07:57.14 | *** join/#storm bigjools (~quassel@canonical/launchpad/bigjools) |
| 08:30.42 | *** join/#storm keppla (~keppla@i577B119E.versanet.de) |
| 08:38.23 | *** join/#storm jdobrien (~john@97.103.22.146) |
| 09:18.45 | *** join/#storm keppla (~keppla@i577B2D69.versanet.de) |
| 09:22.20 | *** join/#storm stub (~stub@ppp-58-8-11-147.revip2.asianet.co.th) |
| 09:22.20 | *** join/#storm stub (~stub@canonical/launchpad/stub) |
| 09:26.21 | jkakar | Hrm. I want the where clause feature philiKON was trying to add to ReferenceSet. |
| 09:27.29 | therve | what does it do? |
| 09:30.22 | jkakar | therve: I want something like carnivores = ReferenceSet(id, User.account_id, User.eats_meat == True) and vegetarians = ReferenceSet(id, User.account_id, User.eats_meat == False). |
| 09:31.33 | jkakar | jml: If you didn't already see it, I posted a thing about the collection pattern: http://jamwow.wordpress.com/2010/06/13/object-collections-in-storm/ |
| 09:32.47 | jkakar | It didn't occur to me until writing the post that ResultSet.find is basically the most rudimentary form of the pattern, which was an interesting realization. |
| 09:33.08 | therve | jkakar: properties! there are great |
| 09:33.32 | therve | they even |
| 09:34.25 | jkakar | therve: Well, sort of. |
| 09:34.50 | jkakar | therve: The problem I have is that I want to expose a ReferenceSet of users, but I want to exclude one type... hence wanting to include a where clause there. |
| 09:35.12 | jkakar | Anyway, yeah, I'll just add some methods get_carnivores and get_vegetarians, but it's slightly annoying. |
| 10:09.13 | jml | jkakar, thanks, I missed that in the deluge of stuff in my reader. |
| 10:09.40 | jml | jkakar, it kind of did occur to me that the collection methods were just specializations of resultset.find |
| 10:10.12 | jml | jkakar, and that if resultset.find().find() ever worked, it'd be a lot more like a collection :) |
| 10:18.05 | jkakar | jml: :) |
| 11:08.05 | *** join/#storm fcorrea (~fcorrea@187.3.31.9) |
| 11:17.19 | *** join/#storm marienz (~marienz@freenode/staff/marienz) |
| 11:22.38 | marienz | would it make sense to compile.add_reserved_words() the list on http://www.sqlite.org/lang_keywords.html in the sqlite backend, much like the mysql backend does? There are 38 words in that list that aren't already in the general sql reserved words list in the expr module. |
| 11:23.42 | marienz | (and I just foolishly tried to name a column "index" and got an OperationalError :) |
| 11:35.21 | jkakar | marienz: Yep, it would. |
| 11:35.30 | jkakar | marienz: Can you file a bug about it please? |
| 11:35.36 | marienz | sure |
| 11:35.44 | jkakar | marienz: Thanks! |
| 11:35.59 | marienz | I'll include a patch for sqlite itself, but I don't know how you prefer to test this |
| 11:43.41 | jkakar | marienz: An example of how to test this is test_mysql_specific_reserved_words in tests/databases/mysql.py. |
| 11:44.00 | marienz | yeah, I saw that but wasn't sure if doing it that way was optimal |
| 11:44.28 | marienz | I was tempted to throw in a test for the specific failure I actually encountered |
| 11:45.40 | jkakar | marienz: If you implement the test first, you should see a failure for 'index', right? |
| 11:45.55 | marienz | I guess there's that |
| 11:46.07 | marienz | give me a few minutes and I'll push a proper branch |
| 11:46.29 | jkakar | marienz: Thanks! |
| 11:54.12 | marienz | jkakar: how do you prefer these, just link the branch to the bug or submit a merge proposal or something else? I don't use launchpad all that often. |
| 11:54.35 | marienz | (pushed lp:~marienz/storm/sqlite-reserved-words) |
| 11:54.40 | jkakar | marienz: Please link the branch to the bug and submit a merge proposal. |
| 11:55.08 | jkakar | marienz: That puts it in the review queue, which will make it visible to Storm hackers: https://code.edge.launchpad.net/storm/+activereviews |
| 11:56.27 | marienz | thanks, both done |
| 11:59.59 | *** join/#storm jamesh_ (~james@canonical/launchpad/jamesh) |
| 12:02.36 | jkakar | marienz: Review delivered! :) |
| 12:04.15 | marienz | different word list was intentional but overly paranoid. Forgetting the docstring was just stupidity/lazyness. Fixing... |
| 12:04.52 | jkakar | marienz: Thanks! |
| 12:07.51 | marienz | jkakar: pushed, do I need to hit a button in the review request webinterface or is that automatic? |
| 12:08.10 | jkakar | marienz: No button pushing necessary. |
| 12:08.40 | jkakar | marienz: We have a two-reviews-per-branch policy. I'll find someone else to take a peak and land it for you ASAP. |
| 12:09.35 | marienz | no rush :) |
| 12:18.24 | *** join/#storm jamesh_ (~james@canonical/launchpad/jamesh) |
| 13:06.09 | m4v | hi, I asked once but I'll try again. I'm trying out storm and I'm confused over something when I use SQL on a sqlite db... |
| 13:06.12 | m4v | http://pastebin.ca/1882846 |
| 13:07.40 | m4v | why does the first SELECT return None? with the str parameter instead of unicode |
| 13:13.23 | *** join/#storm bac (~bac@cpe-065-190-186-163.nc.res.rr.com) |
| 13:13.23 | *** join/#storm bac (~bac@canonical/launchpad/bac) |
| 13:14.56 | *** join/#storm bac` (~bac@cpe-065-190-186-163.nc.res.rr.com) |
| 13:21.18 | marienz | weird |
| 13:21.48 | marienz | m4v: I believe this is related: at the pysqlite level conn.execute('SELECT sql FROM sqlite_master WHERE name=?', ('test',)).fetchall() returns a row, but conn.execute('SELECT sql FROM sqlite_master WHERE name=?', (buffer('test'),)).fetchall() does not |
| 13:22.19 | marienz | and I believe storm's sqlite database backend is converting str to buffer |
| 13:22.42 | marienz | I do not know why, though :( |
| 13:26.12 | m4v | I see now, is pysqlite that doesn't like buffer objects |
| 13:26.49 | marienz | I believe it's more likely matching a buffer against things that aren't blobs doesn't work |
| 13:28.57 | marienz | m4v: see also http://docs.python.org/dev/library/sqlite3.html#introduction |
| 13:30.10 | marienz | I believe the problem is that the column name you're trying to match against is text, while the parameter you're passing in is bytes. And the differing types compare non-equal, so you get no rows. |
| 13:31.20 | marienz | m4v: if I wanted to use the sqlite_master table I'd probably write a storm class matching its schema and query it through that |
| 13:34.33 | m4v | is ok, I only want to check if my tables exist when the program starts, I'll be using unicode instead of str. |
| 13:35.33 | *** join/#storm bac (~bac@cpe-065-190-186-163.nc.res.rr.com) |
| 13:35.33 | *** join/#storm bac (~bac@canonical/launchpad/bac) |
| 13:36.32 | marienz | m4v: something like http://paste.pocoo.org/show/225273/ |
| 13:40.54 | m4v | marienz: ok, I'll use it, thanks for your trouble :) |
| 15:20.28 | *** join/#storm elmom_ (~elmom@vallila-gw.hupnet.helsinki.fi) |
| 16:04.10 | *** join/#storm bac` (~bac@cpe-065-190-186-163.nc.res.rr.com) |
| 17:01.29 | *** join/#storm bac (~bac@canonical/launchpad/bac) |
| 17:32.42 | *** join/#storm deam (deam@is.really.feared.eu) |
| 17:54.46 | *** join/#storm jamesh_ (~james@canonical/launchpad/jamesh) |
| 19:46.12 | *** join/#storm skermajo (~j@ppp118-208-128-21.lns20.bne1.internode.on.net) |
| 20:12.56 | *** join/#storm skermajo (~j@ppp118-208-130-50.lns20.bne1.internode.on.net) |
| 20:30.34 | *** join/#storm keppla (~keppla@i577B2D69.versanet.de) |
| 20:34.45 | *** join/#storm skermajo (~j@ppp118-208-14-96.lns20.bne1.internode.on.net) |
| 21:46.30 | *** join/#storm jdobrien (~john@146.22.103.97.cfl.res.rr.com) |
| 21:53.51 | *** join/#storm niemeyer (~niemeyer@201-25-35-133.pltce701.dsl.brasiltelecom.net.br) |
| 23:24.16 | keturn_ | My test suite discovers that storm has left some database connections open until garbage collection. |
| 23:24.45 | keturn_ | I could run gc.collect in teardown, but it occurs to me that this is an opportunity to clean up some sloppy connection handling. |
| 23:25.42 | keturn_ | is there a way to find all open store connections? |
| 23:28.01 | marienz | to test this kind of thing I think I would take weakrefs to the relevant objects, then test that those weakrefs are dead when they should be |
| 23:28.07 | marienz | or is that not what you're asking about? |
| 23:30.14 | keturn_ | I think the weakrefs would be dead, as everything is fine as long as I run gc.collect(). I am thinking maybe I should be running store.close explicitly in more places, so there are not open transactions hanging out until GC happens. |
| 23:33.19 | *** join/#storm fcorrea (~fcorrea@201-62-110-55-ma.static.vivax.com.br) |
| 23:43.29 | marienz | if gc.collect() is actually doing something the weakref would be alive before it and dead after it. |