| 00:07.04 | radix | tjs: what should storm do in that case? |
| 00:12.01 | tjs | radix: use the currect syntax for creating an empty row for whatever backend |
| 00:12.05 | tjs | correct |
| 00:12.22 | tjs | radix: and happy birthday |
| 00:12.28 | tjs | for the other day |
| 00:13.11 | tjs | radix: It obviously needs to be special-cased |
| 00:13.24 | *** join/#storm dinho_ (n=dinho@201.79.73.173) |
| 00:13.40 | tjs | isnt it something like INSERT INTO table; |
| 00:17.10 | radix | tjs: thanks |
| 00:17.26 | radix | tjs: hmm, ok |
| 00:18.10 | radix | I can't figure out any syntax for that in postgres |
| 00:18.14 | radix | it's not "INSERT INTO table;" |
| 00:26.09 | dinho_ | INSET INTO tablename DEFAULT VALUES; |
| 00:26.26 | tjs | dinho_: thanks |
| 01:05.36 | tjs | well, my crossrefs code seems to work well :) |
| 01:05.56 | tjs | not that I've used it in anger, but the tests are all passing :) |
| 01:06.35 | tjs | it does zodb -> zodb, storm -> storm, storm -> zodb and zodb -> storm relationships |
| 01:07.01 | tjs | atm 1 to 1, but about to add 1 to many |
| 01:10.48 | *** join/#storm jml (n=jml@123-243-65-41.tpgi.com.au) |
| 01:11.54 | tjs | I'll see if I can make it available once its finished |
| 01:12.24 | tjs | I think it provides a nice migration patch from zodb to storm |
| 01:12.32 | tjs | without having to go cold turkey |
| 01:14.23 | tjs | is there any way to get a handle on the db connection id from a storm object? |
| 01:17.26 | radix | what's the db connection id? |
| 01:17.47 | radix | there's a reference from the Store to the connection, but it's private |
| 01:18.23 | tjs | radix: its a zstorm thing I guess |
| 01:18.49 | tjs | IZStorm.get(id) |
| 01:19.56 | radix | ah, that's something you register yourself |
| 01:20.24 | radix | with .create() |
| 01:20.45 | radix | and normal storm objects don't know anything about it |
| 01:21.07 | tjs | hm |
| 01:21.16 | radix | we could perhaps add a .get_name_for_store method |
| 01:21.24 | radix | to get the name for a store you have |
| 01:21.29 | radix | but are you sure you need that? |
| 01:22.04 | tjs | well, I need to identify where a storm object comes from |
| 01:22.26 | tjs | if I want to support multiple storm dbs in my crossrefs thing |
| 04:55.37 | tjs | argh |
| 04:57.01 | tjs | http://rafb.net/p/wvF3Gz88.html |
| 04:57.26 | tjs | ^^^ this is a subclass of storm.properties.Pickle |
| 04:57.51 | tjs | take a look at the __set__ method, fairly mundane right? |
| 04:58.15 | tjs | <PROTECTED> |
| 04:58.16 | tjs | <PROTECTED> |
| 04:58.19 | tjs | <PROTECTED> |
| 04:58.21 | tjs | <PROTECTED> |
| 04:58.54 | tjs | the column its getting has nothing to do with the class |
| 05:05.43 | jkakar | I've often wanted a store.uri property, moreso than the name. |
| 05:06.43 | tjs | jkakar: anything that is unique would be fine by me |
| 05:07.23 | tjs | jkakar: any idea whats going on with my crazy Pickle property? |
| 05:09.26 | jkakar | tjs: No, sorry. :( |
| 05:09.48 | tjs | np |
| 05:10.05 | tjs | hmm :/ |
| 07:27.26 | *** join/#storm mcella (n=michele@ip-176-207.sn2.eutelia.it) |
| 08:16.49 | *** join/#storm rafael (n=rafael@moinmoin/fan/rafael) |
| 08:16.58 | rafael | hoi |
| 12:25.48 | *** join/#storm statik (n=emurphy@189.66.188.72.cfl.res.rr.com) |
| 12:27.05 | statik | hi, https://bugs.launchpad.net/storm/+bug/127812 is ready to be merged I think. Funny thing, I missed the bug report last night and wrote exactly the same trivial patch. I guess that counts as a review, right? |
| 12:27.06 | ubot3_ | Malone bug 127812 in storm "Error on using sqlite3 module from Python2.5" [Undecided,Fix committed] |
| 12:30.13 | *** join/#storm bigdog (n=scmikes@49-66-14-216-arpa.cust.cinci.current.net) |
| 13:07.26 | *** join/#storm nephish_ (n=piv@63.98.191.58) |
| 13:07.47 | nephish_ | morning all ! |
| 13:11.09 | statik | moin |
| 13:19.00 | *** join/#storm dinho (n=dinho@201008076130.user.veloxzone.com.br) |
| 13:28.18 | *** join/#storm Lllama (n=Fingram@217.154.92.11) |
| 13:35.04 | *** join/#storm bigdog (n=scmikes@49-66-14-216-arpa.cust.cinci.current.net) |
| 13:51.48 | *** join/#storm dinho_ (n=dinho@201008110135.user.veloxzone.com.br) |
| 15:28.47 | *** join/#storm rafael (n=rafael@moinmoin/fan/rafael) |
| 16:23.22 | statik | radix: for bug 127812, I looked at the suggested patch, and have pushed a branch with a slight correction. whats the next step to getting it merged, does it need a second review? |
| 16:23.23 | ubot3_ | Malone bug 127812 in storm "Error on using sqlite3 module from Python2.5" [Undecided,In progress] https://launchpad.net/bugs/127812 |
| 16:23.34 | radix | it needs another review, yeah |
| 16:23.50 | radix | in fact I'll review it now, I just finished a landscape review |
| 16:23.55 | statik | awesome |
| 16:27.46 | radix | statik: do the tests pass for you on that branch? |
| 16:28.06 | statik | radix: yeah, it fixes it so that the sqlite tests work on python2.6 |
| 16:28.14 | radix | hm |
| 16:28.18 | radix | statik: they don't seem to pass on 2.5 |
| 16:28.43 | radix | actually whoah, dangit |
| 16:29.16 | statik | dangit sounds bad |
| 16:29.36 | radix | statik: yeah, it definitely does not work on python2.5, whether I have the third-party pysqlite2 installed or I'm using the built-in python2.5 version |
| 16:29.48 | radix | statik: and they have different errors |
| 16:29.51 | statik | odd |
| 16:30.37 | statik | so even on the python2.5 in ubuntu it doesn't work? |
| 16:30.54 | radix | without the package, I get an error doing the install_exceptions. with the package, I get a "DatabaseModuleError not raised" |
| 16:31.02 | radix | right, python2.5 in ubuntu |
| 16:31.24 | statik | hmm |
| 16:31.30 | statik | I get it too |
| 16:31.37 | statik | I don't understand what that test is testing |
| 16:31.39 | radix | unfortunately the install_exceptions thing is pretty hideous |
| 16:31.57 | radix | IMO storm shouldn't really be doing what it does with exceptions but instead catching and reraising exceptions |
| 16:32.01 | radix | but whatever, too late now :) |
| 16:32.22 | statik | so, there is another spot in the code that makes reference to sqlite |
| 16:32.56 | statik | ./tests/databases/sqlite.py has a line about dbapi_module_name = "pysqlite2" |
| 16:32.58 | radix | ah, is that the problem? |
| 16:33.00 | statik | what does that do? |
| 16:33.24 | radix | ahhhh |
| 16:33.25 | statik | ooh, that is even the same test that is failing |
| 16:33.36 | radix | statik: look at UnsupportedDatabaseTest, which is the superclass |
| 16:34.01 | radix | statik: it's testing that storm doesn't blow up horribly when a database module isn't found |
| 16:34.13 | radix | statik: that test may need to be slightly extended to support sqlite, since it tries multiple modules now |
| 16:34.39 | radix | it tests it by hacking sys.modules to make the dbapi module None, which basically causes Python to think the module doesn't exist |
| 16:34.47 | radix | s/dbapi/specified dbapi/ |
| 16:34.53 | radix | :) |
| 16:36.43 | statik | ok, I'm following the code so far. but we assert about a DatabaseModuleError, when the shim is raising an ImportError |
| 16:37.22 | radix | statik: storm is intended to raise the DatabaseModuleError when create_from_uri is called with a URI that refers to a databes for which a module is unavailable |
| 16:37.49 | statik | ah, I see |
| 16:38.03 | radix | statik: the reason it's not being raised is because create_from_uri *is* actually finding the module, because it's not "pysqlite2" which the subclass is specifying as the module name |
| 16:38.50 | radix | statik: I think that UnsupportedDatabaseTest will need to be extended to support *lists* of module names to "patch out of existence" as it were |
| 16:39.26 | radix | whoah. that test does some weird shit :) |
| 16:39.39 | statik | but for sqlite, this test is not valid on 2.5 and 2.6 |
| 16:39.52 | statik | I mean, it's not testing anything meaningful about sqlite |
| 16:39.57 | radix | hmm |
| 16:40.01 | radix | I see what you mean, since it's always there. |
| 16:40.05 | radix | but is it actually always there? |
| 16:40.19 | radix | is it a compile time option? is it there on windows or mac OS X if sqlite isn't installed? etc |
| 16:40.53 | statik | it's in the stdlib, and the source to sqlite is actually included in python, so it's not like the weird optional stdlib modules that disappear sometimes |
| 16:41.35 | radix | oh, I see. ok. |
| 16:41.53 | radix | statik: then perhaps just making that test skip successfully on 2.5 and upwards is teh thing to do |
| 16:42.02 | statik | ok |
| 16:42.15 | statik | I'm double-checking my assertion about the source to sqlite being in python |
| 16:42.21 | statik | how do you skip tests with this testrunner? |
| 16:43.22 | radix | statik: you probably don't |
| 16:43.36 | radix | statik: it's stdlib unittest |
| 16:43.39 | jkakar | I think manipulating Test.is_supported is the only mechanism for determining whether to run tests in a suite. |
| 16:43.40 | *** part/#storm Lllama (n=Fingram@217.154.92.11) |
| 16:43.52 | jkakar | Or rather, is the specialized mechanism Storm uses. |
| 16:44.10 | radix | statik: I suggest simply overriding test_exception_when_unsupported in the SQLite-specific test, and returning early if running on 2.5+ before upcalling. |
| 16:44.29 | radix | jkakar: hm, I hadn't noticed that |
| 16:45.06 | radix | that might actually work, |
| 16:45.34 | radix | def is_supported(self): return sys.version_info[:2] < (2, 5) |
| 16:47.35 | statik | yep, that works |
| 16:47.38 | jkakar | Cool |
| 16:48.45 | statik | pushed |
| 16:50.35 | radix | statik: hmm, that works, accidentally :) |
| 16:50.45 | radix | statik: you should be using (2, 5) not (2.5) |
| 16:51.08 | radix | good old python undefined arbitrary comparison results |
| 16:51.28 | statik | doh! |
| 16:51.36 | statik | I rely on the tests to catch my typos |
| 16:52.37 | radix | and you rely on reviewers to catch the typos in your tests :) |
| 16:53.19 | statik | and you are doing a fine job of it :) pushed again |
| 16:53.29 | mcella | hey guys, can I tell storm to select only some columns? ie avoid select * |
| 16:54.34 | radix | mcella: result_set.values('col1', 'col2') -> iterator of (col1val, col2val) |
| 16:54.48 | mcella | cool :-) |
| 17:09.11 | mcella | radix: tnx, works well |
| 18:42.04 | rafael | hmm - why can't i make c = Customer(name='asdf', field2='...') ? |
| 18:55.13 | *** join/#storm dinho (n=dinho@201008208031.user.veloxzone.com.br) |
| 19:04.43 | radix | rafael: I don't know, why can't you? |
| 19:09.39 | rafael | TypeError: default __new__ takes no parameters |
| 19:09.58 | radix | rafael: maybe you want to provide an __init__ method that takes those parameters |
| 19:10.34 | rafael | ah yes, that's a solution :) |
| 20:14.13 | *** join/#storm nephish_ (n=piv@63.98.191.58) |
| 20:20.16 | *** join/#storm rafael (n=rafael@moinmoin/fan/rafael) |
| 20:26.19 | *** join/#storm oubiwann_ (n=oubiwann@72.164.56.1) |
| 22:39.20 | tjs | morning |
| 22:39.42 | tjs | anyone tried subclassing anything in properties and overriding __set__ with any success? |
| 22:40.28 | tjs | I think there may be something not right with how properties.Property.__set__ finds its class |
| 22:40.37 | tjs | or something |
| 22:40.46 | tjs | http://pastebin.com/d1f41fe29 |
| 22:41.01 | tjs | <PROTECTED> |
| 22:41.02 | tjs | <PROTECTED> |
| 22:41.05 | tjs | <PROTECTED> |
| 22:41.07 | tjs | <PROTECTED> |
| 22:41.39 | tjs | it finds PropertyColumns that have nothing to do with the class the descriptor was bound to |
| 22:41.45 | tjs | I think |
| 22:50.23 | *** join/#storm jamesh (n=james@canonical/launchpad/jamesh) |
| 23:15.42 | tjs | anyone? |
| 23:34.38 | *** join/#storm jml (n=jml@123-243-65-41.tpgi.com.au) |
| 23:40.48 | jkakar | Heya tjs |
| 23:46.58 | tjs | jkakar: hey |
| 23:47.09 | jkakar | tjs: Heya |
| 23:47.52 | tjs | jkakar: how do you guys to url->storm object traversal? |
| 23:47.59 | tjs | is that stuff going to be in zstorm? |
| 23:48.24 | jkakar | tjs: We use internal code for that; it's totally separate from zstorm (and not planned to be release anytime soon, as I know it). |
| 23:48.35 | tjs | fair enough |
| 23:49.22 | jkakar | tjs: Basically, we've implement a routes mapper that integrates with Zope 3's object traversal system. |
| 23:49.53 | tjs | ok |