| 00:00.19 | *** join/#storm wallflower (n=wallflow@ip-205-246-113-216.pool.grokthis.net) |
| 02:15.48 | shaunm | jamesh: so I have field triggers in pulse, functions that get called when a particular field gets changed |
| 02:16.06 | jamesh | eek |
| 02:16.29 | jamesh | database level triggers are a pain to handle with ORMs |
| 02:16.31 | shaunm | and they'd be a *lot* cleaner if they piggybacked off the "changed" event in storm, I think |
| 02:16.38 | shaunm | no no |
| 02:16.45 | shaunm | they're python-level |
| 02:17.06 | jamesh | shaunm: okay. Perhaps validators will do what you want? |
| 02:18.06 | shaunm | no, they're intended to allow plugins to do stuff when the database is updated in ways that are interesting to them |
| 02:30.32 | *** join/#storm jbot (i=ibot@rikers.org) |
| 02:30.32 | *** topic/#storm is The Storm Python ORM - http://storm.canonical.com/ - 0.14 released! || Review branches: https://code.launchpad.net/~storm/storm/trunk/+landing-candidates |
| 02:30.59 | jamesh | hmm. Looks like it hasn't gotten into the tutorial. |
| 02:31.17 | jamesh | <PROTECTED> |
| 02:37.29 | shaunm | I've been largely learning how to use storm by reading the source |
| 02:37.56 | shaunm | which is sometimes annoying, but sometimes informative and fun |
| 02:38.14 | shaunm | but could often be leading me to digging into places you don't want me digging into |
| 02:49.26 | jamesh | shaunm: in some cases, the reason a hook doesn't exist is because it wouldn't be reliable |
| 02:50.16 | jamesh | e.g. a hook for object deletion would not be able to reliably handle objects deleted via ResultSet.delete() |
| 02:51.04 | shaunm | mm, hadn't thought of that |
| 02:51.48 | shaunm | so for a field change notification, it wouldn't exist when doing a bulk update on a ResultSet |
| 02:52.17 | jamesh | probably not. |
| 02:52.25 | jamesh | that is, no. |
| 02:52.36 | jamesh | for ResultSet.set(), that is. |
| 02:52.40 | shaunm | yeah |
| 02:53.03 | shaunm | that's a limitation I hadn't considered, and one that my current code suffers from of course |
| 02:53.41 | shaunm | but, for my purposes, it's something I can deal with |
| 03:05.19 | *** join/#storm bigdog (n=scmikes@72-197-8-8-arpa.cust.cinci.current.net) |
| 03:17.13 | shaunm | storm does not call __init__ for objects it loads from the database? |
| 03:17.19 | shaunm | like from a ResultSet |
| 03:18.14 | shaunm | meaning I'm free to do whatever the hell I want with __init__ and I won't mess up storm? |
| 03:29.53 | jamesh | it'll call __storm_loaded__() though, if you need to do something for objects loaded from the DB |
| 03:30.06 | jamesh | __init__() is yours to do with as you please. |
| 03:30.19 | shaunm | that's handy |
| 03:31.02 | shaunm | that's one more classmethod I can do away with |
| 05:54.39 | *** join/#storm jukart (n=jukart@d91-128-122-97.cust.tele2.at) |
| 08:58.03 | thumper | https://bugs.edge.launchpad.net/storm/+bug/324724 |
| 08:58.04 | mup | Bug #324724: local_variables_are_none() check in Reference.__get__() not always appropriate <Storm:New> <https://launchpad.net/bugs/324724> |
| 08:58.04 | thumper | poke |
| 08:58.33 | thumper | james_h: thanks for all you effort getting this concise bug report |
| 09:07.41 | *** join/#storm niemeyer (n=niemeyer@host217-45-210-54.in-addr.btopenworld.com) |
| 09:10.47 | jamesh | niemeyer: thumper was asking about https://bugs.edge.launchpad.net/storm/+bug/324724 just before you joined. Do you have any thoughts? |
| 09:10.48 | mup | Bug #324724: local_variables_are_none() check in Reference.__get__() not always appropriate <Storm:New> <https://launchpad.net/bugs/324724> |
| 09:14.44 | *** join/#storm mup_ (n=mup@li37-90.members.linode.com) |
| 09:28.12 | niemeyer | jamesh: I'll have to investigate it a bit before providing any feedback.. I'll try to do that today |
| 09:32.06 | *** join/#storm jkakar_ (n=jkakar@host217-45-210-54.in-addr.btopenworld.com) |
| 09:32.31 | jamesh | niemeyer: that'll make thumper happy :) |
| 09:32.49 | *** join/#storm oubiwann (n=oubiwann@host217-45-210-54.in-addr.btopenworld.com) |
| 09:33.00 | niemeyer | jamesh: We'll also try to run over your branches |
| 09:33.22 | jamesh | and that will make me happy :) |
| 09:33.34 | jamesh | I don't think any of them are particularly complicated |
| 09:34.12 | thumper | niemeyer: thanks |
| 09:35.52 | *** join/#storm goschtl (n=goschtl@p5B0BCC34.dip.t-dialin.net) |
| 09:36.13 | niemeyer | jamesh: Sweet.. we'll probably be able to get over all of them then.. being on the same room always helps |
| 09:36.18 | niemeyer | thumper: No problem! |
| 09:37.43 | jamesh | niemeyer: I should be in almost the same time zone as you next week (give or take a few hours) |
| 09:38.23 | jamesh | (a sprint in Montreal) |
| 09:39.51 | *** join/#storm sidnei (n=sidnei@plone/dreamcatcher) |
| 09:41.59 | jamesh | I wonder how difficult it'd be to implement "from __past__ import print_statement" |
| 09:47.58 | niemeyer | jamesh: Oh, nice.. maybe we could take that chance for some further discussions |
| 09:48.09 | niemeyer | jamesh: That IRC meeting, for instance |
| 09:48.46 | niemeyer | jamesh: Where are you missing that magic import? |
| 09:50.05 | *** join/#storm jukart (i=lovely@81.189.156.94) |
| 09:50.34 | jamesh | niemeyer: nowhere. If it wasn't too difficult, it might make a nice april fools patch. |
| 09:50.52 | niemeyer | jamesh: Hmmm.. :-) |
| 09:51.21 | *** join/#storm philn (n=phil@o.bcn.fluendo.net) |
| 09:51.25 | philn | hi |
| 09:52.35 | philn | storm objects don't seem to support pickle.. is it planned to add it? |
| 09:53.12 | jamesh | philn: what would you expect to happen when you unpickle a storm object? |
| 09:55.35 | philn | reload its attributes from db? |
| 09:55.51 | jamesh | with what connection? |
| 09:55.57 | philn | you tell me :) |
| 09:56.27 | philn | i guess it's tricky |
| 09:58.52 | jamesh | and assuming you do find a connection, what if the referenced row doesn't exist anymore? |
| 10:00.55 | philn | yeah... crap |
| 10:02.56 | jamesh | so in answer to your initial question, I don't think anyone is working on pickle support |
| 10:03.20 | jamesh | partly because it is a hard problem, and partly because it isn't clear that it is a good idea |
| 10:04.13 | philn | i understand now.. will have to find another solution for my problem |
| 10:10.41 | niemeyer | jamesh: Hmm |
| 10:11.28 | niemeyer | jamesh, thumper: I think the least disrupting solution for the None reference problem could be to flush the store on the __get__ if the reference is None |
| 10:12.10 | niemeyer | I'll have a try to see if it fixes the test case provided |
| 10:13.04 | jamesh | niemeyer: that's kind of what the branch attached to the bug does. |
| 10:13.36 | jamesh | it does it from local_variables_are_none(), which is called by __get__() |
| 10:13.47 | niemeyer | jamesh: Oh |
| 10:14.19 | niemeyer | jamesh: Let me check it out.. I probably misinterpreted your last comment then |
| 10:14.22 | jamesh | we have a test that checks that we don't touch the DB in this case (for forward refs at least) |
| 10:14.38 | jamesh | I am not sure if it is a useful feature though |
| 10:14.50 | niemeyer | jamesh: I had the impression it was doing the get at all times.. sorry for not reading it upfront instead of guessing |
| 10:15.45 | niemeyer | jamesh: It actually is.. we used to have tons of dummy queries in Landscape because of that |
| 10:15.57 | niemeyer | jamesh: It's most important with nullable references |
| 10:16.21 | jamesh | niemeyer: there is a difference between the local var being None and the local variable being unset though. |
| 10:16.31 | jamesh | s/unset/unknown/ |
| 10:16.32 | niemeyer | jamesh: Without the current behavior, when the reference is NULL whenever someone does foo.ref is None, it touches the database |
| 10:17.04 | jamesh | niemeyer: there is a separate test for that which doesn't fail with my change |
| 10:17.36 | jamesh | I agree that that is a desirable feature to keep |
| 10:18.12 | niemeyer | jamesh: Hmm.. is the issue reported in the bug specific to unset variables? |
| 10:18.13 | jamesh | but e.g. if I invalidate an object and then try to access a reference attribute, I'd expect to hit the DB to look up the local key. |
| 10:18.54 | jamesh | niemeyer: yes. thumper was using a back reference (on_remote=True), and the local key was the primary key |
| 10:19.23 | jamesh | the object was in pending add state, so had an unset primary key |
| 10:19.57 | jamesh | a relation had been established using a reference going in the opposite direction. |
| 10:20.50 | niemeyer | jamesh: Hmm, ok |
| 10:24.04 | niemeyer | jamesh: Alright, I think I understand better the problem |
| 10:24.21 | niemeyer | jamesh: Your branch looks great |
| 10:24.48 | niemeyer | jamesh: test_reference_wont_touch_store_when_key_is_unset indeed doesn't look very useful |
| 10:42.30 | niemeyer | jamesh: I got hold of therve for the second test and he's happy with it too.. so it's ready to go in |
| 10:43.17 | jamesh | niemeyer: cool. I'll drop that test and merge it. |
| 10:46.04 | *** part/#storm philn (n=phil@o.bcn.fluendo.net) |
| 10:52.40 | niemeyer | jamesh: Sweet, thank you! |
| 11:12.51 | niemeyer | jamesh: https://code.edge.launchpad.net/~jamesh/storm/uuid/+merge/3365 is ready to go too |
| 11:44.00 | mup | storm/twisted-integration r215 committed by thomas@canonical.com |
| 11:44.01 | mup | Merge from trunk. |
| 11:57.14 | niemeyer | jamesh: We've got a broken test in Landscape due to the flushing on references |
| 11:58.16 | *** join/#storm jukart (i=lovely@81.189.156.94) |
| 11:59.56 | niemeyer | jamesh: The problem is that we're building certain objects in memory, and passing them around to build further objects |
| 12:00.15 | niemeyer | jamesh: and the additional logic is forcing flushes in the middle of the procedure |
| 12:00.44 | niemeyer | jamesh: Which means that what would be only a few CREATEs become unfinished CREATEs + UPDATEs |
| 12:01.07 | niemeyer | That can potentially break up with IntegrityErrors because it's flushing mid-way through constructors :-( |
| 12:06.29 | *** join/#storm artista-frustrad (n=artista_@201-24-232-112.ctame704.dsl.brasiltelecom.net.br) |
| 14:06.27 | *** part/#storm bigdog (n=scmikes@72-197-8-8-arpa.cust.cinci.current.net) |
| 14:33.16 | *** join/#storm mcayer (n=chatzill@modemcable250.87-70-69.static.videotron.ca) |
| 14:33.23 | mcayer | hi! |
| 14:33.28 | mcayer | i need a little help |
| 14:35.01 | mcayer | could somebody give me a hand |
| 14:35.03 | mcayer | ? |
| 14:35.41 | therve | mcayer: don't hesitate to ask |
| 14:35.48 | mcayer | :) |
| 14:35.53 | mcayer | here's my problem |
| 14:36.12 | mcayer | I'm trying to insert a row in a table |
| 14:36.18 | mcayer | nothing special here |
| 14:36.32 | mcayer | however, on the store.flush(), I get this error |
| 14:36.47 | mcayer | OperationalError: (1205, 'Lock wait timeout exceeded; try restarting transaction') |
| 14:36.53 | mcayer | everytime |
| 14:37.10 | therve | are you calling store.flush() ? |
| 14:37.16 | mcayer | I try this script on another server, and it works flawlesslt |
| 14:37.19 | mcayer | yep |
| 14:37.30 | therve | why? |
| 14:38.01 | mcayer | because I need to have the id of the row generated by the auto_increment to use it in another row insert in another table |
| 14:38.17 | therve | you don't need to do that |
| 14:38.22 | mcayer | ? |
| 14:38.28 | therve | you could use default=AutoReload for example |
| 14:38.37 | mcayer | in the mysql config? |
| 14:38.40 | therve | or don't use the id, and use a reference instead |
| 14:38.50 | therve | no, in your storm object definition |
| 14:38.58 | mcayer | k |
| 14:39.21 | mcayer | the other table has a foreign key with the first |
| 14:39.33 | mcayer | so I need an id |
| 14:39.40 | mcayer | to make it's reference |
| 14:39.56 | therve | if you use a storm Reference object, you can assign directly to it |
| 14:40.29 | mcayer | you mean passing it the first table object instead of just the id? |
| 14:41.21 | therve | right |
| 14:41.34 | mcayer | k |
| 14:41.51 | mcayer | what to see my table definitions? to check if I got it right? |
| 14:41.54 | therve | instead of doing "self.other_object_id = other_object.id", you can do "self.other_object = other_object" |
| 14:42.09 | therve | if other_object is a Reference definition |
| 14:42.26 | mcayer | in the second table |
| 14:43.30 | *** join/#storm vvinet (n=vince@132.210.76.204) |
| 14:44.09 | mcayer | I got ReferenceSet in the first table |
| 14:44.31 | mcayer | I need now to add a Reference in the second table, pointing to the first table? |
| 14:48.59 | therve | or you can add to your ReferenceSet |
| 14:49.48 | mcayer | I don't understand |
| 14:50.01 | mcayer | want to see my code? to see what is done? |
| 14:51.07 | therve | if it's not too huge |
| 14:51.18 | mcayer | no |
| 14:51.29 | mcayer | I'll only give you the two tables definitions :) |
| 14:51.53 | mcayer | where can I paste it? |
| 14:54.08 | therve | pastebin.ubuntu.com |
| 14:56.38 | mcayer | http://pastebin.ubuntu.com/120156/ |
| 14:56.41 | mcayer | there you go |
| 15:03.52 | mcayer | so? what am I missing? |
| 15:06.59 | therve | you can use Plume.contours to link Contour objects to Plume |
| 15:07.24 | therve | if plume1 is an instance of Plume, you can do plume1.contours.add(contour1) |
| 15:07.32 | mcayer | i see |
| 15:07.48 | mcayer | and not doing the store.flush() |
| 15:08.04 | mcayer | only doing the commit() at the end |
| 15:09.38 | therve | right |
| 15:09.53 | mcayer | :) |
| 15:10.09 | mcayer | thanks a lot, I'll modify my code and test it right away |
| 15:14.00 | mup | storm/twisted-integration r216 committed by thomas@canonical.com |
| 15:14.00 | mup | Move the tests to be able to run Postgres/MySQL tests as well |
| 15:15.14 | *** join/#storm bigdog (n=scmikes@72-197-8-8-arpa.cust.cinci.current.net) |
| 15:24.02 | *** part/#storm goschtl (n=goschtl@p5B0BCC34.dip.t-dialin.net) |
| 15:24.13 | mcayer | grrrr |
| 15:24.19 | mcayer | the commit gives me the same error |
| 15:28.11 | niemeyer | thumper: ping |
| 16:35.31 | mup | storm/twisted-integration r217 committed by thomas@canonical.com |
| 16:35.31 | mup | Add postgres tests |
| 16:41.06 | mup | storm/twisted-integration r218 committed by thomas@canonical.com |
| 16:41.06 | mup | Add mysql, some tests failing. |
| 16:44.50 | mup | storm/trunk r294 committed by gustavo@niemeyer.net |
| 16:44.50 | mup | Rolling back last change. |
| 16:44.50 | mup | As a side-effect of the introduced change, objects being constructed |
| 16:44.50 | mup | could be flushed to the database inside their constructors simply by |
| 16:44.51 | mup | accessing a reference which isn't set. Besides being undesired due |
| 16:44.53 | mup | to the additional UPDATE queries that are executed after the CREATE, |
| 16:44.55 | mup | this behavior can actually break code if the database expectations |
| 16:44.57 | mup | (e.g. NOT NULL) are not entirely satisfied by the time the object |
| 16:44.59 | mup | is flushed. |
| 16:45.03 | mup | We have ideas for how to implement the desired behavior without |
| 16:45.05 | mup | introducing this side effect, but this will require additional |
| 16:45.07 | mup | research and time for the implementation. |
| 16:46.53 | *** join/#storm andrea-bs (n=andrea-b@ubuntu/member/beeseek.developer.andrea-bs) |
| 16:48.32 | mup | storm/twisted-integration r219 committed by thomas@canonical.com |
| 16:48.32 | mup | Use InnoDB to be able to use transactions. |
| 17:47.45 | *** join/#storm jukart (n=jukart@d91-128-122-97.cust.tele2.at) |
| 17:54.39 | *** join/#storm bac (n=bac@canonical/launchpad/bac) [NETSPLIT VICTIM] |
| 18:49.57 | *** join/#storm andrea-bs (n=andrea-b@ubuntu/member/beeseek.developer.andrea-bs) |
| 19:03.24 | *** join/#storm artista_frustrad (n=artista_@201-40-203-13.ctame704.dsl.brasiltelecom.net.br) |
| 19:17.50 | *** join/#storm shaunm (n=shaunm@proxyserver.wolfram.com) |
| 19:37.23 | *** join/#storm jukart (n=jukart@d91-128-122-97.cust.tele2.at) |
| 20:41.30 | *** join/#storm mcayer (n=chatzill@modemcable250.87-70-69.static.videotron.ca) |
| 20:41.45 | mcayer | hi there! |
| 20:44.52 | mcayer | I got the following error when using store.commit() |
| 20:45.07 | mcayer | 1205, Lock wait timeout exceeded; try restarting transaction |
| 20:45.13 | mcayer | i'm using mysql |
| 20:45.21 | mcayer | could somebody help me on this? |
| 21:59.24 | mcayer | exit |
| 22:56.40 | *** join/#storm sidnei (n=sidnei@plone/dreamcatcher) |
| 23:03.29 | *** join/#storm sidnei_ (n=sidnei@host217-45-210-54.in-addr.btopenworld.com) |
| 23:51.19 | *** join/#storm sidnei (n=sidnei@plone/dreamcatcher) |