2008-3-11
| 23:13 | Spads | brosner: heh, oddly enough many of those are about the guitarist :) |
| 23:13 | Magus- | mysite.urls = mysite/urls.py, of course |
| 23:13 | Magus- | just like mysite.settings = mysite/settings.py |
| 23:13 | Magus- | standard python module referencing |
| 23:13 | flaco | thks |
| 23:16 | Leonidas | how can I exclude a specific object from a queryset? |
| 23:16 | Magus- | .exclude() of course |
| 23:17 | Leonidas | Magus-: I tried that, but the resulting queryset was broken |
| 23:17 | EyePulp_ | of course. |
| 23:17 | EyePulp_ | (not of course?) |
| 23:17 | Magus- | Leonidas: then you did it wrong |
| 23:17 | Magus- | pastebin some code :) |
| 23:17 | Leonidas | Magus-: yeah, I suppose so, moment |
| 23:17 | Magus- | (or you're using a query that's currently broken) |
| 23:18 | Magus- | there are a few edge cases that don't work right at the moment |
| 23:19 | Leonidas | Magus-: http://dpaste.com/38997/ |
| 23:20 | Leonidas | So, I try to display only series which have more than 0 public_viewable watches |
| 23:21 | Leonidas | The condition was too hard for a simple exlude, at least I couldn't come up with something. |
| 23:22 | Magus- | .exclude(obj) is invalid, of course |
| 23:22 | Magus- | you have to use kwargs like every normal filter/etc |
| 23:22 | Magus- | something like .exclude(id=obj.id) would work |
| 23:22 | Leonidas | Magus-: the fun thing is, it does not crash. It crashes on the print |
| 23:22 | Magus- | well duh, querysets aren't evalutaed until you try to get data :) |
| 23:22 | sillyperson | Hi, I have a Q-related question. Due to a design-issue a object has a start_date field and a start_time field, and i want to get a queryset of objs with a date/time greater than right now, the pseudocode is something like: queryset = filter(start_time__gt=now.time() AND start_date=now.date()) OR start_date__gt=now.date(), -- as i read it, Q would do the job with: filter(Q(start_time__gt=now.time |
| 23:22 | sillyperson | (), start_date=now.date()) | Q(start_date__gt=now.date())) but I must be mistaken because that produces faulty output... any pointers? |
| 23:22 | Magus- | so you go and build the invalid one, and on the print it dies |
| 23:23 | notanumber | Can someone tell me why, when I update 'reports_to' in this: http://dpaste.com/38998/, 'supervises' also updates with the exact same content? |
| 23:23 | Magus- | because you didn't turn off the symmetrical option |
| 23:24 | Magus- | sillyperson: try using & to combine the first two, instead of foo,bar |
| 23:25 | gordonjcp | hrm |
| 23:25 | gordonjcp | I'm trying to use modelforms to edit a model |
| 23:26 | gordonjcp | how do I get it to populate the existing values of that particular object into the form? |
| 23:26 | notanumber | Magus, really? I read the description of symmetricl, but it didn'tseem to apply in this case. It did work when I put it on though, but thought I might it be a bug. Guess I was just reading it wrong. |
| 23:26 | sillyperson | christ!!! ... Magus- you're a saint :-) |
| 23:26 | Magus- | notanumber: symmetrical = what you add to one m2m is added to the other side of it |
| 23:26 | cccp | what should i do if i want to format a datetimefield in a newforms form(modemform) before outputting (i need 12.3.07 instead of 2007-3-12)? the boundfield.data just returns None - so what do i need to change? |
| 23:26 | Magus- | in this case, it's likely interpreting the 2nd m2m as the other side |
| 23:26 | Magus- | and any m2m to 'self' is symmetrical by default |
| 23:27 | cccp | modelform* |
| 23:27 | brosner | gordonjcp: pass in the instance. |
| 23:28 | Magus- | cccp: as far as I know you cannot change the initial display of the data in the form |
| 23:28 | Magus- | you could try a custom datefield and/or widget though |
| 23:28 | gordonjcp | brosner: pass it in where? |
| 23:28 | cccp | hm :/ |
| 23:28 | brosner | gordonjcp: MyModelForm(instance=obj) as the docs say |
| 23:29 | notanumber | Magus, but that's still the case for two seperate m2m fields? |
| 23:29 | cccp | can i easily replace the normal widget with it? (i use a subset of the fields via a modelform) |
| 23:29 | Magus- | notanumber: every m2m to 'self' is symmetrical=True |
| 23:29 | Magus- | you could make 50, they'd all have that set |
| 23:29 | AzMoo | OK, so I need to authenticate with our AD server and mattmcc recommended a custom backend for the auth middleware. However, I need support for our AD groups as well. From what I've seen of the backend it's for user authentication only and groups are implemented as a db thing. Is there a way I can extend the Group functionality? |
| 23:30 | gordonjcp | brosner: aha, thanks |
| 23:30 | notanumber | Magus, oh. I see what what you're saying. That's an odd behaviour, isn't it? WHy would you want that. |
| 23:30 | Magus- | no it isn't |
| 23:30 | gordonjcp | brosner: next problem is that dropdowns don't work |
| 23:30 | Magus- | for instance, suppose you had a userprofile with a 'friends' m2m to self |
| 23:30 | Magus- | if user A is friends with B, B is friends with A |
| 23:31 | Magus- | symmetrical handles the other end automatically |
| 23:31 | wolfe | tada |
| 23:31 | wolfe | :P |
| 23:31 | gordonjcp | hrmm, this is going to take either more beer or more sleep to fix |
| 23:31 | brosner | gordonjcp: more info please |
| 23:32 | Magus- | notanumber: but since that's not /always/ desired, the option can be disabled |
| 23:33 | notanumber | Magus, that part makes sense. My case has "user" with two m2m fields: supervises & reports_to. Changing either one in a user records, changes the other. That's the part that seems odd to me. |
| 23:34 | notanumber | I would have thought they'd be distinct. |
| 23:34 | Magus- | notanumber: and normally they would be |
| 23:34 | Magus- | I think the symmetrical is getting in the way |
| 23:34 | Magus- | it may be a bug with that option, I don't know |
| 23:35 | notanumber | Magus, I thought it might be a bug. Didn't see a report for it though, so thought I'd confirm. |
| 23:35 | Magus- | I'm not entirely sure :) |
| 23:35 | Leonidas | Magus-: thanks, I got it working properly. |
| 23:35 | notanumber | Symmetricl=false did fix it though. |
| 23:39 | notanumber | Maybe I'll submit a bug report either way. Let the developers decide if it's a bug or not. |
| 23:48 | notanumber | Magus, actually, I lied. S=symmetrical=False didn't work. Now adding a user, "Bill |
| 23:49 | notanumber | , whom reports to "John". Added Both "John", & "Bill" to 'reports_to' & 'supervises' for John. |
| 23:49 | notanumber | and Bill |
| 23:52 | Magus- | hmm |
| 23:52 | Magus- | you could try using db_table to specify different tables for each m2m |
| 23:52 | Magus- | but definitely file a bug for that one |
| 23:53 | notanumber | Well, at least I know I'm not going crazy. heh. Can't type for crap tonight though... |
| 23:55 | speedy | whats the best way to extend user class? |
| 23:55 | Magus- | behold, the power of google |
| 23:55 | Magus- | |
| 23:55 | Magus- | first hit |
| 23:56 | speedy | thanks magus, i went thru this post, but there are other methods, i'm asking about the best |
| 23:56 | Magus- | no there aren't |
| 23:56 | empty | what other methods? |
| 23:56 | Magus- | that's the only method |
| 23:57 | speedy | |
| 23:57 | Magus- | that method is absolute shit |
| 23:57 | Magus- | so yah, use the b-list one |
| 23:58 | speedy | alright |
| 23:58 | malibu | lol.. |
| 23:58 | empty that's just awful | |
| 23:58 | speedy | yes it is lol |
| 23:59 | brosner | omg, seriously... what a shame |
Page 25 of 25
← Previous
(2,495 total)