/irc-logs / w3c / #css / 2013-07-26 / end

Options:

# Session Start: Fri Jul 26 00:00:00 2013
# Session Ident: #css
# [00:01] * Quits: dbaron (~dbaron@public.cloak) ("8403864 bytes have been tenured, next gc will be global.")
# [00:05] * Quits: abucur (~abucur@public.cloak) ("Leaving")
# [00:30] * Joins: tobie (tobie@public.cloak)
# [00:44] * Quits: plh (plehegar@public.cloak) (Ping timeout: 180 seconds)
# [01:09] * Quits: tobie (tobie@public.cloak)
# [01:10] * Joins: dbaron (~dbaron@public.cloak)
# [01:10] * Quits: dbaron (~dbaron@public.cloak) (Client closed connection)
# [01:10] * Joins: dbaron (~dbaron@public.cloak)
# [01:55] * Joins: jdaggett (~jdaggett@public.cloak)
# [02:05] <TabAtkins> plinss: Yo, your dfn data seems slightly weird. I'm getting multiple definitions for the term "initial" and "none" for specs that do *not* export those terms.
# [02:06] <TabAtkins> I thought only props and descriptors were auto-exported.
# [02:06] <plinss> got an example?
# [02:08] * Quits: cabanier (~cabanier@public.cloak) (Ping timeout: 180 seconds)
# [02:09] <plinss> looking in the DB the only valuedef for initial I see is in css-cascade-3
# [02:09] <plinss> I do see s bunch of 'none's though...
# [02:10] <TabAtkins> I see "initial" defined twice in both values and cascade.
# [02:10] <TabAtkins> FATAL ERROR: Too many xrefs for "initial":
# [02:10] <TabAtkins> {"url": "#initial", "type": "valuedef", "spec": "css-values-3"}
# [02:10] <TabAtkins> {"url": "#lsquoinitialrsquo", "type": "valuedef", "spec": "css-values-3"}
# [02:10] <TabAtkins> {"url": "#initial-keyword", "type": "valuedef", "spec": "css-cascade-3"}
# [02:10] <TabAtkins> {"url": "#initial0", "type": "valuedef", "spec": "css-cascade-3"}
# [02:11] * Joins: danielfilho (~danielfilho@public.cloak)
# [02:12] <TabAtkins> (I'm heading out for the day now - I'll continue this tomorrow.
# [02:12] <TabAtkins> )
# [02:12] <plinss> mkay
# [02:19] <plinss> TabAtkins: FWIW, in both of those cases, they're each defined in the official version and the draft with different anchors, so you're getting one of each
# [02:26] * Quits: lmclister (~lmclister@public.cloak) (lmclister)
# [02:26] <plinss> TabAtkins: also, looking at the code, all valuedefs are auto-exported
# [02:27] <plinss> I just tried turing that off and then they _all_ wind up being internal
# [02:28] <plinss> so it doesn't look like there are any explicitly exported value dfns at the moment…
# [02:32] * liam looks at paris hotels... how definite is the paris f2f do you think?
# [02:57] * Quits: danielfilho (~danielfilho@public.cloak) (Client closed connection)
# [03:34] * Quits: dbaron (~dbaron@public.cloak) ("8403864 bytes have been tenured, next gc will be global.")
# [03:56] * Joins: tobie (tobie@public.cloak)
# [04:02] * Quits: tobie (tobie@public.cloak)
# [04:05] * Joins: cabanier (~cabanier@public.cloak)
# [04:20] * Quits: sgalineau (~sgalineau@public.cloak) (Client closed connection)
# [04:38] * Quits: liam (liam@public.cloak) (Ping timeout: 180 seconds)
# [04:48] * Joins: liam (liam@public.cloak)
# [05:37] * Joins: cabanier1 (~cabanier@public.cloak)
# [05:43] * Quits: cabanier (~cabanier@public.cloak) (Ping timeout: 180 seconds)
# [06:33] <plinss> liam, I believe the paris dates are solid at this point
# [06:50] * Joins: dbaron (~dbaron@public.cloak)
# [07:03] * Joins: sgalineau (~sgalineau@public.cloak)
# [07:05] * Quits: sgalineau (~sgalineau@public.cloak) ("Textual IRC Client: www.textualapp.com")
# [07:25] * Joins: krit1 (~krit@public.cloak)
# [07:26] * Quits: krit (~krit@public.cloak) (Ping timeout: 180 seconds)
# [08:02] * Joins: teoli (~teoli@public.cloak)
# [08:06] * Joins: teoli_ (~teoli@public.cloak)
# [08:13] * Quits: teoli (~teoli@public.cloak) (Ping timeout: 180 seconds)
# [08:17] * Joins: teoli (~teoli@public.cloak)
# [08:21] * Quits: krit1 (~krit@public.cloak) ("Leaving.")
# [08:23] * Quits: teoli_ (~teoli@public.cloak) (Ping timeout: 180 seconds)
# [08:35] * Joins: krit (~krit@public.cloak)
# [09:01] * Joins: nvdbleek (~nvdbleek@public.cloak)
# [09:13] * Quits: dbaron (~dbaron@public.cloak) ("8403864 bytes have been tenured, next gc will be global.")
# [09:20] * Joins: plh (plehegar@public.cloak)
# [09:27] * Quits: plh (plehegar@public.cloak) ("Leaving")
# [09:31] * Joins: michou (~mibalan@public.cloak)
# [09:47] * Quits: nvdbleek (~nvdbleek@public.cloak) (nvdbleek)
# [09:47] * Joins: nvdbleek (~nvdbleek@public.cloak)
# [09:47] * Quits: nvdbleek (~nvdbleek@public.cloak) (nvdbleek)
# [09:47] * Joins: nvdbleek (~nvdbleek@public.cloak)
# [09:49] * Quits: nvdbleek (~nvdbleek@public.cloak) (nvdbleek)
# [09:49] * Joins: nvdbleek (~nvdbleek@public.cloak)
# [10:13] * Quits: jdaggett (~jdaggett@public.cloak) (jdaggett)
# [11:03] * Quits: nvdbleek (~nvdbleek@public.cloak) (nvdbleek)
# [11:13] * Joins: antonp (~Thunderbird@public.cloak)
# [11:28] * Quits: michou (~mibalan@public.cloak) (Client closed connection)
# [11:28] * Joins: michou (~mibalan@public.cloak)
# [11:41] * Quits: antonp (~Thunderbird@public.cloak) (antonp)
# [12:32] * Quits: liam (liam@public.cloak) (Ping timeout: 180 seconds)
# [12:43] * Joins: liam (liam@public.cloak)
# [13:29] * Joins: darktears (~darktears@public.cloak)
# [13:29] * Quits: {Darktears} (~darktears@public.cloak) (Client closed connection)
# [13:34] * Joins: {Darktears} (~darktears@public.cloak)
# [13:35] * Quits: darktears (~darktears@public.cloak) (Client closed connection)
# [13:47] * Joins: jdaggett (~jdaggett@public.cloak)
# [13:55] * Joins: plh (plehegar@public.cloak)
# [14:16] * leaverou_away is now known as leaverou
# [14:22] * Quits: jdaggett (~jdaggett@public.cloak) (jdaggett)
# [15:28] * Quits: krit (~krit@public.cloak) (Ping timeout: 180 seconds)
# [15:30] * Joins: krit (~krit@public.cloak)
# [15:31] * Joins: teoli_ (~teoli@public.cloak)
# [15:31] * Quits: teoli (~teoli@public.cloak) (Client closed connection)
# [16:09] * Quits: plh (plehegar@public.cloak) ("Leaving")
# [16:20] * Joins: antonp (~Thunderbird@public.cloak)
# [16:26] <liam> plinss, thanks [Paris]
# [16:31] * Quits: antonp (~Thunderbird@public.cloak) (antonp)
# [16:44] * Joins: sgalineau (~sgalineau@public.cloak)
# [16:44] * Quits: sgalineau (~sgalineau@public.cloak) (sgalineau)
# [16:45] * Joins: sgalineau (~sgalineau@public.cloak)
# [16:55] * Joins: teoli (~teoli@public.cloak)
# [16:57] * Joins: plh (plehegar@public.cloak)
# [17:00] * Joins: teoli__ (~teoli@public.cloak)
# [17:00] * Quits: teoli_ (~teoli@public.cloak) (Client closed connection)
# [17:05] * Quits: teoli (~teoli@public.cloak) (Ping timeout: 180 seconds)
# [17:08] * Quits: michou (~mibalan@public.cloak) ("Leaving.")
# [17:08] * Joins: teoli (~teoli@public.cloak)
# [17:11] * Quits: teoli (~teoli@public.cloak) (Client closed connection)
# [17:12] * Joins: teoli (~teoli@public.cloak)
# [17:15] * Quits: teoli__ (~teoli@public.cloak) (Ping timeout: 180 seconds)
# [17:22] * Quits: teoli (~teoli@public.cloak) (Client closed connection)
# [17:22] * Joins: teoli (~teoli@public.cloak)
# [17:44] * Quits: cabanier1 (~cabanier@public.cloak) (Client closed connection)
# [17:51] * Quits: krit (~krit@public.cloak) ("Leaving.")
# [17:57] * Joins: lmclister (~lmclister@public.cloak)
# [17:59] * leaverou is now known as leaverou_away
# [18:05] * Quits: rhauck (~Adium@public.cloak) (Client closed connection)
# [18:06] * Joins: cabanier (~cabanier@public.cloak)
# [18:08] * Joins: tobie (tobie@public.cloak)
# [18:20] <TabAtkins> plinss: Ah, the draft/official split didn't occur to me. That's definitely what it is. I can deal with that.
# [18:20] <TabAtkins> plinss: Hmm, I guess there's a reasonable argument for valuedefs being auto-exported like propdefs.
# [18:21] <TabAtkins> It's unlikely they'll be defined for internal use, like many normal dfns are.
# [18:25] <plinss> TabAtkins: note that 'draft' anchors only appear in drafts, 'official' anchors appear in /TR and may or may not have a match in the draft
# [18:26] <TabAtkins> Right.
# [18:26] <plinss> I can track it the official anchors appear in drafts too if you need that...
# [18:26] <plinss> but I don't currently
# [18:26] <TabAtkins> I'm wondering what policy to use. I probably want to default to choosing the 'draft' anchor, but what about for when preparing for publication?
# [18:26] <TabAtkins> Or for doing CR/etc level drafts?
# [18:27] <plinss> I think if you need to know if you're processing a draft or /TR publication candidate
# [18:27] <plinss> and for /tr pub candidates always link to /tr if possible
# [18:28] <TabAtkins> I can tell that - the draft status is one of the required metadata fields, and anything other than ED or DREAM is a /TR draft.
# [18:28] * leaverou_away is now known as leaverou
# [18:28] * Joins: krit (~krit@public.cloak)
# [18:28] <TabAtkins> And yeah, I think you're right. That means possibly throwing fatal linking errors when you prepare a document for /TR, unfortunately.
# [18:28] <plinss> the issue if you default to linking to drafts is that you can't tell if an 'official' anchor appears in the draft (sometimes they don't)
# [18:29] <TabAtkins> I'm not sure what you mean.
# [18:29] <plinss> when preparing for /tr you can either ignore the draft anchors or actually link them to the ED
# [18:30] <plinss> when linking to the drafts by default, if you have an 'official' anchor, it might not be present in the current draft
# [18:30] <TabAtkins> Oh, I see what you mean.
# [18:30] <plinss> sometimes anchor names change, sometimes they go away, sometimes they change type (turn into sections, etc)
# [18:30] * Joins: rhauck (~Adium@public.cloak)
# [18:31] <TabAtkins> Hm, perhaps then I should actually *enforce* linking to ed when preparing an ed.
# [18:31] <SimonSapin> TabAtkins: your processor has a DREAM status?
# [18:31] <TabAtkins> (Unless you manually specify otherwise - I plan for that to be controllable.
# [18:31] <TabAtkins> SimonSapin: Yes, I use it for my github specs. ^_^
# [18:31] <SimonSapin> hehe
# [18:31] <plinss> as I said above, I _can_ track if official drafts also appear in the draft, but I don't currently
# [18:31] <plinss> s/official drafts/official anchors/
# [18:32] <TabAtkins> plinss: Presumably I can track that myself, just be seeing if the same linking text points to different versions of the same spec.
# [18:33] <plinss> that'll tell you if the anchor name changed, but not if the anchor went away in the draft entirely
# [18:34] <TabAtkins> Wouldn't it, though? If the anchor went away entirely in the ED, it just won't be present in the data.
# [18:34] <SimonSapin> plinss: shouldn’t we avoid that?
# [18:34] <TabAtkins> Rather, it won't be present under the "draft" status.
# [18:35] <plinss> SimonSapin: it can happen, say if an entire section is removed from one level and moved to another… (or just removed)
# [18:36] <plinss> TabAtkins: my point is that if the anchor appears in both the draft and the /tr version identically, you won't see a draft status version of it currently
# [18:36] <TabAtkins> Oh! Yeah, that obviously changes things. ^_^
# [18:36] <plinss> so you can't tell if it's the same in boh places or just not preent in the draft
# [18:36] <plinss> s/boh/both/
# [18:36] <plinss> s/preent/present/
# [18:37] <TabAtkins> Okay, so having that knowledge would be good. Either with a separate entry or with some addition to status, whatever.
# [18:38] <plinss> I can deal with that two ways, either add a new publishing status (something like 'both'), or add a separate entry for the draft anchor even if present in the /tr. But if I do the latter that may have interesting implications to other systems that I have to think about...
# [18:38] <TabAtkins> Whatever's easier; I can handle whatever.
# [18:39] <plinss> ok, the 'both' status is way easier I think
# [18:44] <plinss> so in the API I currently add a 'draft' boolean field if it's in draft state only, I'll add an 'official' boolean field too
# [18:46] <TabAtkins> Okay, so now it's just two booleans which are true if it's in a given status.
# [18:47] <plinss> right
# [18:47] <plinss> online shortly
# [18:47] <TabAtkins> Some of these boolean fields - are they only present when true? In Python I have to worry about key errors, so I'm wondering if I can just get away with "foo in ref" or if I have to do "foo in ref and ref[foo] == true"
# [18:47] <plinss> these two will only be present if true, yes
# [18:48] <TabAtkins> Same with export?
# [18:48] <plinss> (keeps the api payload smaller)
# [18:48] <plinss> yes, same with export
# [18:48] <TabAtkins> kk
# [18:49] <plinss> linking_text is also only present if it has a value, but for that one you obviously care about the value
# [18:49] <TabAtkins> Yeah, I just have to look at both that and title.
# [18:49] <plinss> all the others ar always present
# [18:50] <TabAtkins> k
# [19:01] <SimonSapin> TabAtkins: re Python key errors some_dict.get(key, default) is your friend
# [19:01] <SimonSapin> default defaults to None
# [19:01] <TabAtkins> Ah, that's right!
# [19:02] <TabAtkins> I find Python dicts slightly confusing. I prefer JS's Map semantics.
# [19:02] <SimonSapin> confusing meaning different to what you’re used to? :)
# [19:03] <TabAtkins> Kinda? I'm always trying to use .foo to pull things off of them.
# [19:03] <SimonSapin> http://docs.python.org/2/library/stdtypes.html#dict.setdefault is useful also at times
# [19:03] <TabAtkins> Yeah, I've used that one plenty.
# [19:03] <TabAtkins> It's ridiculously useful.
# [19:03] <TabAtkins> Though I hate the name, because it's really a get operation that might have a set side effect.
# [19:05] * Joins: dbaron (~dbaron@public.cloak)
# [19:25] <plinss> TabAtkins: ok, changes are online, specs are re-parsing as I type
# [19:25] <TabAtkins> plinss: Cool, thanks!
# [19:25] <plinss> np
# [19:31] * Quits: tobie (tobie@public.cloak)
# [19:32] <plinss> ok, parsing complete...
# [19:36] * Quits: krit (~krit@public.cloak) ("Leaving.")
# [19:42] <plinss> TabAtkins: I have a Q for you...
# [19:42] <TabAtkins> yo
# [19:43] <plinss> when parsing the specs, I'd like to be able to associate the value dfns with the properties they're valid for
# [19:43] <TabAtkins> I would also appreciate this!
# [19:43] <plinss> any thoughts how to accomplish that?
# [19:43] <plinss> often there's just prose
# [19:43] <TabAtkins> Since you know the section info, I'm betting you could auto-associate the valuedfn with the property in the same section, assuming there is a single propdef or descdec in that section.
# [19:44] <TabAtkins> Otherwise, the spec'll need an explicit annotation.
# [19:44] <plinss> right, it's the non-single property cases I'm concerned about, e.g.: http://drafts.csswg.org/css-align-3/#center
# [19:44] <TabAtkins> I'm fine with, say, having a <dfn data-for-prop=foo>''value''</dfn>.
# [19:45] <TabAtkins> Yeah, I just wouldn't auto-associate those.
# [19:46] <plinss> data-for-prop sounds like a win, but we also need to handle cases like in align where the values are for a type of value (like <item-position> )
# [19:47] <plinss> (and then the propdef needs to list the value types in a way I can parse...)
# [19:47] <TabAtkins> Oh, hm. That would be a much more involved proposition.
# [19:47] <plinss> my goal is to be able to generate something like: http://www.w3.org/TR/CSS21/propidx.html except for all properties and all values across all specs…
# [19:48] <TabAtkins> Yeah, definitely.
# [19:48] <TabAtkins> Hmm!
# [19:48] <plinss> with each value linking to the spec where it's defined...
# [19:48] <TabAtkins> So, a propdef should be easily parseable to get the <foo> types out.
# [19:49] <TabAtkins> And we could have a different property for a <dfn> to associate it with a specific type rather than a specific property.
# [19:49] <TabAtkins> data-for-type, I guess.
# [19:49] <TabAtkins> And then you just build a tree.
# [19:49] <plinss> I think so, yeah, except possibly for cases where a new level adds a <foo> type to an old property...
# [19:49] <TabAtkins> Ah yeah, the "additional values" thing.
# [19:49] <TabAtkins> That doesn't happen too often, so we can make sure to standardize it so it's machine-parseable.
# [19:50] <plinss> yeah, maybe a value-type dfn type or something...
# [19:51] <TabAtkins> Oh, I already recognize that in my processor anyway.
# [19:51] <TabAtkins> If a <dfn> contains something like "<foo>", I assume it's a type.
# [19:51] <plinss> ok that can work
# [19:51] <TabAtkins> I currently just use that to give it a better id.
# [19:52] <plinss> I can even add that as an anchor type, what should we call it?
# [19:52] <TabAtkins> typedef
# [19:52] * Joins: rhauck1 (~Adium@public.cloak)
# [19:52] <plinss> works for me
# [19:52] <plinss> so all we need to add to the spec markup is data-for-prop and data-for-type ?
# [19:53] * Quits: teoli (~teoli@public.cloak) (Client closed connection)
# [19:53] <TabAtkins> I think so. With gradual upgrades, we can build a type tree with roots in the properties.
# [19:53] <plinss> cool. I'll add support to the spec parser today.
# [19:53] <plinss> thx!
# [19:53] * Joins: teoli (~teoli@public.cloak)
# [19:54] <TabAtkins> ^_^ I really enjoy this collaboration. I'll mean lots of good things both for us spec authors and for web authors, as we get to publish better documentation.
# [19:54] <TabAtkins> (And better for implementors, too, so they're not constantly being surprised by new definitions.)
# [19:55] <plinss> yeah, I'm liking it too :-)
# [19:55] <plinss> I should also be able to have the spec parser flag a value def that it can't associate to a prop or type so we can find the specs that need new markup fast...
# [19:56] <TabAtkins> Excellent.
# [19:56] <TabAtkins> That'll be a big list, but easy to burn down.
# [19:56] <plinss> yeah...
# [19:57] <plinss> maybe the pre-processor can flag it too?
# [19:58] * Quits: rhauck (~Adium@public.cloak) (Ping timeout: 180 seconds)
# [20:00] * Quits: teoli (~teoli@public.cloak) (Ping timeout: 180 seconds)
# [20:01] <TabAtkins> Sure.
# [20:02] <TabAtkins> Well, I can't as easily find things that can be auto-associated, since I don't run the outline algorithm.
# [20:02] <TabAtkins> (I cheat instead.)
# [20:03] <TabAtkins> Hm, but it wouldnt' be hard to fake it in a way that would still be correct.
# [20:12] * Quits: plh (plehegar@public.cloak) ("Leaving")
# [20:14] * Quits: rhauck1 (~Adium@public.cloak) ("Leaving.")
# [20:20] * Joins: rhauck (~Adium@public.cloak)
# [20:24] * fantasai wants this nonsense to just go away
# [20:29] * leaverou is now known as leaverou_away
# [20:30] * leaverou_away is now known as leaverou
# [20:38] * Quits: dbaron (~dbaron@public.cloak) (Ping timeout: 180 seconds)
# [20:52] * Joins: plh (plehegar@public.cloak)
# [21:07] <TabAtkins> fantasai: What nonsense?
# [21:07] * Joins: krit (~krit@public.cloak)
# [21:10] <fantasai> ED vs. TR
# [21:47] * Joins: plh3 (plehegar@public.cloak)
# [21:53] * Quits: plh (plehegar@public.cloak) (Ping timeout: 180 seconds)
# [21:55] <fantasai> in my ideal world, ED is literally just scratch space
# [21:55] <fantasai> that we can share and comment on
# [21:55] <fantasai> and then publish when it's no longer half-finished edits etc.
# [21:57] <hober> in my ideal world, publication of an ED to /TR is just C-x C-s
# [21:59] <fantasai> I would not want to do that. I check in some not-so-solid things sometimes :)
# [22:00] <fantasai> and even beyond that, I Ctrl+S *all the time* while I'm editing to check it
# [22:00] <hober> yeah, and i don't mind the idea of all of those intermediate saves being visible on /TR
# [22:00] <fantasai> and I really don't want that showing up on the server until I'm done and have a coherent spec again
# [22:00] <hober> (in an ED)
# [22:01] <fantasai> Well, here's the thing, I don't think the live spec on /TR shoudl be called an ED. It shoudl just be called "Working Draft"
# [22:01] <fantasai> and it should just be kept up-to-date
# [22:01] <hober> ahh. yeah, WDs are different
# [22:01] <fantasai> ED should only exist if the editor wants a separate scratch space
# [22:02] <fantasai> and then, yeah, check in whatever whenever you want, nobody should be looking at it unless collaborating with you
# [22:19] * Quits: cabanier (~cabanier@public.cloak) ("Leaving.")
# [22:23] <TabAtkins> hober: I definitely mind intermediate saves being visible on /TR.
# [22:23] <TabAtkins> hober: Plus, I'd like to have a safer scratch-space for things like my github specs, without implicitly gaining W3C endorsement.
# [22:24] <liam> I wonder if a compromise would be a possible link from the TR page to an ED page for each WG
# [22:33] <fantasai> That's ridiculous
# [22:33] <fantasai> If people are expected to look at a draft, that should be the official draft
# [22:34] <fantasai> If everyone official tells you "don't look at the official draft, look at this scratch space over here" then maybe that means the thing you're calling the "official" draft *doesn't deserve that designation*
# [22:35] <liam> the difficulty is when people see an interim draft and go ff and implement it, or complain when "it doesn't work"
# [22:36] <liam> it's already a problem that Member submissions look like Recommendations to most people, I've seen books about the "W3C so-and-so spec" when it was just something submitted to us and published before there was even a working group
# [22:37] <TabAtkins> That problem doesn't block the problem fantasai's worried about.
# [22:37] <TabAtkins> Which is that TR is short for TRash. ^_^
# [22:38] <fantasai> plinss: Tantek doesn't seem to have edit permissions on the wiki, and I can't seem to figure out why
# [22:38] <fantasai> TabAtkins: btw, can you edit the wiki?
# [22:38] <fantasai> TabAtkins: e.g. the main page
# [22:38] <TabAtkins> lessee...
# [22:39] <TabAtkins> i can definitely get to the edit screen and then hit save without making changes.
# [22:39] * Joins: dbaron (~dbaron@public.cloak)
# [22:41] <fantasai> ok, then I have no idea what's going on
# [22:44] * Joins: tobie (tobie@public.cloak)
# [22:49] * Joins: florian (~florian@public.cloak)
# [22:54] <fantasai> plinss: So, i fixed Tantek's problem by doing something that shouldn't make a difference
# [22:54] <fantasai> plinss: But it seems like anyone in the public + css-wg groups can't edit the wiki
# [22:54] <fantasai> plinss: but if you're in public + css-wg + something-else-even-with-no-extra-permissions, it works
# [22:54] <fantasai> plinss: using 'foo' as a test user
# [22:55] <fantasai> plinss: feel free to play with it... but I have no idea what to do about that :(
# [23:02] <plinss> fantasai: just got back from lunch, looking in to it
# [23:07] <plinss> wow, that's really weird...
# [23:10] * Quits: dbaron (~dbaron@public.cloak) ("8403864 bytes have been tenured, next gc will be global.")
# [23:17] <plinss> fantasai: ok, fixed
# [23:17] <plinss> I was returning the list of a user's group in an (php associative) array keyed by group id, apparently it doesn't like that
# [23:18] <plinss> removed the keys and it's fine now
# [23:18] <plinss> must be something weird about the way they iterate the array not liking certain combinations of indexes
# [23:20] * Joins: cabanier (~cabanier@public.cloak)
# [23:21] * Quits: sgalineau (~sgalineau@public.cloak) ("Textual IRC Client: www.textualapp.com")
# [23:23] * Quits: tobie (tobie@public.cloak)
# [23:26] * Quits: florian (~florian@public.cloak) ("Leaving.")
# [23:30] * Quits: krit (~krit@public.cloak) ("Leaving.")
# [23:34] * Quits: cabanier (~cabanier@public.cloak) ("Leaving.")
# [23:37] <fantasai> plinss: ok
# [23:37] <fantasai> plinss: BTW, happy Sysadmin Appreciation Day! :)
# [23:38] * fantasai is glad you're around to fix these problems :D
# [23:39] <plinss> Heh, thanks.
# [23:39] <plinss> This one was a problem that I introduced unfortunately...
# [23:39] * Joins: dbaron (~dbaron@public.cloak)
# [23:47] * Quits: plh3 (plehegar@public.cloak) ("Leaving")
# [23:53] * Joins: tobie (tobie@public.cloak)
# Session Close: Sat Jul 27 00:00:00 2013

The end :)