[07:21:33] * aloisius is back (gone 62:00:44) [12:37:13] --> jack-e (nobody@pD9EBCFE1.dip.t-dialin.net) has joined #peak [14:15:56] <-- jack-e has quit ("2.8.2-EPIC3.004 -- just do it.") [15:53:03] --> pje (~pje@n00.bcrtfl01.us.wh.nameservers.net) has joined #peak [15:53:10] Howdy. [16:11:50] Boy it's quiet in here. [16:25:31] --> Maniac (~daManiac@209-5-247-105.mb.skyweb.ca) has joined #peak [16:25:42] hey pje is here [16:26:42] Yep. [16:26:47] hi [16:27:22] i'm a new peak wannabe posting as 'darryl' on the mailing list [16:27:38] * pje nods [16:28:13] So who's the *other* Maniac? :) [16:29:06] my 'home' irc client [16:29:23] pretty much permanatly partaking of 1/2 dozen or so IRC channels [16:29:45] Ah. [16:30:08] pje: so is PEAK something you do for projects @work [16:30:10] ? [16:30:28] Yep. [16:30:43] so it is currently used in production in that context? [16:30:58] Yep. [16:31:11] Only some very trivial parts of it, though. [16:31:14] * Maniac notes pje loves short answers :) [16:31:35] Ha. [16:31:42] you would consider it very early in development and too fluid for your consulting practice at thist stage? [16:32:10] * Maniac notes you are not accepting any new contracts business must be good [16:32:59] PEAK doesn't have stable web application support yet. [16:33:10] So it's hard to write web apps with it. [16:36:21] and your consulting is web-app-centric? [16:36:40] You could say that. [16:38:45] :) [16:41:19] are u the only peak developer? [16:41:41] No. [16:44:22] Oh. [16:44:27] In the words of Yoda, "There is one other." [16:44:45] * Maniac notes many cvs commit logs indicate pje [16:45:22] Yes, and a few mention someone else. [16:45:58] The domain name of the CVS repository might give you a clue as to who that someone would be. :) [16:47:56] eby? [16:48:28] That's me. [16:48:41] sarna? [16:48:52] Yep. [17:08:25] Hi [17:09:10] Does ZConfig allow multiline values? Is there anybody here with ZConfig experiences? [17:09:50] I don't think it supports multiline values, but I could be wrong. [17:10:43] I have read refernce and howto and didn't find it. [17:13:57] I would also appreciate richer key alphabet than [-._a-z0-9] [17:14:27] I imagine the authors would be open to patches. :) [17:14:59] Depending on your intent, an XML-based configuration language might be better. [17:16:51] Yes, it would be better. [17:20:48] I would configure fact that "service '/contact' offers itself to '/accounting' as 'lib.accounting:IContact'" and it is naturaly triplet - not only key-value. [17:21:50] I can group these triplets into sections but it will end with xmlish ZConfig. [17:25:54] That sounds like programmer-ish configuration anyway. [17:26:38] That is, not something an end-user is going to be able to understand or care about. [17:27:30] Pretty much I've noticed that when I try to do something in ZConfig, and it's difficult, it's a sign I'm trying to make the end-user do something that they shouldn't be doing. :) [17:27:54] But I haven't used it so much that I can say that it would always be the case. [17:32:36] And I have another sign. Pages of ZConfig manual a totaly glued by some chewing gum :-) [17:41:59] * pje sighs [17:42:18] I'm still trying to sort out the development plan for peak.web... [17:42:37] pje: I have studied your reply on my friday mail. Your solution gives me greater flexibility. Now I can have context bound database connection (two application works on one database and share cache), context bound to service or context bound to application itself. [17:42:46] But having a lot of difficulty getting natural milestones in place. [17:43:04] Greater flexibility than what? [17:43:23] pje: Than my proposed solution. [17:43:39] Ah. I'm not sure I ever understood what that was. :) [17:43:54] I was very confused when we talked. [17:45:00] But as long as you understand then I guess I don't have to now. :) [17:45:24] pje: Yes it was valuable for me. Thanks. [17:48:17] I have read your peak.web plan. I have missed some form abstraction. I have a plan to make some sort of server side xforms as defined on w3c. [17:51:38] Forms can be implemented as views to be placed in templates. [17:52:10] The main trouble I've run into at this point is that although I've designed a framework that can support very sophisticated things... [17:52:35] In itself, it doesn't comprise much... so little, in fact, that it's difficult to visualize the "kernel" [17:52:59] It feels like if I write the kernel it will be just a few scraps of code that don't do anything... [17:53:24] But on the other hand, if I try to implement all the layers of advanced functioning I envision, that seems too complex, too YAGNI. [17:54:10] I'm beginning to think I need a project to drive requirements, but it has to be a really *simple* project. [17:54:20] And I don't really have any of those. :) [17:54:28] I have learn from PEAK that interfaces are very important itself without code. [17:54:55] Me too. :) [17:56:51] I need to go AFK for a bit... [17:58:26] From my point of view, web development is form handling, table listing, tree traversing and some views. [18:02:07] I have a remark to HTML Tidy: it refuses clean lot of word files exported to HTML. [18:06:03] I'm back now. [18:06:48] Adding to your list: what about navigation? Menus? Breadcrumbs? [18:07:04] i18n? [18:19:40] YAGNI? [18:19:54] "You Ain't Gonna Need It" [18:20:25] ah [18:20:26] Frameworks designed before they're understood usually have to be thrown away. :( [18:20:26] http://xp.c2.com/YouArentGonnaNeedIt.html [18:20:26] Menus and navigation are so application specific that a do not see any chance to make any general interface. Another case is navigation inside isolated components as forms and such. [18:20:53] i18n is important, of course. [18:21:05] But menus for example have item titles and URLs, and that's pretty universal. [18:21:26] Breadcrumbs also require a title and URL. [18:21:37] And for i18n, presumably the titles must be translatable. [18:29:57] Menu and navigation appearence is often driven only by page template provided by designer. [18:30:45] i18n can be incorporated by substitution in template and URL is given. [18:31:06] Appearance *should* be driven by page tempaltes. [18:31:17] But that's "view", not "model" [18:31:47] And the model would need to provide what the view needs for rendering, no? [18:34:02] mm, using peak.storage and a pgsql connection can i connect with domain sockets or only tcp/ip? [18:34:19] Beats me. [18:34:23] yes, but I cannot see any space for modeling title and url tuple. [18:35:20] Even for dynamic resources? That's really what I'm talking about. [18:35:36] Maniac: I don' [18:35:49] t know much about PostgreSQL on a practical level yet. [18:36:31] But if you need it, it should be pretty easy for you to test it, and let me know what needs changing/adding in PEAK. [18:36:58] Maniac: I have modified connection URL parsing so it accepts pgsql:///... but I canot find it now. [18:37:48] aloisius: ah [18:40:07] Maniac: most probably it has disapeared with cvs update but I will restore it in my head. [18:43:36] I think that pattern in storage.SQL:GenericSQL_URL must be changed such way that it accepts empty server part. [18:44:54] Then server part is None and pgsql connector interpretes it as connection via UNIX domain socket. [18:46:00] When you get it working, feel free to submit a patch. :) [18:46:18] I was not sure if this solution has not some drawbacks and delayed patch and lost it :-( [18:49:06] pje: May I change GenericSQL_URL or it would be better make PG specific URL class? I think that mysql or other SQL server would operate same way. [18:49:45] domain sockets perform better and there is security settings you can use to restrict to domain based sockts [18:49:58] A specific syntax would certainly be appropriate for something that doesn't fit the "generic" mold. [18:50:42] GenericSQL_URL was just a quick shortcut; I imagine that over time every DB will grow its own variant, eventually. [18:56:33] It doesn't work for me now. Maybe something has changed in PEAK since my first attempt. I am getting infinite recursion. [18:56:50] Odd. [18:57:14] Especially since I just rewrote parts of the parsing framework this weekend to make recursive grammars work. [18:58:18] I get it when referencing attribute con = binding.bindTo("pgsql://localhost/test") [18:59:07] I did cvs update before while. [18:59:18] With unpatched PEAK? Or was that patched? [18:59:48] with upatched [19:00:15] Weird. Got a traceback sample? [19:03:39] Are you able to receive it via IRC as file? [19:03:42] --> _Maniac (~User@h24-77-230-121.wp.shawcable.net) has joined #peak [19:04:01] Maybe... give it a try. [19:04:16] where would i change from pgdb to a different DB-API driver? [19:04:20] But hurry; I have to go to a meeting. [19:04:42] Maniac: see peak.storage.SQL and 'peak.ini' [19:04:54] --- Offering url_tb.txt to pje [19:05:02] gracious [19:05:47] --- Offering x.py to pje [19:06:13] --- DCC SEND connection established to pje [208.55.254.110:33884] [19:06:16] --- DCC SEND url_tb.txt to pje complete [40980 cps]. [19:06:44] <-- _Maniac_ has quit (asimov.freenode.net irc.freenode.net) [19:08:05] Radek... add ',adaptTo=ISQLConnection' to your bindTo. [19:08:48] --- DCC SEND x.py to pje timed out - aborting. [19:09:10] the same [19:09:58] Okay... I think i found the problem... [19:10:18] In peak.storage.connections: [19:10:19] address = binding.requireBinding( [19:10:19] "Address used to create the actual connection", [19:10:19] suggestParent=False [19:10:19] ) [19:10:27] Add the , suggestParent=False [19:10:39] --- Offering x.py to pje [19:10:48] I've got to go... drop a note to the list about whether it worked or not. [19:11:13] Thanks. [19:11:20] <-- pje has quit () [19:13:40] --- DCC SEND x.py to pje timed out - aborting. [19:45:30] --> azaad (~jxj@sv-fw.looksmart.com) has joined #peak [19:45:39] Hi all [19:51:39] Hi azaad! I am just leaving. [19:52:04] Hi and bye then. :) [19:52:45] _Maniac: I have posted promised patch to transwarp list. [19:52:48] * aloisius is away: mangiare le scarpe [20:32:16] aloisius: cool [21:43:58] <-- Maniac (~daManiac@209-5-247-105.mb.skyweb.ca) has left #peak ("home time") [23:28:29] --- _Maniac is now known as Maniac [23:30:04] --- Disconnected (Remote host closed socket). [23:30:40] --> You are now talking on #peak [23:30:40] --- Topic for #peak is PEAK development (http://peak.telecommunity.com) [23:30:40] --- Topic for #peak set by jack-e at Fri Jun 20 11:58:28 [23:30:54] --> jolby_1 (~jolby@12-230-14-63.client.attbi.com) has joined #peak [23:43:42] -lilo- [Global Notice] Hi all. We're experiencing severe attacks and we'll be shutting down the network piece by piece. Thanks. [23:46:07] <-- azaad has quit (calvino.freenode.net irc.freenode.net) [23:49:56] -lilo- [Global Notice] Looks as if that may have been a bit premature, folks, at least for now things seem to be a bit better. We'll try bringing the two dropped main rotation servers back up.