Sorry, local time in this log (GMT +10) **** BEGIN LOGGING AT Tue Oct 4 16:36:34 2005 Oct 04 16:36:34 --> You are now talking on #for-oct Oct 04 16:36:34 --- zelazny.freenode.net sets mode +n #for-oct Oct 04 16:36:34 --- zelazny.freenode.net sets mode +s #for-oct Oct 04 16:36:34 --- #for-oct :[freenode-info] why register and identify? your IRC nick is how people know you. http://freenode.net/faq.shtml#nicksetup Oct 04 16:38:00 --- xley2 has changed the topic to: ForrestTuesday: cleanup Jira Oct 04 16:49:22 --> crossley (n=crossley) has joined #for-oct Oct 04 17:32:35 --> rgardler (n=chatzill) has joined #for-oct Oct 04 17:32:58 Morning/Evening David Oct 04 17:33:36 Evening/Morning Ross Oct 04 17:33:54 --> tscherler (n=thorsten) has joined #for-oct Oct 04 17:34:20 --- tscherler is now known as tscherler_just_l Oct 04 17:35:27 Hey Thorsten, good to see you, i was just asking a question on list Oct 04 17:36:09 yeah hi all Oct 04 17:36:30 I too much a junky to this damm proect ;-) Oct 04 17:36:39 :-) Oct 04 17:36:50 Lets try and make it a more welcoming place for you then Oct 04 17:37:02 (no need to coment Oct 04 17:37:10 just acknowledging how you feel right now) Oct 04 17:37:19 My question... Oct 04 17:37:24 I just need to learn when to shut up and think my part Oct 04 17:37:38 (that makes two of us) Oct 04 17:37:49 About views status... (re davids mail) Oct 04 17:37:51 that is hard to know in real life too, eh. Oct 04 17:38:08 Hi david, I thought you were asleep Oct 04 17:38:20 me alseep, no Oct 04 17:38:23 --- tscherler_just_l is now known as tscherler Oct 04 17:38:33 hi david Oct 04 17:38:33 only 4:30 pm here anyway. Oct 04 17:38:46 gidday thorsten glad you could be here Oct 04 17:39:28 hi david Oct 04 17:39:45 like written above I am an addict to FOR Oct 04 17:39:53 good Oct 04 17:40:19 Guys, I'm only here for another 20 minues Oct 04 17:40:30 Will be back later, around mid day BST Oct 04 17:40:34 * tscherler writting an answer to Miroslav Mocek Oct 04 17:40:58 My question for Thorsten (when he comes back) Oct 04 17:41:04 me here Oct 04 17:41:07 regarding status of views Oct 04 17:41:23 what is the plan, how do we help? Oct 04 17:41:38 jxpath in short Oct 04 17:41:45 was my summary on list correct Oct 04 17:41:46 but that is not easy Oct 04 17:42:16 david wrote Documentation will need to be changed. Oct 04 17:42:35 for the new plugins yes Oct 04 17:42:49 since you Oct 04 17:43:08 will need to request all data through the view Oct 04 17:43:19 there is no standard data model anymore Oct 04 17:43:46 rgardler I would love to see an example how you use contracts in an early state Oct 04 17:43:58 Yeah, I was thinking about that Oct 04 17:44:11 What we hae done with the Resume plugin Oct 04 17:44:21 is basicaly put an alternative path through the sitemap Oct 04 17:44:29 if you want to help give me an example and I can think about it Oct 04 17:44:32 This is no good for core as it Oct 04 17:44:43 will make maintenance a nightmare Oct 04 17:44:48 As I've been working through the LM stuff Oct 04 17:44:55 I've realised our sitemaps are crappy Oct 04 17:45:01 loads of repitition Oct 04 17:45:02 jeje Oct 04 17:45:09 The LM is helping alot Oct 04 17:45:10 then we are 2 Oct 04 17:45:14 but not enough Oct 04 17:45:24 I noticed that when starting with views Oct 04 17:45:29 I've been thinking of starting an Oct 04 17:45:37 experimental branch in which I will Oct 04 17:45:43 start building a new pipeline Oct 04 17:45:47 using your contracts Oct 04 17:45:59 ross Oct 04 17:46:04 In this I'll do the contracts first thing Oct 04 17:46:22 > >>0.9 - core as a plugin framework (all input/output code to plugins) Oct 04 17:46:22 the sitemaps have grown and grown. We already knew that they have become cumbersome. Oct 04 17:46:43 IMO we should not use branching Oct 04 17:47:01 better build everything in plugings Oct 04 17:47:16 > >>0.9 - core as a plugin framework (all input/output code to plugins) -> yes Oct 04 17:47:28 they can have dependency under each other Oct 04 17:47:28 I could use the XHTML2 plugin instead (remove structurer.xmap) Oct 04 17:47:53 My concern though... Oct 04 17:48:05 If I do this and your work on views diverges too far Oct 04 17:48:13 (or I diverge too far) Oct 04 17:48:19 one of us will be wasting effort Oct 04 17:48:22 thoughts? Oct 04 17:48:43 we need to bring them together Oct 04 17:49:14 still thinking how but they have to grow together Oct 04 17:49:45 the problem with the new version of views is that they are not doing the linkrewritting Oct 04 17:50:04 is that ite: and ext: only? Oct 04 17:50:12 yeah Oct 04 17:50:17 or does it include ;m:? Oct 04 17:50:24 lm you mean Oct 04 17:50:24 sortty site: and lm: Oct 04 17:50:35 lm should work Oct 04 17:50:41 I have not tried it Oct 04 17:50:54 but hinking about it Oct 04 17:51:10 have to try it Oct 04 17:51:12 so why aren't we focussing on the bug with Cocoon use of JXPath/Linkrewriter Oct 04 17:51:29 +1 for me Oct 04 17:51:35 crossley so why aren't we focussing on the bug with Cocoon use of JXPath/Linkrewriter Oct 04 17:51:55 let me be radical for a moment... Oct 04 17:52:01 the problem is quite heavy cocoon/jx stuff not trivial Oct 04 17:52:02 Do we need site: and ext:? Oct 04 17:52:14 lm: gives us the same but is not dependant on site.xml Oct 04 17:52:26 we no longer ave to restrict the elements in site.xml Oct 04 17:52:36 and we can thus create a DTD for site.xml Oct 04 17:52:57 really radical but worth a second thought Oct 04 17:53:30 Are you sure that LM gives you the same functionality as Linkrewriter Oct 04 17:53:39 Yes: see... Oct 04 17:54:09 http://forrest.apache.org/docs_0_80/locationmap.html#linkrewriting Oct 04 17:54:29 site and ext uses the structure of site.xml to build the name of the link Oct 04 17:54:38 i.e. site:path/to/node Oct 04 17:54:53 this means that you cannot change the structure of your navgation document Oct 04 17:54:59 lm uses a completely separate namespace Oct 04 17:55:13 thus we can change site.xml as much as we like Oct 04 17:55:23 We could even modifiy the lm so that it used ID's in site.xml Oct 04 17:56:17 this doesn'r fix the problem, but means we can proceed Oct 04 17:56:20 and we have two less protocols to deal with Oct 04 17:56:33 and we can finally create a dtd for site.xml Oct 04 17:57:09 I feel depressed thinking about all the documentation changes Oct 04 17:57:24 Yes, that is a big effort Oct 04 17:57:29 but then so is the XHTML2 move Oct 04 17:57:40 and views Oct 04 17:57:44 and sitemap refactoring Oct 04 17:57:56 in otherwords we have to rewrite every document already' Oct 04 17:58:07 I'm thinking of this as a 1.0 release Oct 04 17:58:42 so many changes at once doesn't help, just makes it worse Oct 04 17:59:07 It need not be at once Oct 04 17:59:14 *if* we do this Oct 04 17:59:25 it is to enable us to move on with views Oct 04 17:59:33 and to remove duplicated functionality Oct 04 17:59:40 it can happen in the 0.8 release Oct 04 17:59:50 others parts change in 0.9 and on Oct 04 18:00:12 Of course, fixing a bug is better for the community Oct 04 18:00:22 but do we need to fix this particular bug? Oct 04 18:01:04 Anyway, I'm +1 for fixing the bug or removing the functionlaity Oct 04 18:01:16 right now I've got to go, I'll be back later Oct 04 18:01:20 for dev of views no, for getting rid of skins yes Oct 04 18:01:37 There is surely far more to the site: and ext: protocols than just linkrewriting Oct 04 18:02:12 (please expand David, and post a current log before you go to bed, I have to take my machine with me) Oct 04 18:02:14 thanks Ross, we will regroup later Oct 04 18:03:58 thorsten: >>for dev of views no, for getting rid of skins yes .... pleas expand Oct 04 18:04:53 why is the JXPath problem related to skins Oct 04 18:05:16 Actually I did not even noticed that the linkrewritting was broken in run mode till you pointed it out Oct 04 18:05:42 because linkrewritting is big part of forrest Oct 04 18:05:56 <-- rgardler has quit ("Chatzilla 0.9.68.5 [Firefox 1.0.6/20050716]") Oct 04 18:06:09 yes, but not specific to "skins" Oct 04 18:06:20 if we want to officially state that views are replacing skins then everything that was working with skins should work with views Oct 04 18:06:28 that is right david Oct 04 18:07:17 i suppose so, we can't just dump existing users Oct 04 18:08:46 Anyway, i am going to try to spend some more time today trying to fix the JXPath issue in Cocoon Oct 04 18:09:19 :) Oct 04 18:09:23 but, i have already been trying for weeks. Oct 04 18:09:25 you are the best dude Oct 04 18:09:30 me too Oct 04 18:09:49 it is just too much touching jx specific code Oct 04 18:10:41 why so, it is just that the Linkrewriter (or Input Modules) are not dealing with the namespace Oct 04 18:11:05 the linkrewriter code actually has a fixme comment about that. Oct 04 18:11:32 namespace on the site.xml input file, i mean Oct 04 18:14:08 but actually site: , ...: is not a ns Oct 04 18:14:25 it is a trigger for the rewritting Oct 04 18:15:48 no that is not the problem Oct 04 18:16:23 ... from site.xml ... Oct 04 18:16:35 Oct 04 18:16:40 ... Oct 04 18:16:47 if that is instead ... Oct 04 18:17:53 Oct 04 18:18:04 then no problem Oct 04 18:18:09 oh Oct 04 18:18:11 interessting Oct 04 18:18:20 rt Oct 04 18:18:23 but doing that creates a different problem Oct 04 18:18:30 which? Oct 04 18:18:32 ... Oct 04 18:18:46 I wait with the rt remark Oct 04 18:19:08 okay i am getting a reference ... Oct 04 18:20:10 issues.apache.org/jira is so slow ... Oct 04 18:20:41 http://issues.apache.org/jira/browse/FOR-675 ... see the fourth comment Oct 04 18:21:23 moment Oct 04 18:21:25 "Removing the namespace declaration from our site.xml lets the "site:" etc. links get resolved properly. However doing that removes the left-hand menus too. Obviously not the solution, but shows that namespace is something to do with the problem." Oct 04 18:24:35 ah Oct 04 18:25:41 let me have a look into the code Oct 04 18:25:53 I may have an idea Oct 04 18:27:25 crossley the tabs are still working? Oct 04 18:27:51 i think so ... i will check. Oct 04 18:28:36 http://svn.apache.org/viewcvs.cgi/forrest/trunk/whiteboard/plugins/org.apache.forrest.plugin.internal.structurer/menu.xmap?view=markup Oct 04 18:29:42 here are the transforms that are responsible for rewriting of the menu Oct 04 18:29:57 e.g. Oct 04 18:31:16 verified, tabs still work but the left-hand menu disappears Oct 04 18:32:13 ... i am talking convention skins there, e.g. in site-author Oct 04 18:32:15 then IMO we can limit the debugging for now to above linked xmap Oct 04 18:32:26 ah Oct 04 18:32:44 then it is the menu.xmap in main Oct 04 18:34:31 checking behaviour in views ... Oct 04 18:35:01 crossley I will need to do some work (will still be in this chat but reaction time will be high) Oct 04 18:35:22 --- tscherler is now known as tscherler_afk Oct 04 18:35:23 ... verified, same there Oct 04 18:35:26 ok Oct 04 18:35:34 okay Oct 04 18:35:43 that makes sense because we are using the same stylesheets Oct 04 18:36:04 means if we fix it for views or skins we fixed it for both Oct 04 18:36:16 yep Oct 04 18:38:06 removing the namespace from site.xml is not really fixing the issue though. But it would enable us to continue work. Oct 04 18:39:09 I will have a look in the java classes after work Oct 04 18:40:04 great, if you mean Linkrewriter, then Cocoon trunk has a new sample to test it Oct 04 18:40:11 ok Oct 04 18:40:14 cheers Oct 04 18:40:27 happy working, see you later Oct 04 18:40:43 btw what happends if we do not the standard namepace Oct 04 18:41:37 same, there is no standard. Oct 04 18:42:18 but Oct 04 18:42:22 i gather the namespace is not being handled at all Oct 04 18:42:29 ok Oct 04 18:42:38 oh wait Oct 04 18:43:02 i misunderstood your question, will try ... Oct 04 18:43:07 cheers Oct 04 18:51:35 in forrest/site-author/ Oct 04 18:53:20 either qualifying the namespace as xmnls:site of removing the namespace attribute Oct 04 18:54:09 makes the "site:" etc links work properly. The side-effect is that the left-hand menu disappears. Oct 04 18:54:52 s/of removing/or removing/ Oct 04 18:55:02 david Oct 04 18:55:11 hmm Oct 04 18:55:16 did you add site: to all nodes? Oct 04 18:55:38 bleah, no Oct 04 18:55:42 ;-) Oct 04 18:55:58 try with a couple Oct 04 18:56:08 same thought Oct 04 18:56:34 jeje Oct 04 18:57:57 no change Oct 04 18:58:04 ok Oct 04 18:58:14 svn revert site.xml Oct 04 18:59:28 i have some other thing to investigate with that bug, so will go quiet on IRC Oct 04 18:59:55 ok Oct 04 19:00:07 and cook dinner/ watch Dr Who Oct 04 19:07:30 lol Oct 04 19:21:44 --> JennyCurran (n=eggdrop@80.224.64.103) has joined #for-oct Oct 04 19:23:34 <-- JennyCurran (n=eggdrop@80.224.64.103) has left #for-oct Oct 04 19:23:40 --> JennyCurran (n=eggdrop@80.224.64.103) has joined #for-oct Oct 04 19:26:45 --> cheche (n=cheche@80.224.64.103) has joined #for-oct Oct 04 19:26:53 good morning! Oct 04 19:43:41 hi cheche Oct 04 19:44:51 hey man.......... Oct 04 19:45:00 I camejust to setup JennyCurran Oct 04 19:45:08 :) Oct 04 19:45:11 how is everything? Oct 04 19:45:15 fine Oct 04 19:45:22 working Oct 04 19:45:35 y tu? Oct 04 19:45:40 bueno tambien! Oct 04 19:45:42 jajajajajajajjaaj Oct 04 19:45:45 look at it: Oct 04 19:45:47 http://casa.che-che.com/pootle/ Oct 04 19:45:51 is a language server Oct 04 19:46:12 the files are in po Oct 04 19:46:25 but is a web front end for people to translate Oct 04 19:46:32 kind of colaboration tool Oct 04 19:47:12 jeje nice logo ;-) Oct 04 19:47:25 but it is not mine... Oct 04 19:48:04 i know Oct 04 20:43:45 --> xley (n=crossley@apache/committer/xley) has joined #for-oct Oct 04 20:44:39 <-- crossley has quit () Oct 04 22:20:11 --> ipv6guru (n=gavinmcd) has joined #for-oct Oct 04 22:20:52 Hi Guys Oct 04 22:22:23 --- ipv6guru is now known as _Gav_ Oct 04 22:51:37 --> kronenpj (n=kronenpj) has joined #for-oct Oct 04 23:01:47 hi gav Oct 04 23:01:50 hi all Oct 04 23:08:08 <_Gav_> Aha, we are alive Oct 04 23:08:31 <_Gav_> Whats happening? Oct 04 23:08:41 I am working Oct 04 23:08:59 <_Gav_> I not long got back from work Oct 04 23:09:33 <_Gav_> I have had a couple of weeks break from forrest and am a bit lost at the moment, and after an svn up the other day Oct 04 23:09:43 <_Gav_> I can not even run a forrest test Oct 04 23:10:11 what is the prob? Oct 04 23:10:28 <_Gav_> I'd like to being again helping with views, maybe css or .ft or whatever you need, but need to have a working forrest Oct 04 23:10:46 <_Gav_> will paste the output here in a sec if thats ok Oct 04 23:11:09 use better pastbin.org Oct 04 23:12:08 pastebin.com I mean Oct 04 23:12:33 otherotherwise you may get an overflow and kicked from the channel Oct 04 23:13:01 <_Gav_> ok Oct 04 23:18:48 <_Gav_> http://pastebin.com/382568 Oct 04 23:19:37 <_Gav_> I seem to have broken something. It is ok to have latest svn isn't it , I don't need to rollback or anything ? Oct 04 23:20:23 file:/C :/Apache2/forrest/main/webapp/WEB-INF/cocoon.xconf:30:62 Oct 04 23:20:51 can you see whether you have a conflict there? Oct 04 23:21:02 <_Gav_> ok Oct 04 23:21:13 better paste the row 30 Oct 04 23:21:19 here ;-) Oct 04 23:24:17 <_Gav_> Line 30 says :- Oct 04 23:24:19 <_Gav_> Cocoon Version 2.2 ----> is this correct Oct 04 23:24:53 yes Oct 04 23:24:57 hmm Oct 04 23:25:06 what does a svn st say? Oct 04 23:27:16 <_Gav_> whilst svn st is running , I noticed I have a corrupt forrest-core.xconf Oct 04 23:27:59 <_Gav_> <<<<<<< .mine Oct 04 23:27:59 <_Gav_> @forrest.home@/build/plugins/org.apache.forrest.plugin.internal.xhtml2 Oct 04 23:27:59 <_Gav_> @forrest.home@/build/plugins/org.apache.forrest.plugin.internal.xhtml2/resources/views Oct 04 23:27:59 <_Gav_> ======= Oct 04 23:27:59 <_Gav_> @forrest.home@/build/plugins/@project.view.internal.plugin@ Oct 04 23:28:00 <_Gav_> @forrest.home@/build/plugins/@project.theme.plugin@/resources/views Oct 04 23:28:02 <_Gav_> >>>>>>> .r28921 Oct 04 23:28:16 that is the one Oct 04 23:28:17 ;-) Oct 04 23:29:03 <_Gav_> I will delete all forrest-core.xconf and diffs and re-do svn up , anything else ? Oct 04 23:30:48 make sure you do not have more C (svn conflicts) ;-) Oct 04 23:31:50 <_Gav_> I thought I read on-list somewhere about having to delete some files or such, do I still need to so this Oct 04 23:32:56 <_Gav_> now at revision 293592, will give this a try Oct 04 23:34:54 do you have the common-jxpath-1.2.jar in $forrest/lib/core ? Oct 04 23:40:44 <_Gav_> I have : commons-jxpath-20030909.jar Oct 04 23:41:01 <_Gav_> and : commons-jxpath-20030909.jar.license.txt Oct 04 23:41:11 <_Gav_> but not what you said. Oct 04 23:41:18 good as gold, mate Oct 04 23:43:47 Quick question - From where can a common-jxpath jar be downloaded? Oct 04 23:44:22 http://jakarta.apache.org/commons/jxpath/ Oct 04 23:44:44 Thanks... I'll go back to sleep now ;) Oct 04 23:44:57 good night, good fight ;-) Oct 04 23:45:23 Oh, it's 8:45am here. For some reason I hadn't been able to find it. Oct 04 23:45:52 Thanks! Oct 04 23:46:16 you're welcome Oct 04 23:46:35 --> rgardler (n=chatzill) has joined #for-oct Oct 04 23:46:58 I'm back if anyone is around... if not I'm working on the LM stuff Oct 04 23:51:00 i am going to have lunch cya Oct 04 23:54:15 <_Gav_> thanks for your help ts Oct 04 23:54:18 <_Gav_> Hi Ross Oct 05 00:02:14 HI gavin, what you guys up to? Oct 05 00:02:24 (what is the URL for JEnny logs?) Oct 05 00:02:52 <_Gav_> Unfortunatly I am having probs with my forrest, so am currently deleting the lot and starting again Oct 05 00:03:09 :-( that happens occasionally Oct 05 00:03:27 <_Gav_> dont know where the log is, I only got here an hour or so ago Oct 05 00:03:49 OK, I'll mail Cheche Oct 05 00:05:04 <_Gav_> I was hoping to get up to speed and help with something, CSS or .ft but will have to wait until I have svn up'ed the lot Oct 05 00:06:00 OK let me know when you svn'upped Oct 05 00:06:06 we can work together on something Oct 05 00:07:02 <_Gav_> ok Oct 05 00:17:25 <-- _Gav_ has quit (Read error: 104 (Connection reset by peer)) Oct 05 00:35:41 --> _Gav_ (n=gavinmcd) has joined #for-oct Oct 05 00:36:16 <_Gav_> damn windows, crashed on me, started again getting all of trunk Oct 05 00:45:57 rgardler: http://casa.che-che.com/~bot/forrest/forrest.log.04Oct2005 Oct 05 00:46:15 I just came back home... Oct 05 00:46:49 --> add1sun (n=add1sun) has joined #for-oct Oct 05 00:47:15 hi add1sun Oct 05 00:47:27 hey all, it's addi Oct 05 00:47:39 I can't stay on long Oct 05 00:47:51 but was wondering if there is a live log available? Oct 05 00:49:05 http://casa.che-che.com/~bot/forrest/forrest.log.04Oct2005 Oct 05 00:53:04 <_Gav_> Just ignore my problems, they can be cut-out Oct 05 00:53:08 <_Gav_> :) Oct 05 00:53:26 Thanks for log Cheche Oct 05 00:53:28 Hi Addi Oct 05 00:54:59 <-- add1sun has quit (Read error: 104 (Connection reset by peer)) Oct 05 00:55:06 --> add1sun (n=add1sun) has joined #for-oct Oct 05 00:55:33 whoops, didn't mean to disconnect Oct 05 00:55:38 thanks for the link Oct 05 00:55:51 i'll try to come back later Oct 05 00:56:05 See you later Addi Oct 05 00:56:17 Gavin, you got Forrest working yet? Oct 05 00:56:25 <-- add1sun (n=add1sun) has left #for-oct Oct 05 01:00:15 <_Gav_> would you believe svn is still getting it Oct 05 01:00:40 <_Gav_> Computer crashed during first attempt (see 21:36) Oct 05 01:01:24 <_Gav_> What are you working on anyway? Oct 05 01:02:43 <_Gav_> Should be 15:36 in log Oct 05 01:02:45 I'm still working on refactoring the sitemaps for LM Oct 05 01:03:26 <_Gav_> What do you have to do for that Oct 05 01:03:38 It's pretty simple really... Oct 05 01:03:42 do you know how the LM works? Oct 05 01:04:23 <_Gav_> Hmm, not really, having had a couple of weeks off, I have not really looked at LM, but I can when I am running again Oct 05 01:04:46 OK, here's the whistelstop tour - full details are at Oct 05 01:05:07 http://forrest.apache.org/docs_0_80/locationmap.html Oct 05 01:05:14 The locationmap allows us to Oct 05 01:05:27 separate the client URLspace from the source urlspace Oct 05 01:05:39 so a client can ask for domain.org/some/file.html Oct 05 01:06:00 but the source can come from http://svn.apache.org/an/unrelated/location (for example) Oct 05 01:06:13 This allows us to build the src structure to suit Oct 05 01:06:16 the editor Oct 05 01:06:23 and build the urlspcae to suit the user Oct 05 01:06:31 (with me so far?) Oct 05 01:07:10 <_Gav_> Yep, with you there. Same file, different location Oct 05 01:07:17 OK Oct 05 01:07:27 This magic is done through the lm: protocol in the sitemap and Oct 05 01:07:31 the locationmap.xml file Oct 05 01:07:36 You can see a location map at Oct 05 01:08:12 http://svn.apache.org/repos/asf/forrest/trunk/main/webapp/locationmap.xml Oct 05 01:08:25 In there there are no locators, only mounts, they mount files like this one: Oct 05 01:08:35 http://svn.apache.org/repos/asf/forrest/trunk/main/webapp/locationmap-transforms.xml Oct 05 01:08:49 The mounting is like an import Oct 05 01:08:52 (still with me?) Oct 05 01:09:30 <_Gav_> ok, yep, will understand more by doing, but this theory is good Oct 05 01:09:40 We are very nearly at the doing ;-) Oct 05 01:09:50 <_Gav_> cool Oct 05 01:09:57 If you take a look at the second link above, you will see Oct 05 01:10:09 Oct 05 01:10:13 right at the end of the file Oct 05 01:10:26 in there you will see a number of loctions like this: Oct 05 01:10:33 Oct 05 01:10:47 Notce the match is like those in the sitemap Oct 05 01:10:59 and we can use the elements matched in the same way with {1} and {2} Oct 05 01:11:10 OK, no to see it used... Oct 05 01:11:12 Take a look at Oct 05 01:12:07 http://svn.apache.org/viewcvs.cgi/forrest/trunk/main/webapp/sitemap.xmap?rev=293018&r1=293014&r2=293018&diff_format=h Oct 05 01:12:19 Notice how a big chunk of the sitemap is now replaced with Oct 05 01:12:30 Oct 05 01:12:44 (do you see how it works?) Oct 05 01:15:32 <_Gav_> Sort of, I am not familar with what {1} and {2} are though Oct 05 01:15:44 OK Oct 05 01:15:47 <_Gav_> in the coding sense Oct 05 01:16:00 it's the same in the locationmap as it is in the sitemap... Oct 05 01:16:16 say we have a pattern of "somefile.*" Oct 05 01:16:36 this will match a request for "somefile.html" and for "somefile.xml" Oct 05 01:16:44 In the first instance {1} = "html" Oct 05 01:16:51 int the second {1} = "xml" Oct 05 01:17:01 so... Oct 05 01:17:22 lm:transform.document.fo will be matched by transform.*.* Oct 05 01:17:35 {1} = document {2} = fo Oct 05 01:17:50 Does that help? Oct 05 01:18:57 <_Gav_> ok, {1}2{2} == document2fo Oct 05 01:19:50 You got it! Oct 05 01:19:52 <_Gav_> Thats a lot less coding at http://svn.apache.org/viewcvs.cgi/forrest/trunk/main/webapp/sitemap.xmap?rev=293018&r1=293014&r2=293018&diff_format=h Oct 05 01:20:01 <_Gav_> :) Oct 05 01:20:17 (hong on phone ringing) Oct 05 01:22:30 Yes, it has advantages for amount of code, but also Oct 05 01:22:40 in that we can now change the location without touching the Oct 05 01:22:41 sitemaps Oct 05 01:22:53 very useful for customising individuals sites Oct 05 01:23:16 What I am trying to do now is go through all the sitemaps and use the locationmap Oct 05 01:23:47 See http://issues.apache.org/jira/browse/FOR-200 Oct 05 01:24:20 <_Gav_> Ok, yep, so is the replace process identical for all sitemaps Oct 05 01:24:34 <_Gav_> Oct 05 01:24:35 <_Gav_> Oct 05 01:24:35 <_Gav_> Oct 05 01:24:40 <_Gav_> replace with Oct 05 01:24:46 <_Gav_> Oct 05 01:24:54 Yes, but it goes further Oct 05 01:24:57 <_Gav_> except for the names] Oct 05 01:25:07 <_Gav_> ok Oct 05 01:25:31 My aim is to replace all {forrest:*} and {project:*} variables Oct 05 01:25:42 this is to facilitate the issue at Oct 05 01:25:54 http://issues.apache.org/jira/browse/FOR-588 Oct 05 01:26:01 for example: Oct 05 01:26:18 {project:skinconf} becomes {lm:project.skinconf} Oct 05 01:26:47 There is a naming convention defined at http://forrest.apache.org/docs_0_80/locationmap.html Oct 05 01:26:51 but it is not complete yet Oct 05 01:27:25 btw the lm stuff and views together make forrest a renderer that could be used in other appz without major changes Oct 05 01:27:37 Hi Thorsten... Oct 05 01:27:51 Yes, I have not touched on the massive power the two together create Oct 05 01:28:03 <_Gav_> Hi T Oct 05 01:28:25 hi again Oct 05 01:28:30 back from lunch Oct 05 01:28:46 <_Gav_> I think a few brain cells are doing something Oct 05 01:29:44 We seem to have that effect on each other here :-) Oct 05 01:30:05 <_Gav_> yep Oct 05 01:30:08 So, Thorsten, I read the logs regarrding site: and ext: Oct 05 01:30:17 care to start where we left off? Oct 05 01:30:47 problem is that I am still at work :( Oct 05 01:31:09 can help with throwing in ideas but not with coding :( Oct 05 01:31:16 atm Oct 05 01:31:28 OK, no probs Oct 05 01:31:39 Can you talk for a few mns? Oct 05 01:31:47 sure Oct 05 01:31:51 OK Oct 05 01:32:13 I see both David and yourself were concerned that if we got rid of site: and ext: Oct 05 01:32:22 we would be in danger of not supporting current users Oct 05 01:32:27 this is, of course, valid Oct 05 01:32:36 I was thinking that if we did it Oct 05 01:32:59 we could provide some XSL and scripts to upgrade Oct 05 01:33:04 What do you think? Oct 05 01:33:24 so to get you right Oct 05 01:33:50 you would transform site:bla to lm:site:bla ? Oct 05 01:34:04 something like that? Oct 05 01:34:31 yes, something like that Oct 05 01:34:36 what we have not test it is whether lm is working with jxpath-1.2 Oct 05 01:35:07 very good point :-)) Oct 05 01:35:13 ;-) Oct 05 01:35:29 Although, my motivation for suggesting deprecation of site: and ext: Oct 05 01:35:37 is not only as a workaround for the JXPath issue Oct 05 01:35:48 I don't like duplication and lm: does both Oct 05 01:37:34 good point Oct 05 01:37:53 and it is easier extensible Oct 05 01:38:07 and it allows us to use a DTD for site.xml Oct 05 01:39:06 OK. I'll let you get some work done Oct 05 01:39:20 If you can test lm: with JXPath on your local installation that would be great Oct 05 01:39:29 I'll carry on with the LM refactoring for now Oct 05 01:40:14 ok Oct 05 01:40:27 Gavin, fancy giving me a hand with the LM stuff? Oct 05 01:46:13 <_Gav_> yep, svn still downloading, hopefully wont be much longer :| Oct 05 01:46:37 <_Gav_> Am taking a browse though FOR:200 now. Oct 05 01:50:37 <_Gav_> Those 2 patches of Tims seem to be putting in similar to what you are taking out ?# Oct 05 01:50:55 <_Gav_> Was these patches earlier thoughts Oct 05 01:51:44 YEs, the patches have been applied now Oct 05 01:51:50 We are working on the subtasks Oct 05 01:52:12 i.e. using the lm: in each of the sitemaps Oct 05 01:53:17 <_Gav_> woohoo Revsion 293601, svn complete. Oct 05 01:53:39 <_Gav_> How are you going about testing these LM changes? Oct 05 01:54:27 <_Gav_> what do you want me to have a go at? Oct 05 01:55:04 All testing is done by: Oct 05 01:55:10 <_Gav_> P.S - It's nearly 11:PM over here, so I'll give it an hour and then carry on in the morning Oct 05 01:55:17 cd FORREST_HOME/main Oct 05 01:55:32 ./build.[sh|bat] test Oct 05 01:55:42 This will do some tests but they are not complete Oct 05 01:55:58 An hour is good, we'll be well up to speed by then. Oct 05 01:57:06 I just did issues.xmap, how about you have a go at linkmap.xmap Oct 05 01:57:13 any problems, just holler Oct 05 01:57:53 <_Gav_> ok, will do Oct 05 01:58:24 NB - I have not looked at the linkmap.xmap - so I'm not sure if it is a simple one or not Oct 05 01:58:43 <_Gav_> cheers :) Oct 05 01:59:46 do SVN up now, I just committed a couple more changes, will save conflicts in your edits of locationmap.xml Oct 05 02:00:40 <_Gav_> ok, is it good to do a build clean, build, build test - after every svn up? Oct 05 02:01:06 Generally you only need do that if there has been a change to Java (which there hasn't this time) Oct 05 02:01:18 I only do it if java has changed or I am seeing strange effects Oct 05 02:02:25 <_Gav_> ok, as an aside, why do I still get output.pdf not found error Oct 05 02:02:55 hen running waht command? Oct 05 02:03:46 <_Gav_> build test, forrest site etc Oct 05 02:06:07 <_Gav_> configure cocoon: path/../org.apache.forrest.plugin.output.pdf\conf not found Oct 05 02:06:47 <_Gav_> I have had this error since I first encountered forrest, and a clean install today has not changed it Oct 05 02:07:06 <_Gav_> doesn't seem to effect anything I have been doing though. Oct 05 02:08:40 Hang on I'm trying to duplicate Oct 05 02:09:24 I'll be back to this issue, I just realised I broke SVN head with my last commit Oct 05 02:09:28 must fix that immediately Oct 05 02:12:24 <_Gav_> oops Oct 05 02:16:28 OK, fixed now (forgot to svn add one of the files_) Oct 05 02:16:42 This shows that in answer you your question: Oct 05 02:17:24 How do we test the LM stuff? Oct 05 02:17:33 I should have said... Oct 05 02:17:41 build test __AND__ Oct 05 02:17:48 forrest site in site-author Oct 05 02:18:00 back to you output.pdf problem Oct 05 02:18:07 let me try and reproduce Oct 05 02:23:05 damn, I have still broken trunk! Oct 05 02:28:23 --> diwaker (n=diwaker) has joined #for-oct Oct 05 02:30:16 --> ferdinandSoethe (n=ferdinan) has joined #for-oct Oct 05 02:30:38 <_Gav_> Hi Guys Oct 05 02:31:39 Hi everybody. Sorry to be so late, appointments all day ... Oct 05 02:32:18 <_Gav_> no probs, you just missed a confused .self Oct 05 02:32:49 what does that mean? Oct 05 02:33:26 <_Gav_> I was confused, but the cloud is slowly lifting Oct 05 02:34:04 <_Gav_> Ross: in linkmap.xmap, the first change I see is Ah, I see. Hope not one of the many clouds I created around some silly mistakes :-) Oct 05 02:34:10 btw: who is Paul Kronenwetter? Oct 05 02:35:10 Gav: that would be {lm:transform.linkmap.document} and would be matched by the matcher we were referring to earlier Oct 05 02:35:21 High Ferdinand - welcome Oct 05 02:35:34 Good question (re Paul) Oct 05 02:35:49 Welcome Paul, care to introduce yourself? Oct 05 02:36:26 <_Gav_> ok Oct 05 02:36:49 are you guys discussing cleaning up the sitemaps? should I, can I chip in a question re the html pipeline or should I wait till you get to forrest.xmap? Oct 05 02:37:14 Ferdinand, I think I just broke head... Oct 05 02:37:19 let me revert my changes and I'll come back to you... Oct 05 02:38:04 Sure, no problem. I'll take another look at my problem in html-transformations Oct 05 02:40:25 Ross: I finally read the docs on locationmaps. Really liked the docs and the mechanism. Reading it make me wonder why we should still have all those variables for directories in Forrest.properties. Oct 05 02:41:18 Shouldn't /couldn't we do away with a few layers of indirections and base all of it on locationmaps? Oct 05 02:42:14 Oh, Hi! I'm Paul... I'm lurking, for lack of a better word. Just trying to get a better grasp on views. Oct 05 02:43:04 Welcome! Oct 05 02:43:10 <_Gav_> Aren't we all :) Oct 05 02:43:20 Thanks! I hope to be able to help someday, but that is not today... Oct 05 02:43:49 Well you could help us break head, couldn't you :-) Oct 05 02:43:52 _Gav_: Perhaps, but some code... :> Oct 05 02:44:17 I suppose I could, but I'm not sure I have a working version of forrest to start with... Oct 05 02:47:00 hi everyone Oct 05 02:48:50 * diwaker is back. Oct 05 02:48:59 Hi Diwaker. How is San Diego? Are you planning on attending ApacheCon? Oct 05 02:49:10 san diego's warm and sunny as ever Oct 05 02:49:44 apachecon -- not sure. i'm going to be out of country at that time i think Oct 05 02:50:12 Hi diwaker - good to see you again Oct 05 02:50:35 All: head is a (little) broken and I have no idea why Oct 05 02:50:53 is there anyone here who has not updated from SVN today? Oct 05 02:50:58 i haven't Oct 05 02:51:05 Cool, can you do a test for me? Oct 05 02:51:08 sure Oct 05 02:51:09 shoot Oct 05 02:51:13 (saves me reverting changes) Oct 05 02:51:19 me too Oct 05 02:51:23 do a forrest run in site-author Oct 05 02:51:28 then try getting... Oct 05 02:51:35 <_Gav_> Ross : Entry in locationmap-transforms.xml has already a Oct 05 02:51:36 http://localhost:8888/docs_0_80/changes.rss Oct 05 02:52:38 Gav. that match will correctly match the one you are adding, i.e. Oct 05 02:52:39 says resource not found Oct 05 02:52:54 {lm:trasnform.linkmap.document} Oct 05 02:53:21 <_Gav_> cool, next. Oct 05 02:53:24 will be translated to {forrest:stylesheets}/linkmap2document.xsl by the line you quote Oct 05 02:53:41 diwaker - thanks it appears I didn't break it today then :-) Oct 05 02:53:45 what revision do you have? Oct 05 02:54:02 293498 Oct 05 02:54:13 OK, anyone got any earlier than that? Oct 05 02:54:25 <_Gav_> 294747 for me Oct 05 02:55:09 OK, I'm trying to find the problem Oct 05 02:55:18 Ferdinand, fire with your problem when you are ready Oct 05 02:56:18 Well my problem re sitemaps is really question: Why do we use IDgen twice in the html pipeline? Oct 05 02:56:46 <_Gav_> My time is up, will turn into a pumpkin anytime now, will continue with linkmap.xmap in the morning. Oct 05 02:57:25 hi all bye all Oct 05 02:57:31 I will go home Oct 05 02:57:32 Good nite Oct 05 02:57:40 OK Gav, I'll leave it for you then ;-) - thanks for your help Oct 05 02:57:40 maybe I have internet Oct 05 02:57:46 Bye Torsten Oct 05 02:57:51 See you later Thorsten (we hope) Oct 05 02:58:14 <-- tscherler_afk (n=thorsten) has left #for-oct Oct 05 02:58:25 <_Gav_> NO probs, will check my work with you then, I guess you'll be gone when I get up. Oct 05 02:58:46 Probably, 5 pm here now Oct 05 02:59:02 just attach a patch to jira, I'll look at it when I get up Oct 05 02:59:52 * diwaker is away: off to school Oct 05 03:00:51 Ferdinand - ID gen twice? Hmmm, let me look Oct 05 03:01:36 I only see it once (in forrest.xmap) should I be looking elsewhere? Oct 05 03:02:00 yes, in sitemap is a second one Oct 05 03:02:16 OK, hang on... Oct 05 03:02:38 Oct 05 03:03:12 And this one - as far as I can tell - is called for all documents Oct 05 03:03:41 Yes, it looks like an error to me Oct 05 03:03:52 have you tried removing the i18n one in forrest.xmap? Oct 05 03:04:01 no Oct 05 03:05:00 I think it can be safely removed, give it a try if it is causing you problems Oct 05 03:05:32 OK, so I'll take that out right now and commit strait away? Oct 05 03:05:50 After testing of course... Oct 05 03:06:21 The only problem I see is that if someone requests *.xml and it is generated from Oct 05 03:06:30 an *.html file it will not have ids Oct 05 03:06:39 perhaps we should do it the other way around Oct 05 03:06:44 take it out of body-* Oct 05 03:06:49 and add it to all *.xml matchers Oct 05 03:06:52 WDYT? Oct 05 03:07:28 hmmm, not very elegant. Why would sdy need ids in XML. This is typically a final processing step, not? Oct 05 03:07:46 Not necessarily... Oct 05 03:08:32 Someone may use the *.xml request to get some content for inclusion elswhere via xi:include Oct 05 03:09:01 (that ay be why it was added in the first place - looking to see who added it) Oct 05 03:10:16 ok, but why not generate the ids at the end of their post processing. After all, these IDs are really to fac. local navigation, they are no something I'd like to have in my Oct 05 03:10:40 original xml (just bloating tsorage) Oct 05 03:10:52 Yes, you are correct Oct 05 03:11:10 and besides, as diwaker pointed out, our id's are not even xml conforming Oct 05 03:11:14 I say remove it in the i18n matchers, we can always check with whoever did the original commit Oct 05 03:11:47 ok, will do that. meanwhile, here is sth I asked earlier while you were fixing things: Oct 05 03:11:49 Ross: I finally read the docs on locationmaps. Really liked the docs and the mechanism. Reading it make me wonder why we should still have all those variables for directories in Forrest.properties. Oct 05 03:11:50 Shouldn't /couldn't we do away with a few layers of indirections and base all of it on locationmaps? Oct 05 03:13:04 You are right Oct 05 03:13:08 it is a two stage process though Oct 05 03:13:30 Quick test everythings looks ok. I'll check in the change. Oct 05 03:13:35 See http://issues.apache.org/jira/browse/FOR-200 Oct 05 03:13:46 which blocks http://issues.apache.org/jira/browse/FOR-588 Oct 05 03:13:56 The second of these removes the indirection Oct 05 03:14:01 the first is what I am currently doing Oct 05 03:15:26 ok will check that out. is there time for a question re the html cleanup? Oct 05 03:15:48 yes - you are keeping me from finding the changes.rss bug and that is annoying e now Oct 05 03:16:28 it is what? Oct 05 03:16:57 sorry - it is annoying me now (i.e. I wnat to be distrcted from it) Oct 05 03:17:15 ok, no problem ... Oct 05 03:17:34 so here is the remaining problem (I posted the rest to the list) Oct 05 03:18:36 This template here is not doing what is should be Oct 05 03:18:38 Oct 05 03:18:40 Oct 05 03:18:42 Oct 05 03:18:43 Oct 05 03:18:45 Oct 05 03:18:46 Oct 05 03:18:48 Oct 05 03:18:49 Oct 05 03:18:51
Oct 05 03:18:52 Oct 05 03:18:54 Oct 05 03:18:55 Oct 05 03:18:57 Oct 05 03:18:58 Oct 05 03:19:00 Oct 05 03:19:01 Oct 05 03:19:03 Oct 05 03:19:04 Oct 05 03:19:06 Oct 05 03:19:08 Oct 05 03:19:10 Oct 05 03:19:11 Oct 05 03:19:13 Oct 05 03:19:15 Can you see why it would always go into the otherwise-branch even if I set the test to true? Oct 05 03:20:49 Did you see the commit by david: http://svn.apache.org/viewcvs.cgi/forrest/trunk/main/webapp/skins/common/xslt/html/document2html.xsl?rev=293531&view=log Oct 05 03:21:48 not yet. will check it out right away. Oct 05 03:22:10 (incidentally a tip from Thorsten: http://www.pastebin.com - very useful when using IRC) Oct 05 03:22:21 Take a look at that commit and see if it answers your question Oct 05 03:23:22 hmmm. Interesting change. will try that right away. Oct 05 03:25:17 Basically, if a node does not exist it will return false Oct 05 03:25:31 if a node does exist but is empty then you need the ='' test Oct 05 03:25:49 so not(@class} will return true if the node doesn't exist Oct 05 03:26:02 in programming terms null != '' Oct 05 03:27:42 works like a charm. I had assumed that non existant attributes can be tested like non existant elements. Now I know better. Thanks David. Oct 05 03:28:05 Stung me a while back too - wish David sat by me at all times ;-) Oct 05 03:29:17 Do you remember wanting a schema for site.xml and not being able to have one becuase of site: and ext: Oct 05 03:30:50 yes, and I still do ... Oct 05 03:32:16 OK, well I'm here to tell you that you can have it, but at a cost... Oct 05 03:32:18 that pastebin stuff is nice. wish I could have that in eclipse and offline Oct 05 03:32:41 ok, set the trap, I'm ready to jump in :-) Oct 05 03:32:46 :-) Oct 05 03:33:41 Can you read the logs, from when I first arrived this morning to when I leave (http://svn.apache.org/viewcvs.cgi/forrest/events/forrest-tuesdays/20051004-log.txt?rev=293577&view=markup) Oct 05 03:34:36 will do Oct 05 03:35:18 and also from around 16:30 to when Thorsten goes back to work (http://casa.che-che.com/~bot/forrest/forrest.log.04Oct2005 Oct 05 03:35:27 Then I'd like your comments on what we were discussing Oct 05 03:36:17 (I'll be gone for 10-15mins) Oct 05 03:40:30 Having a good laugh about David saying "I feel depressed thinking about all the documentation changes" Oct 05 03:43:34 Feedback on the morning part: I'm getting a rough idea of what you proposing ... and I like it so far. Not sure that I still like it when I understand the details but for the moment this is what I see: Oct 05 03:47:06 we kick out ext and site and start out with site.xml as being a very simple document with a strict grammar that defines a navigation structure with a hierarchie of menu items that link to pathnames. These pathnames can be real pathnames (simple case) or they can be highjacked to a different resource by using a location map. Oct 05 03:50:51 What I like is the 2 step approach with clear separation of complexity layers (what I mean is that novice users never have to deal with indirections until they need them). What I don't like so much is that once you start using location maps you always have to look at two documents to know what a menu item will actually do. In that sense I think that site.xml is much more elegant. BTW. During... Oct 05 03:50:53 ...the last discussion we hand in this sbdy (Nicola Ken?) suggested that we could rather easily switch from unique element names to id's in site because the original problem was not longer there. Oct 05 03:53:50 anybody still following or am I on my own with Jenny once more? Oct 05 03:54:17 Did you miss that I said I'd be away for a short while, just come back anyway - reading your comments Oct 05 03:55:13 No I didn't. I just thought I'd have time to flirt ... Oct 05 03:55:36 :-)) Oct 05 03:55:47 OK, to your comments.. Oct 05 03:55:56 You have the basic idea, yes Oct 05 03:56:22 regarding the problem of now knowing where to look for a file see... Oct 05 03:57:00 http://marc.theaimsgroup.com/?l=forrest-dev&m=112814359618995&w=2 Oct 05 03:58:22 With respect to id's, that could (theoretically be done, but by who?) however... Oct 05 03:58:51 I don't see this improving things much. As long as we hand edit site providing a report is just not enough. Oct 05 03:58:52 See my comments in the log at 1635 (http://casa.che-che.com/~bot/forrest/forrest.log.04Oct2005) Oct 05 03:59:57 (i.e. site: ext: and lm: all duplicate functionlaity) Oct 05 04:01:59 \stalk ferdinand Oct 05 04:02:09 (sorry wrong slash ) Oct 05 04:02:15 pardon? Oct 05 04:02:29 it should have been a forward slash Oct 05 04:02:44 then it would have been a command to IRC - nothing to do with real life - honest ;-) Oct 05 04:03:20 ok, you give me a minute with jenny and I won't tell ... Oct 05 04:03:32 :-)) Oct 05 04:04:18 About your concerns re readability in sitemap Oct 05 04:04:22 (sorry site) Oct 05 04:04:38 It is not making it any worse since Oct 05 04:04:54 site:forrest/index is no less readable than lm:forrest.index Oct 05 04:05:17 read the 1635 stuff. don't see how this changes anything. Oct 05 04:05:20 the difference is with site.xml we can't create the document mapping user space to src space Oct 05 04:05:43 (we have two topics going at once, let me come back to the 1635 stuff) Oct 05 04:06:24 ah, i see. I need to explain more. my problem is not that I find site:sdfsdfsda hard to read or understand. With a bit more explanation that step of going down the virtual path in site is actually pretty easy. Oct 05 04:08:01 What I find hard is that I see lm:xyz in site.xml and have to look that up in lm to know what file is really the source. Now I can all do that in site then I'd have to open the second file lm or maybe even a stack of files. Oct 05 04:08:38 On the other hand: You get a lot more power using lm instead of site. So perhaps it does make sense. Oct 05 04:08:50 Ah, no, now I need to explain something. Oct 05 04:09:00 In site.xml we define the *user* space Oct 05 04:09:21 i.e. Oct 05 04:09:41 In locationmap we define the *source* space Oct 05 04:10:08 i.e. index.html = http:svn.apache.org/repos/forrest/turnk/blah Oct 05 04:12:12 yes, I understand that.So if you send me your Forrest and I look at site.xml I need to look at site and the lm to know what file is behind a menuitem. And perhaps I even have to understand and regexp to figure it out. Oct 05 04:12:39 Anyway. I see the tempting benefits. Oct 05 04:12:56 No... Oct 05 04:13:04 if I send you my forrest site.xml Oct 05 04:13:22 you see the user space, if you need to know the source space Oct 05 04:13:29 you need my LM yes, but... Oct 05 04:14:00 we make Forrest respond to domain.org/filemappings.html (or whatever) Oct 05 04:14:05 with a list like this: Oct 05 04:14:16 index.html -> http://.... Oct 05 04:14:30 a/page.html -> file:// Oct 05 04:14:32 etc. Oct 05 04:14:59 We could even create a contract that added it to the meta-info Oct 05 04:15:03 in the generated page Oct 05 04:15:10 OK, I did understand you correctly and I think that the above list is not a great help (though better than nothing). Oct 05 04:15:18 However: How about Oct 05 04:15:29 --> Ron_ (n=rblasch) has joined #for-oct Oct 05 04:15:33 i.e. meta name="sourceLocation" content="blah blah"/> Oct 05 04:15:45 feeding back the info from the lm as a commented site.xml. Oct 05 04:16:16 How would the comments be added? Oct 05 04:16:17 For example: Oct 05 04:16:35 Hi Ron, good to see you Oct 05 04:16:53 Hi Ross Oct 05 04:17:08 Oct 05 04:17:32 looks good, but how is @lmdoc added? Oct 05 04:18:38 adding a transformation that - when called - resolves the links and adds the resolved url as an additional attribute to the original element. Oct 05 04:18:59 so it is not in the original site.xml file? Oct 05 04:21:18 no, it wouldn't have to be. although, if possible, I'd try and write it back right into the original site so that I have the documentation right where I'm editing. Oct 05 04:22:03 I'm -1 for writing it back that would confuse users (they would edit in site.xml) Oct 05 04:22:30 If it is not in the original source then what benefit does this give over having a specific index page? (no reason why we can't have both though) Oct 05 04:23:05 Confuse: Maybe, so you could make it a comment instead. Oct 05 04:24:37 I'm haiving trouble understanding what benefit this gives Oct 05 04:24:55 if someone edits the LM then you have to regenerate the page to see the results Oct 05 04:25:11 benefit: I look at site.xml and see a menuitem with its label and the final url right there. With the list I'd have to 'store' the virtual url in my memory (bad) and find it in the list. Imagine doing that in a list of 5000 pages, 500 of which are called index.html Oct 05 04:25:18 I see it as being equal to Davids suggestion Oct 05 04:25:34 OK - got it Oct 05 04:25:48 even with incremental search you'd have to type a lot to look up the final url. Oct 05 04:26:04 I have the solution I think ;-) Oct 05 04:26:10 Think views... Oct 05 04:26:20 confusion starting Oct 05 04:26:29 ;-) Oct 05 04:26:52 In develope/editorr view the menu you have the src file location in the meta informatiopn Oct 05 04:26:56 you also have it in the footer Oct 05 04:27:06 you also have it in a yellow hint box when you hover over a link Oct 05 04:27:14 you also have it sung to you by the voice plugin Oct 05 04:27:23 you also have it... Oct 05 04:27:42 In "reader" view, you just don't need to see any of these Oct 05 04:28:00 So, you don't even need to open site.xml anymore Oct 05 04:28:07 wait, I think you are talking the other direction of lookup. Oct 05 04:28:13 (unless you want to edit it of course) Oct 05 04:28:33 Ah, yes, I am Oct 05 04:28:42 You mean I have a src and I want to know where it appears Oct 05 04:28:50 Now I understand the problem... Oct 05 04:29:10 The locationmap tells you this though Oct 05 04:29:13 look at... Oct 05 04:29:59 right! and that is pretty simple since site.xml and the url of the page in my browser are the same. Oct 05 04:30:15 http://svn.apache.org/repos/asf/forrest/trunk/main/webapp/locationmap-transforms.xml Oct 05 04:30:30 and search for pattern="transform.*.*" Oct 05 04:31:08 Now think of Davids "index" Oct 05 04:31:36 it can also show that the file html2document.xsl is used by .... Oct 05 04:31:54 (bad example, I should have picked an XML LM entry, but you get the point I hope) Oct 05 04:32:12 OK, in a minute. Just let me finish that thought: I'm thinking that I (Editor) will maintain something like site-author while David is on holiday (David, don't get any ideas!). Oct 05 04:33:06 And now want to move all documents for version 0.8 to a new physical location. So I need a quick way to tell from site where the actual document is located. Oct 05 04:33:16 OK, now I read your link. Oct 05 04:34:12 so far so good... Oct 05 04:34:43 (I'm now on baby duty so my responses may be delayed) Oct 05 04:35:24 I read that link but I didn't see this change my mind. Oct 05 04:36:51 in your scenario you have changed physical location of the source files? Oct 05 04:37:23 I guess the point I'm traing to make is that (ideally) a quick look at any menuitem in site.xml should tell me where the document/data is coming from. Any need to look that up will make maintenance harder. Oct 05 04:38:21 yes I assumed that I needed to restructure parts of a site. Otherwise I wouldn't much care about the actual location of files. Oct 05 04:39:00 In 99.9% of cases the user will not change the defaultlocation of a file... Oct 05 04:39:16 so you already know that they are in content/xdocs Oct 05 04:39:45 using lm only gives you the *option* to get them from somewhere else Oct 05 04:40:32 having a document map from site.xml to locationmap (as David suggests) solves the problem if you use that option Oct 05 04:40:36 hmmm. you have a point there. Oct 05 04:41:15 Even more important, this is irrelevant to the whole "do we replace site: and ext: with lm:" discussion because... Oct 05 04:41:38 the problem is still the sam in that Oct 05 04:41:50 site:index gets resolved to index.tml Oct 05 04:42:01 which may be redirected by the lm anyway Oct 05 04:42:24 (although readability is still an impotant issue) Oct 05 04:43:52 Wait. The fact that is can alreay happen does not mean that it is w/o problem. But as I said. I see the power and I'm in favour of doing it Oct 05 04:44:37 ok, so we need to solve the readability issue... Oct 05 04:44:48 in my view it is only a problem for non-technical users Oct 05 04:44:57 but I'd try to document the resulting uris as close to the site-entry as possible. Oct 05 04:45:38 Well, not really. Even if you consider me a non technical user Oct 05 04:46:36 I'd say that even a technical user would take much longer to figure out where a uri is pointing at when he deals with a lm that he hasn't designed. Oct 05 04:46:45 OK, lets consider your annotated site.xml route Oct 05 04:47:00 if we write it back as you suggest... Oct 05 04:47:12 how does the user know if it is up to date? Oct 05 04:48:08 write a datestamp with it. It doesn't have to be perfect. Oct 05 04:48:45 In my view it could even me an optional command "forrest selfdoc" that you only call when you need it. Oct 05 04:49:21 I don't think a timestamp helps, that means you have to remember when the lm was last edited Oct 05 04:49:30 even worse in a multi-user environemnt Oct 05 04:49:42 the selfdoc is and option Oct 05 04:50:04 I'd be OK for that to happen Oct 05 04:51:13 since it is optional (I don't want my own files being overwritten automatically) Oct 05 04:52:53 True. Multiuser enviroments are a concern. But then: the whole concept of site.xml is questionable in a mu-environment. Oct 05 04:54:05 Once there is more than one editor we should really move site and tabs into a database with proper locking of individual records (objects) Oct 05 04:54:09 good point - in an multi-user I use Daisy and the navigation structure in there Oct 05 04:54:44 OK, decision made, we need a selfdoc target - care to make an issue? Oct 05 04:55:31 and if you do, you can simply add the extra info in a r/o table that is linked to the edited records and shown along with them. Oct 05 04:58:30 ok, will do that right away. Oct 05 04:59:05 q: how do I get rid of the debug line that jetty is currently throwing at me by the dozent? Oct 05 05:00:14 Thats's Cooon debug info Oct 05 05:00:42 I think we need to change the log level' Oct 05 05:00:51 not loked into it though Oct 05 05:03:24 Ron, are you still about? Oct 05 05:09:37 Yup. Oct 05 05:10:17 What's up, Ross? Oct 05 05:11:08 I figured that but didn't figure out where. In forrest-core.xconf that setting is commented out so I figured that would default to minimal logging Oct 05 05:11:30 issue is created Oct 05 05:12:59 Ron: nothing up, just wondered if had plans for the profiling plugin Oct 05 05:14:14 Sure. Here's what I've come up with: Oct 05 05:14:27 - It should be an internal plugin Oct 05 05:14:50 cocoon-profiler-block goes to lib/. Oct 05 05:15:43 - The profiler roles should move to forrest.xconf. Oct 05 05:16:29 ok, will sign off for now. Oct 05 05:16:41 see you later ferdinadn Oct 05 05:16:52 I'm still a bit shaky on the internal plugin side Oct 05 05:17:16 Ron, I'm just aboiut to have my tea, I'll be back in 10-15 mins, please outline your plan I'll catch up when I return Oct 05 05:17:20 \away Oct 05 05:18:42 Ross: enjoy your tea Oct 05 05:19:42 ... and main/webapp/profiler.xmap will also move to the plugin, along with the stylesheets Oct 05 05:20:56 One problem I haven't completely figured out yet is that the profile result got to be rendered last. Oct 05 05:21:32 A possible solution would be to enable profiling via a command line switch, which would add the page to cocoon.Main. Oct 05 05:22:00 (see main/targets/site.xml) Oct 05 05:22:02 ... Oct 05 05:22:04 Oct 05 05:22:36 Oct 05 05:22:37 ... Oct 05 05:22:55 This isn't real Ant syntax, but you should get the idea. Oct 05 05:25:58 That's about it. Oct 05 05:27:33 I'm back... let me see... Oct 05 05:27:57 forrest.xconf should not be touched by a plugin since it makes the plugin Oct 05 05:28:20 Should be forrest-core.xconf Oct 05 05:28:22 ? Oct 05 05:28:38 However, I paln to enable plugins to add their own *.xconf Oct 05 05:28:45 (yes forrest-core.xconf) Oct 05 05:29:00 in fact it is already enabled if you use the XPatch functionlaity in Cocoon Oct 05 05:29:06 although this is due to be removed Oct 05 05:29:16 It will be OK, during development Oct 05 05:29:20 There's only one line that has to be moved out. Oct 05 05:29:21 Oct 05 05:29:53 Ok, we need an issue to remind us to do it Oct 05 05:29:59 (take it out that is) Oct 05 05:30:25 getting profiler output last... Oct 05 05:30:41 give me a few moments to look at what you refer to (I'm not familiar with it) Oct 05 05:33:12 Hmmm... I'm not sure what you are suggesting here? Oct 05 05:33:34 How does Cocoon know to render that page last? Oct 05 05:34:14 There's only 1 page given as argument, I think it's the linkmap. Oct 05 05:34:42 If we add another, it should get rendered after everything contained in linkmap. Oct 05 05:34:50 At least, that's the theory. Oct 05 05:35:03 Oh, OK, I didn;t realise we could do that with Cocoon. Oct 05 05:35:07 That's easy then :-) Oct 05 05:35:14 I hope so. Oct 05 05:35:35 I'll have to check the Cocoon source to be sure things work like I want them to be. Oct 05 05:35:49 Otherwise we'll have to come up with something else. Oct 05 05:36:49 <-- ferdinandSoethe has quit (Read error: 110 (Connection timed out)) Oct 05 05:38:12 My suggestion for "something else" would be to use Oct 05 05:38:27 a webtest script (or AntEater or whatever) Oct 05 05:38:42 One advantage of doing this is that we could create a specific set of profiling tests Oct 05 05:38:53 this would ensure that we have consistent output for comparison Oct 05 05:39:04 However, your approch is much easier and quicker to implement Oct 05 05:39:12 so should be the first version Oct 05 05:39:46 Just to see if I understood you correctly Oct 05 05:40:00 This would mean to start the webapp Oct 05 05:40:10 and run the script against the server? Oct 05 05:40:16 Yes Oct 05 05:40:43 Hmm, I like the idea. Oct 05 05:41:19 I'd be happy to help implement it if you can get the initial plugin together Oct 05 05:41:32 I've been playing with webtest on an interal project recently Oct 05 05:41:38 In fact... Oct 05 05:41:54 I'm hoping this may be a way of improving the tests within Forrest Oct 05 05:42:07 Perhaps your plugin should be a "test" plugin rather than "profile" plugin>? Oct 05 05:42:50 Or "benchmark?" Oct 05 05:43:50 One thing I don't know yet is how to contribute the code. Oct 05 05:44:04 Create an issue and attach patches? Oct 05 05:46:45 benchmark? misses the "test to make sure it works" Oct 05 05:46:51 test misses "benchmarking" Oct 05 05:47:02 We need a name capturing both, but it's only a name Oct 05 05:47:12 Contributing - issue + patch - yes Oct 05 05:47:48 Maybe I am missing the scope of the plugin. Oct 05 05:47:57 What do you have in mind? Oct 05 05:49:08 That is, what is the scope of the "test." Oct 05 05:50:22 (Contributing: ok) Oct 05 05:50:55 Are you familiar with Unit testing? Oct 05 05:51:12 Yes. Oct 05 05:51:28 OK, WebTest is JUnit for webapps Oct 05 05:51:36 It's not ideal for testing Forrest, but... Oct 05 05:51:53 we can at least use it (or something similar) to retrieve a page Oct 05 05:52:00 and check that it contains the right content. Oct 05 05:52:19 Ii.e. div class="foo" exists, title="bar" etc. Oct 05 05:52:30 It can also test pdF geeration Oct 05 05:52:59 If we double this as the scripts for your profiling we "kill two birds with one stone" as we say in Englan Oct 05 05:53:00 Oh, I see. Oct 05 05:53:49 Like I said, I think your first version should do the profiling Oct 05 05:53:50 I thought the profiling plugin should only provide some timings, how fast different pipelines are. Oct 05 05:54:38 I'm just tossing an idea around Oct 05 05:54:45 not necessarily making a recomendation Oct 05 05:54:53 And would you add tests like "/index.html must be rendered within xxx ms?" Oct 05 05:54:53 really I'm wondering... Oct 05 05:55:33 (I was wondering) if having a standard set of files being processed would make the prfiling more accurate Oct 05 05:56:11 (index.html within xxx ms) is also a good idea, for quickly ensuring we are not killing performance with new additions Oct 05 05:57:07 Are we talking about a benchmark site here? Oct 05 05:57:34 I think I may be confusing things by mixing the two issues Oct 05 05:58:13 I wasn't intending it to be a benchmarking site, but it could be used as such Oct 05 05:58:54 David and I have talked about this on the list - the benchmark thingy, that is. Oct 05 05:59:22 OK, I'll find the discussion in the archives - I have to admit I only skimmed those threads Oct 05 05:59:42 however, first I need to batch my son and put him to bed. Oct 05 05:59:49 back in about 30 mins Oct 05 06:00:00 \away being a dad Oct 05 06:00:15 Good. I'll grab something to eat in the meantime. Oct 05 06:00:19 one day I'll type the right slash for a command Oct 05 06:00:29 ;-) Oct 05 07:00:09 I think we are talking about two things, really. Oct 05 07:00:29 (1) "Are we serving the right content?" Oct 05 07:01:21 Yes, that is what I refer to as "tests" Oct 05 07:02:48 --> Ron__ (n=rblasch) has joined #for-oct Oct 05 07:03:17 wow, two rons! Oct 05 07:03:39 Oops, lost my connection for a moment. Oct 05 07:05:03 Strange, I feel tempted to talk to myself now. ;-) Oct 05 07:05:22 OK, I'll dissapper for a while... Oct 05 07:05:39 I'm just reviewing the mail archives re: benchmarking Oct 05 07:05:49 I note that David has made a start on a benchmarking site: http://marc.theaimsgroup.com/?l=forrest-dev&m=112523204216082&w=2 Oct 05 07:06:34 Although if that is the only commit then there is nothing in there yet (not checked source) Oct 05 07:07:12 There is a template site. Oct 05 07:07:59 Yes, that is what David has done, I realise now. Oct 05 07:08:33 So should the benchmark sites, also be the test sites? Oct 05 07:09:03 I haven't thought about this yet, really. Oct 05 07:09:10 I may well be. Oct 05 07:09:21 How about this... Oct 05 07:09:28 The idea was that the benchmark sites should change rarely. Oct 05 07:10:00 Ah, yes, that is important, whereas test sites Oct 05 07:10:09 whereas test sites will change when features change or Oct 05 07:10:12 when bugs are uncovered Oct 05 07:10:45 Right. Oct 05 07:11:07 I think they should be separate becuase of that Oct 05 07:11:17 that solves the naming issue - benchmark it is Oct 05 07:11:39 Lets leave testing for another day Oct 05 07:12:45 Right, but I like the idea. Didn't know it was easily possible to do that. Oct 05 07:13:25 http://webtest.cannoo.com Oct 05 07:13:36 Definitely would want to have this for my pod-input plugin. Oct 05 07:14:16 sorry - http://webtest.canoo.com/manual/WebTestHome.html Oct 05 07:14:44 It's not ideal for testing intermediate stages (like XML) but good for the end result Oct 05 07:14:54 If you want it in your pod-input plugin Oct 05 07:15:11 we should work out a centralised testing framework Oct 05 07:15:25 a standard way for plugins to provide tests Oct 05 07:15:43 currently all we do is require that the docs build Oct 05 07:15:54 see the "test" target in plugins/build.xml Oct 05 07:16:23 Sounds good. Oct 05 07:19:24 So, for the benchmark I'll start with a plugin in "org.apache.forrest.plugin.internal.benchmark," Oct 05 07:19:44 and submit an issue as soon as things are somewhat running. Oct 05 07:20:24 <-- Ron_ has quit (Read error: 110 (Connection timed out)) Oct 05 07:20:36 --- Ron__ is now known as Ron_ Oct 05 07:21:13 sounds great - I'll be on and off IRC for at least another hour or two. Oct 05 07:21:19 holler if you wnat help with anything Oct 05 07:22:09 Cool, thanks. Oct 05 07:26:19 But I guess I won't get too far today, as it is already late in Austria (UTC + 2). Oct 05 07:55:22 --> Ron__ (n=rblasch) has joined #for-oct Oct 05 08:03:14 <-- Ron_ has quit (Read error: 110 (Connection timed out)) Oct 05 08:03:40 --- Ron__ is now known as Ron_ Oct 05 08:03:55 Ross, you still there? Oct 05 08:13:33 Yes Oct 05 08:14:09 I've got the basic setup running. Oct 05 08:14:23 Wow - that was pretty damn quick Oct 05 08:14:44 Yup, things went faster than I thought. Oct 05 08:14:51 There's 1 thing missing, though. Oct 05 08:15:07 go ahead... Oct 05 08:15:08 The map:pipe. Oct 05 08:15:33 sorry, not following... Oct 05 08:15:38 I would need to replace the map:pipes of sitemap.xmap. Oct 05 08:15:49 There's set to the non-profiling parts. Oct 05 08:15:49 let me look at that... Oct 05 08:16:08 The profiling page currently only shows cprofile.html. Oct 05 08:16:35 Ahhh.. I see the problem, the map:pipe is defined in sitemap.xmap and therefore cannot Oct 05 08:16:37 It works as soon as I set the map:pipes in sitemap.xmap to the profiling parts. Oct 05 08:16:45 be redefined in the plugins xmap Oct 05 08:17:01 Right. Oct 05 08:17:07 that's a problem Oct 05 08:17:14 let me think.... Oct 05 08:18:33 I can't think of any way to get around that Oct 05 08:18:40 I think we'll need to ask the Cocoon community Oct 05 08:18:46 are you subscribed to their dev list? Oct 05 08:18:51 Yup. Oct 05 08:19:00 I'll take care of it. Oct 05 08:19:06 Do you want to ask them, or shall I (it might be better on the user list actuallY) Oct 05 08:19:38 Cool - thanks Ron Oct 05 08:20:51 I'll look into this and keep the forrest dev list posted. Oct 05 08:21:53 I'll leave it there for today. Oct 05 08:22:23 Thanks for your help, Ross. Oct 05 08:22:36 Bye, everyone. Oct 05 08:27:18 <-- Ron_ (n=rblasch) has left #for-oct Oct 05 08:40:11 anyone still around? Oct 05 08:44:23 I think there are a few. Oct 05 08:45:27 hi, i am reading back through the log to see what's been happening Oct 05 08:49:00 Wow, Paul you are a dedicated lurker :-) Oct 05 08:49:06 Morning David Oct 05 08:49:17 Loads of fruitful discussion Oct 05 08:49:24 :) Oct 05 08:49:35 SO Paul, do you have a working Forrest yet? Oct 05 08:51:05 Paul: i just found your message to dev sitting in the moderation queue. Deleivered now. Oct 05 08:51:14 Hi Ross Oct 05 08:51:42 Thanks. It's probably something I should know, but I don't. Oct 05 08:52:00 Paul: you need to move old jxpath.jar completely out of the way, cannot just rename it Oct 05 08:52:27 Hmm. Will try. BRB. Oct 05 08:53:09 Nothing else needs to change, it just finds it. Right? Oct 05 08:53:23 (Same error) Oct 05 08:55:24 I'm afraid I can't help with that one Paul Oct 05 08:55:40 David has been playing with JXPath, he may help when he's finished reading the logs Oct 05 08:56:36 I think I found the problem (between chair & keyboard...) Oct 05 08:57:39 :-)) we get alot of those around here (I've broken trunk twice today) Oct 05 08:57:51 Just in case anyone listening is curious - a .tar.gz file is not the same as a .jar... Just for reference... Oct 05 08:58:07 :-)) Oct 05 08:58:45 I get a bunch of 'No pipeline matches request' errors but it works... Oct 05 08:59:41 I know why that happens... Oct 05 09:00:04 There shouldn't be any of those errors (unless you have not done svn up today, I fixed an issue with that this afternoon) Oct 05 09:00:45 One bunch was because I didn't have the output.pdf plugin requested. Oct 05 09:01:02 The only ones left are these: Oct 05 09:01:04 X [0] samples/error:ext:forrest/webapp BROKEN: No pipeline matched request: samples/error:ext:forrest/webapp Oct 05 09:01:25 OK, that is a known issue with the new JXPath library Oct 05 09:01:34 I guess that is what you meant by "I know why"? Oct 05 09:01:59 Actually that was more for the PDFs but, I'll take it.. :) Oct 05 09:02:10 -) Oct 05 09:02:17 So you are jumping into views then? Oct 05 09:03:36 I've sort of been there since Diwaker announced his site about 4 months ago... Oct 05 09:03:55 OK, you been there since day one then Oct 05 09:04:01 It's just been broken for a bit and I wanted to see you good folks at work. Oct 05 09:04:15 they've been around for longer but until Diwaker built his site I think only Thorsten understood them Oct 05 09:04:44 These Forrest Tueday things are new Oct 05 09:04:52 we are still working out how to best use the time Oct 05 09:04:56 I've been following .dev since about the same time. Oct 05 09:05:13 Any suggestions with what you have seen? (both here and onlist) Oct 05 09:05:26 They are difficult to organize. I work virtually 3 days/week so I understand some of the difficulties. Oct 05 09:05:41 Probably having an agenda of things to fix with people lined up to help with their part would be good. Oct 05 09:05:53 Maybe holding them on a weekend but that can be tough too. Oct 05 09:06:48 agenda + people: I agree this is important, but... Oct 05 09:07:02 being voluntary it's hard to get people to commit time Oct 05 09:07:11 especially on a work day Oct 05 09:07:19 which brings us to weekends... Oct 05 09:07:37 quite a few of the devs resist weekends (or is it just me?) Oct 05 09:07:43 Getting a good momentum running can help quite a bit too... Oct 05 09:07:59 Once a sense of accomplishment hits, it's easier to keep going. Oct 05 09:08:33 momentum: I agree Oct 05 09:08:39 this was one of the great things about Oct 05 09:08:44 our test with Gobby Oct 05 09:08:54 we could all *see* something happening Oct 05 09:09:03 in code Oct 05 09:09:10 I didn't get a chance to try it, but it sounds very interesting. Oct 05 09:09:27 That'd probably help me get oriented to the code too. Oct 05 09:09:31 You were here when I was explaining the locationmap to Gav, right? Oct 05 09:09:53 I've been on during the day, but working. I've looked back a little. Oct 05 09:10:06 OK, in short... Oct 05 09:10:16 Gav wanted to know what I was doing with the locationmap Oct 05 09:10:21 he didn't know how it worked Oct 05 09:10:29 I was half way through explainng Oct 05 09:10:34 pointing at files in svn Oct 05 09:10:39 cutting snippets to explain them etc. Oct 05 09:10:49 Yep, got it. Oct 05 09:10:55 and he said "I think I'd get it better if I worked on the code" Oct 05 09:11:13 With Gobby I could have explained it Oct 05 09:11:18 opened the files for him to see Oct 05 09:11:27 changed a file to show him how it worked Oct 05 09:11:35 Yep, that would have been helpful. Oct 05 09:11:39 run a version of forrest and shown him the results in a browser Oct 05 09:12:39 Maybe next month... Oct 05 09:12:41 :) Oct 05 09:12:42 So, what did you have in mind when you said we need to get momemtum going? Oct 05 09:13:47 Find something that's relatively easy to fix but is "pissing off" a number of people and get it working... Then try to tackle something a little harder. Oct 05 09:14:04 (I'm still trying to remember how to format stuff for IRC) Oct 05 09:14:27 Even solid progress on understanding how Thurston thinks could do it ;) Oct 05 09:14:38 lol Oct 05 09:15:02 Yeah, we need a Gobby that gives us a direct link to his bran patterns Oct 05 09:15:03 J/K... Something to give people a sense of accomplishment... Oct 05 09:15:07 too damn smart for me! Oct 05 09:15:17 :> Oct 05 09:15:45 xley: My skins site is "compiling" rather nicely now. Thanks! Oct 05 09:15:54 good Oct 05 09:16:09 I'd still like to see cocoon cache things effectively, but that's another day. Oct 05 09:17:13 BTW the faster i go, the behinder i get. Try to read logs and follow current and read last night email :-) Oct 05 09:18:06 just read the logs - eventually you'll catch up with chatter here - don't forget your coffee though, I'm about to get my whiskey Oct 05 09:18:19 I think Paul makes a valid point about momentum on these Forrest Tuesdays Oct 05 09:18:44 The problem I see is that people come and go throughout the day Oct 05 09:18:57 any ideas how we can get around that? Oct 05 09:19:18 rgradler: ah, i was worried that you wouldn't be around for much longer Oct 05 09:19:37 More interaction (Gobby-style) could help. Oct 05 09:19:48 come and go is good too Oct 05 09:20:09 Sure, if that's what lets people participate. Oct 05 09:20:16 But you need to capture their attention while they are here. Oct 05 09:21:06 (Sweet - Gentoo includes a build for Gobby!) Oct 05 09:21:27 If folks wander in and don't see anything really going on, they'll leave more quickly. Oct 05 09:22:04 they need to learn to wait, Oct 05 09:22:06 I agree, but Oct 05 09:22:15 let me start again... Oct 05 09:22:27 If there's no activity for 15 minutes, they'll likely go do something else. Oct 05 09:22:31 I agree that we need people to participate Oct 05 09:22:42 but at the same time one of the goals is to attract new people Oct 05 09:22:53 * kronenpj waves. Oct 05 09:22:55 provide some "entry level" help with forrest Oct 05 09:23:17 I was really pleased that Gav wanted to know about the Locationmap stuff Oct 05 09:23:25 and that Paul is also lurking Oct 05 09:23:38 If we had a more limited scope would we be in danger of killing that? Oct 05 09:24:07 (Paul *was* lurking that is) Oct 05 09:24:23 True, but if you have a purpose and are active in pursuing it, I think you'll keep the people you attract. Oct 05 09:24:34 So, if this is a dev activity, gauge the activity for devs. Oct 05 09:24:45 If you feel a user activity is needed, start another one. Oct 05 09:25:08 You'll have a tougher time keeping users in a dev channel and vice-versa. Oct 05 09:25:23 Unless users have a keen interest in becoming devs... Oct 05 09:25:24 yeah, it is a dev activity (at least *I* think it is) Oct 05 09:25:43 How about we do something like Oct 05 09:25:46 And with the things *I* understood you wanted to accomplish, I agree. Oct 05 09:26:05 have a theme for each FT that we will work on through the day Oct 05 09:26:18 but encourage potential and existing devs to come Oct 05 09:26:30 with specific issues they wish to address Oct 05 09:26:42 when they raise them we break off from the main task Oct 05 09:26:43 * xley almost up to speed with current topic Oct 05 09:26:55 That works, or devote a Tuesday to answering user questions... Oct 05 09:27:13 OK. Oct 05 09:27:25 This is only the third time I have used IRC Oct 05 09:27:33 It is great for Talk - but rubbish for doing Oct 05 09:27:41 enter Gobby... Oct 05 09:27:51 Right. Oct 05 09:28:17 Lets see what David thinks when he catches us, since he s nearby Oct 05 09:28:31 Ok. A slight tangent, if I might. Oct 05 09:28:37 go ahead Oct 05 09:28:47 If someone comes up with a question that doesn't quite Oct 05 09:28:55 fit the theme, but someone is willing to answer it. Heave Oct 05 09:29:08 Damn, Jenny's automated log broken Oct 05 09:29:12 Have the two or however many go off onto another IRC channel. Oct 05 09:29:14 http://casa.che-che.com/~bot/forrest/forrest.log.05Oct2005 Oct 05 09:29:18 empty Oct 05 09:29:20 Then come back when they're done. Oct 05 09:29:41 the live log is so important Oct 05 09:29:50 jenyy dead :-(( Oct 05 09:29:53 Oops.. Oct 05 09:29:56 We need two machines running in future Oct 05 09:30:33 "Split to separate channel" -> I'm OK with this but... Oct 05 09:30:40 It happened before to, the rollover after midnight in Spain Oct 05 09:30:50 it is critical that all communications are logged Oct 05 09:30:56 JennyCurran is still here Oct 05 09:31:06 i have local local Oct 05 09:31:07 Agreed, so there should be a process for doing it. Oct 05 09:31:10 log log Oct 05 09:31:23 :-) Oct 05 09:31:38 I have *one* whisky, maybe I should have a whiskey whiskey Oct 05 09:31:41 will upload, but manual is wasted effort Oct 05 09:32:01 if that keeps you going, yes yes Oct 05 09:32:34 Paul: I think Gobby may help here s well Oct 05 09:32:45 There is an IRC like chat mechanism Oct 05 09:33:09 that could be used for the prelim chat "can someone help me with XYZ" typw things Oct 05 09:33:16 Sure. Oct 05 09:33:22 but when something needs to be done a new file Oct 05 09:33:24 could be created Oct 05 09:33:30 it would be brilliant if that was actual IRC, so gobby piggybacked Oct 05 09:34:13 oin the new file we write a document about what to do and why Oct 05 09:34:16 Agreed, but it *sounds* like it's not... Oct 05 09:34:21 self documenting "discussion" Oct 05 09:34:37 Gobby piggy-back: it is not. but it is only version 0.2.2 Oct 05 09:34:54 unfortunately it is GPL and uses lots of tech I am not familiar with Oct 05 09:35:03 if we have a focussed discussion, with people waiting to give each other a chance, the irc can be very useful, even to solve an issue Oct 05 09:35:44 Well, the chat in Gobby is the same as IRC except Oct 05 09:35:49 it is not over IRC channels Oct 05 09:35:56 I mean it looks the same and behaves the same Oct 05 09:36:42 this morning i explained to Thorsten the JXPath bug. I could correct his misconceptions on the spot. Oct 05 09:36:59 * kronenpj needs to run for dinner... Oct 05 09:37:17 Paul, thanks for your input - really helpful - enjoy dinner Oct 05 09:37:18 Ross - It's likely you'll be asleep before I return.. Oct 05 09:37:33 I can continue with you on the dev list if you'd like... Oct 05 09:37:42 Yes, I think that would be good Oct 05 09:38:11 please do participate as much as you can manage Paul Oct 05 09:38:35 David, I agree IRC style chat *is* useful Oct 05 09:38:37 However Oct 05 09:39:00 the only advantage of IRC over Gobby is most people have at least heard of it Oct 05 09:39:16 and there are many clients Oct 05 09:39:41 What Gobby gives is the ability to share files, illustrate what you are talking about and Oct 05 09:39:53 actually *do* stuff at the same time achatting Oct 05 09:40:12 i was looking for a way to enable those who can't run gobby a way to participate (for short term solution) Oct 05 09:40:33 Yes - that is the *real* problem with it Oct 05 09:40:34 i like the sound of gobby-like tools Oct 05 09:40:44 and until we solve that it is *not* an option Oct 05 09:41:18 You know I wish I had the time to write one in Java and put it out under the apache license Oct 05 09:42:17 amazes me that there isn't yet Oct 05 09:42:36 yeah, me too Oct 05 09:42:42 I'm jut looking at Jabber Oct 05 09:42:48 should be one using thier open protocols Oct 05 09:43:04 that had occurred to me too Oct 05 09:43:05 (I wish GSoC were happening now I know the project I'd want) Oct 05 09:44:03 (the inkboard addition to Inkscape was a GSoC project, did you know) Oct 05 09:45:16 I din't know that - GSoc projects did so many cool things Oct 05 09:45:52 see the press release at http://www.openclipart.org/ Oct 05 09:46:42 OK, while I read that you might want to read http://wiki.inkscape.org/cgi-bin/wiki.pl?Inkboard2.0 Oct 05 09:48:48 thanks for the ref, they have a way to go then. Oct 05 09:49:02 unfortunately, yes Oct 05 09:49:28 note the opensource lesson though ... Oct 05 09:50:09 He says: "I inherited this codebase, and although I have the design documents, I don't have the original designers and coders to talk to. " Oct 05 09:51:17 at least he has design documents Oct 05 09:51:39 yeah, and they'll look *exactly* like the code! Oct 05 09:51:39 we have some pretty good doco at forrest Oct 05 09:52:12 Only for some parts Oct 05 09:52:22 I'm amazed at the number of things I'm finding in the sitemap Oct 05 09:52:27 that I have no idea what they do Oct 05 09:52:32 e.g. Jeff's linking docs Oct 05 09:52:38 but we don't have Jeff Oct 05 09:53:22 Unfortunately, many of our docs are out of date Oct 05 09:53:24 svn log helps a little bit Oct 05 09:53:56 but shows that we need to be better commit descriptions Oct 05 09:54:28 I really like the "auto documenting" of Jira Oct 05 09:54:30 as usual we need for people to constantly fine-tune the docs Oct 05 09:54:40 Just put the Jira Issue number in the commit message Oct 05 09:54:43 and hey presto Oct 05 09:54:52 doens't help with the docs Oct 05 09:54:59 but helps decipher what is going on Oct 05 09:55:00 yes ... brilliant Oct 05 09:55:12 we must be more diligent as committers Oct 05 09:55:27 yes. Oct 05 09:55:45 we should have an svn template Oct 05 09:56:04 ... Oct 05 09:56:32 would people use it? I always type svn ci foo -m "bar" Oct 05 09:56:35 and what about GUI's Oct 05 09:56:57 yes that is problem Oct 05 09:57:29 i always go 'svn commit' and rely on my editor for message Oct 05 09:57:41 that also gives a chance Oct 05 09:57:58 to abort the commit and edit the message Oct 05 09:58:13 i wish people would not use GUIs Oct 05 09:58:42 Yes, too many spelling errorin my commits because of -m "brr" Oct 05 09:58:56 as for GUI's they are here to stay :-( Oct 05 09:59:06 grrr Oct 05 10:00:20 so... Oct 05 10:00:34 anyway back to those sitemaps Oct 05 10:00:52 sorry, interrupted you Oct 05 10:01:01 go ahead (I was going to site: and ext:) Oct 05 10:01:24 i have yet read today's IRC log about site: etc. Oct 05 10:02:03 (back to those sitemaps...)? Oct 05 10:03:05 you said that parts confuse you, anything that i can help with ... Oct 05 10:03:13 or just a general comment. Oct 05 10:03:23 No just a general comment Oct 05 10:03:32 thought so Oct 05 10:03:33 it's not that the pipelines confuse me Oct 05 10:03:35 ... Oct 05 10:03:43 just that I didn't know they were there Oct 05 10:03:49 or even if they are used Oct 05 10:04:08 shows that we need to sprinkle some comments in our code. Oct 05 10:04:34 yes, I'm hoping Ferinand will find the time to make his recomendations (an onlist discussion) Oct 05 10:04:41 yes our sitemaps have grown and grown Oct 05 10:05:08 I'm working on the resources.xmap right now Oct 05 10:05:11 boy is it complex Oct 05 10:05:17 and it does nothing! Oct 05 10:05:26 (apart from read a few files) Oct 05 10:05:37 with lm it is going to shrink loads Oct 05 10:08:10 good work, thanks. On another topic ... Oct 05 10:08:23 Paul mentioned earlier about IRC techniques ... Oct 05 10:08:55 people tend to go away when there is a quiet spot Oct 05 10:09:24 yep Oct 05 10:09:36 that might be so for general IRC constant sessions Oct 05 10:09:48 but for a day like today Oct 05 10:10:09 it should not happen Oct 05 10:10:18 ... Oct 05 10:10:38 does ... mean you are finished talking? Oct 05 10:10:47 i for one just get on with something related to the day Oct 05 10:11:06 (no it doesn't ;-) Oct 05 10:11:08 (drat sorry means im still talking) Oct 05 10:11:17 carry on Oct 05 10:11:22 :-) Oct 05 10:11:36 i for one just get on with something related to the day ... Oct 05 10:12:23 when the IRC window beeps, then check back. Oct 05 10:12:48 ... Oct 05 10:13:24 we need less talk and more doing. Oct 05 10:13:34 (over) Oct 05 10:13:40 :-) Oct 05 10:14:02 Yeah, I've been working on the LM all day Oct 05 10:14:08 but I have also been talking an awful lot Oct 05 10:14:12 (suprise!) Oct 05 10:14:15 great Oct 05 10:14:34 To be honest, I have found this FT far better than the first and our interim one Oct 05 10:14:41 Why? because... Oct 05 10:14:56 Gav and Paul were here Oct 05 10:15:01 (new folk) Oct 05 10:15:22 Ferdinand came with a valuabule question about the sitemaps and what to do Oct 05 10:15:28 (see something happening) Oct 05 10:15:55 I've gathered lots of peoples views on site: and ext: Oct 05 10:16:03 (probably could have been done onlist though) Oct 05 10:16:12 but... Oct 05 10:16:27 I've got almost nothing *done* on the locationmap stuff Oct 05 10:16:46 What w need is Oct 05 10:16:56 a task that can be broken into a number of small parts Oct 05 10:17:05 the LM stuff is ideal (each xmap is separate) Oct 05 10:17:25 Gavin and I started on that, but he is on the wrong timeframe for me Oct 05 10:17:33 (and not as addicated as I am) Oct 05 10:17:38 (over) Oct 05 10:18:09 (that is why i didn't want to deal with "site:" removal today) Oct 05 10:19:21 i reckon we need to come to these days with a specific goal, ... Oct 05 10:19:45 and generally do the rest on the list ... Oct 05 10:20:15 Sounds like Gobby-like tools would help keep that focus. Oct 05 10:21:23 Yes, although... Oct 05 10:21:36 there is a danger with a "specific goal" that Oct 05 10:21:47 we get into a mess like the XHTML2 stuff Oct 05 10:21:57 I was really pissed off about that... Oct 05 10:22:02 not *at* anyone... Oct 05 10:22:15 but that the approach was considered incorrect... Oct 05 10:22:37 Thorsten was nto there so... Oct 05 10:22:46 his vision could not be included... Oct 05 10:23:02 if we have a specific goal I think it needs to be either Oct 05 10:23:09 1) very clearly mapped out (is that possible) Oct 05 10:23:24 2) someone is there to "drive" the objective forward and ensure that Oct 05 10:23:35 when people arrive they understand the overall vision Oct 05 10:24:46 team should keep it on track. Newcomers need to read the backlog Oct 05 10:25:05 no never can clearly map beforehand Oct 05 10:25:43 (over) Oct 05 10:26:25 team keep on track requires there to be a clear vision in the first place doesn't it? Oct 05 10:28:20 yes, but let it begin and then evolve. I dunno Ross, what will work. Oct 05 10:29:17 There was an interesting mail in the Cocoon dev list today.. let me find a reference... Oct 05 10:31:19 "And we have stopped waisting time and energy in the "my way or no way" ego battles that stopped so much development initiatives before. from... Oct 05 10:31:26 http://marc.theaimsgroup.com/?l=xml-cocoon-dev&m=112841737012179&w=2 Oct 05 10:31:49 I wonder if we are suffering the same problem (and it could be my ego so people reading the logs shuld not get paranoid) Oct 05 10:32:57 interesting that was the last post that i read yesterday. Oct 05 10:34:00 (each morning, like now, i have a deluge of eurpean mail to catch up on.) Oct 05 10:38:18 do you think it could be relevant to us as a project? (go on stick your neck out, I like others with big egos have thick skins) Oct 05 10:39:16 ego probably traps everyone, even if we think that we are beyond it. Oct 05 10:40:06 you are such a diplomat :-) Oct 05 10:40:13 opensource projects really put this in our face Oct 05 10:41:08 yes i think our project suffers too, not as bad as some. Oct 05 10:42:08 we are only a small community - time for more yet ;-) Oct 05 10:42:56 OK, so, concluding the Forrest Tuesday discussion... Oct 05 10:43:05 I think they have worked well, but Oct 05 10:43:14 agree that we have not got much done... Oct 05 10:43:29 For the next one we need to start planning at least a week in advance Oct 05 10:43:47 If I had suggested a release a week ago Oct 05 10:43:58 we could have made the LM the focus of this one Oct 05 10:44:09 is that about what you are saying? Oct 05 10:45:25 yes that is sufficient to give some focus ... Oct 05 10:45:42 \unaway Oct 05 10:46:04 not that we i opened this IRC channel i didn't know what to set as the Topic Oct 05 10:48:04 (Hi Paul) Oct 05 10:48:25 Yeah, I think with us having the "interim" IRC meeting we forgot (I know I did) Oct 05 10:48:26 A related topic from earlier ... Oct 05 10:48:53 about when to hold these gethogethers ... Oct 05 10:49:11 >>rgardler: quite a few of the devs resist weekends (or is it just me?) Oct 05 10:49:37 I, for one, am usually fine with anytime, no aversion to weekends. Oct 05 10:50:38 Why don't we have it weekends, I don't remember (but do know I prefer not to, so am worried it might be jsut me) Oct 05 10:51:16 what about Friday so that is overlaps with the weekend Oct 05 10:51:59 you ask why ... Oct 05 10:52:18 * kronenpj didn't. Oct 05 10:52:54 IIRC it was because Ross had a "slight preference for Tuesdays". Oct 05 10:53:13 OK, well that is *not* good Oct 05 10:53:32 I certainly have a preference for a weekday, and Tuesday is good for me Oct 05 10:53:51 however, if weekend is better for most, my "preference" is not important Oct 05 10:54:33 Both work for me, but probably because self-employed and work from home. Oct 05 10:55:08 yeah, that is my concern too, we are in the minority Oct 05 10:55:20 Perhaps some sort of a vote or poll on the dev list? Oct 05 10:55:32 weekends are OK, but I will much spend less time here Oct 05 10:55:38 that is not necessarily a bad thing Oct 05 10:55:47 Remember the trouble we had trying ... Oct 05 10:55:49 Paul is right, lets review this onlist Oct 05 10:56:19 If a few (~3) options are polled a decision might be more forthcoming. Oct 05 10:56:38 to get people to decide a time for the XHTML2 IRC was difficult. Oct 05 10:57:15 And it always will be... Oct 05 10:57:31 I think you're looking for a good number of people who can make it. Oct 05 10:57:48 There also may be critical people, depending on the objectives. Oct 05 10:58:07 In that case, it'd be somewhat important for them to be there... Oct 05 10:58:26 (over) Oct 05 10:58:55 Oh, sometimes changing it can attract a different set of people Oct 05 10:59:02 than you might get if it's *always* on day X. Oct 05 10:59:06 (over) ((again)) Oct 05 10:59:09 The session day is known well in advance, so that people can plan for it. Oct 05 10:59:57 "critical" people - there shouldn't be any, we are a team, but... Oct 05 11:00:06 if someone has a vested interest in the objective, then Oct 05 11:00:15 they will make themsleves available (I know I would) Oct 05 11:00:21 There are people who are more familiar with, say views, than others are there not? Oct 05 11:00:27 "plan for it" I agree Oct 05 11:00:59 "poll dev list" - we do alright on Tuesdays, but some people who would probably like to participate can't due to work Oct 05 11:00:59 yes, this is a topic that we should be able to do on list. Oct 05 11:01:30 then they can do it after work, that is why it is a whole day. Oct 05 11:01:36 perhaps we should say on list "1) stay Tuesdays, 2) switch to Friday/saturday overlap" Oct 05 11:02:18 See what responses you get & go from there... Oct 05 11:02:46 yep, lets try that, but not as a Vote. Oct 05 11:02:55 +1 Oct 05 11:03:04 :-) Oct 05 11:03:08 +0.01 (non voting member) Oct 05 11:03:28 OK, I'm testing the resources.xmap conversion to lm: Oct 05 11:03:38 it's a little broken right now Oct 05 11:03:49 I'm going to get it working then ask you both to test because it is Oct 05 11:03:54 a very big set of changes Oct 05 11:03:56 ok? Oct 05 11:03:56 Paul: you mean +1 Oct 05 11:04:24 good, something we can collaborate on. Go. Oct 05 11:05:33 I'm currently trying to figure out how to work locationmap Oct 05 11:05:42 so let me know when to update & I can see what happens. Oct 05 11:06:07 (http://cfas-new.kronenpj.dyndns.org if you're interested) Oct 05 11:06:42 Take a look at the log at http://casa.che-che.com/~bot/forrest/forrest.log.04Oct2005 from 16:04 for the whistlestop tour of locationmaps Oct 05 11:06:53 I'll be back shortly when it is working locally Oct 05 11:07:52 At least I *think* it's locationmap... Searching... Oct 05 11:14:34 Ok, I'm lost. Would someone look at http://cfas-new.kronenpj.dyndns.org Oct 05 11:14:49 and tell me what's generating the breadcrumb trail at the top? Oct 05 11:15:05 And the myGroup/myProject trail under the header? Oct 05 11:16:56 src/documentation/skinconf.xml Oct 05 11:17:38 Ok, but I can't find where myGroup/myProject is in any source file... Oct 05 11:18:12 I don't mind the breadcrumbs, but I'd like them to have my content... Oct 05 11:19:28 Ah, this is a views-constructed site. Sorry can't help. Oct 05 11:21:06 its in the *.fv file Oct 05 11:21:57 Ah... Let me take a look... Oct 05 11:25:58 (just running a forrest site on site-author will commit in a moment if it works) Oct 05 11:26:20 David, if you are looking for something to do, a document on how to use IRC to work collaboratively Oct 05 11:26:33 would be great, really saying more do, less talk as you were saying before Oct 05 11:26:49 are you kidding ... me "looking" for something to do. Oct 05 11:26:58 :-)) Oct 05 11:27:14 Just thought I'd try :-) Oct 05 11:27:22 :-) Oct 05 11:27:27 I have a couple of broken links - back to resources.xmap Oct 05 11:28:39 you are a mind-reader. I have a todo.txt file for today's IRC remind me about things to ask or todo and that doc is one of them. Oct 05 11:29:18 broken links? Is that changes.rsss ... that was there yesterday. Oct 05 11:30:11 No, I fixed changes.rss earlier today, this is one I just introduced locally (damned image resolving again) Oct 05 11:38:25 If you get stuck then send your patch to Jira and i can try. Oct 05 11:38:52 I was just thinking the same Oct 05 11:39:08 I tried in site-author and got one set of errors now Oct 05 11:39:16 I'm trying fresh-site and get another Oct 05 11:39:21 I'm tired and confused Oct 05 11:39:24 patch on its way Oct 05 11:42:55 http://issues.apache.org/jira/secure/ManageAttachments.jspa?id=12317495 Oct 05 11:43:14 let me know whe you have it installed, I'll tell you symptoms Oct 05 11:45:50 done, now running 'build test' Oct 05 11:46:16 ok, under build test we can't find "skin/profile.css" Oct 05 11:46:24 but other css is found Oct 05 11:47:03 mine is also missing skin/basic.css ... Oct 05 11:47:21 oh... right light me check that locally Oct 05 11:49:05 and also gives strange error about the link for that css. Oct 05 11:49:28 starange error is a known issue (I have it on Jira) Oct 05 11:49:34 cat brokenlinks.xml Oct 05 11:49:37 at least I am assuming you get "access denied" Oct 05 11:49:56 no ... Oct 05 11:50:16 skin/basic.css BROKEN: /svn/asf/forrest/main/webapp/. (No such file or directory) Oct 05 11:50:16 OK, can you paste your output to http://www.pastebin.com Oct 05 11:50:36 Same error, different report, must be operating system Oct 05 11:51:08 I don't understand why you can't get skin/basic.css, I have that OK Oct 05 11:51:17 also errors with "site:" links, do you get that too? Oct 05 11:51:24 No Oct 05 11:51:39 Are you using the "views" JXPath library? Oct 05 11:52:07 i though that i removed that, was just going to verify. Oct 05 11:52:44 all okay with that. Oct 05 11:53:14 stranger and stranger Oct 05 11:53:28 hang on I'm looking into my local profile.css problem Oct 05 11:54:36 I am too tired silly mistake... Oct 05 11:54:46 ah found the "site:" problem. It is mine .. Oct 05 11:55:00 OK, you not woken up yet ;-) Oct 05 11:56:03 testing the jxpath bug with a namespace-uri parameter. Doing svn revert on that. Oct 05 11:57:18 --> __Gav_at_work (n=ipv6guru) has joined #for-oct Oct 05 11:58:04 <__Gav_at_work> Hi All, Just a quickie, howe long to go, I can carry on in about 3 hours Oct 05 11:59:10 the day officially goes for 5 more hours. We can decide at the time. Oct 05 12:00:22 <__Gav_at_work> Ok, good, I want to carry on with LM conversions, I guess RDG wont be around to guide though Oct 05 12:00:40 Hi Gavin Oct 05 12:00:52 David, do you still have the same problem? Oct 05 12:00:53 <__Gav_at_work> B**ger me, you still up! Oct 05 12:01:01 rgardler: the only errors now with my 'build test' are for each of "basic.css and print.css and screen.css" Oct 05 12:01:02 I'm a mug! Oct 05 12:01:15 good Oct 05 12:01:29 strange, I was only getting profil.css missing, are you finding that? Oct 05 12:02:04 ha, that one is okay for me. LOL. Oct 05 12:02:14 <__Gav_at_work> Sorry for butting in, why the need for all these different CSS files, I havent worked that out yet. Oct 05 12:02:15 ??? Oct 05 12:02:39 no error is reported for profile.css Oct 05 12:03:06 yeah, I understood that, ??? is toatl confusion Oct 05 12:03:24 I have another patc coming in a moment, perhaps Paul or Gavin can try it as well Oct 05 12:03:33 may be an OS issue Oct 05 12:04:05 <__Gav_at_work> I can't try atm while at work, back home on 3 hours. Oct 05 12:04:40 I'm back... I'm on Linux, will that help? Oct 05 12:05:01 Paul, great, I'm windows, David is Mac Oct 05 12:05:52 Ok. Let me go back and figure out what to do... Oct 05 12:06:10 just apply the patch I linked to earlier Oct 05 12:06:16 then cd FORREST_HOME/main Oct 05 12:06:19 ./build.sh test Oct 05 12:06:41 Ok, would it be bad if test didn't work so well from svn HEAD? Oct 05 12:08:06 we try and keep head working, we *know* this patch breaks it and I am going to bed pretty soon (can't think anymore) Oct 05 12:09:27 Is this a 'site-author' directory patch? Oct 05 12:09:41 Ross i tried again, and get four broken skin/*.css including profile.css Oct 05 12:10:00 sorry, I forgot I was doing it from withn a directoy, it is main/webapp Oct 05 12:10:04 Paul: no it is in main/webapp Oct 05 12:10:38 build clean / build? Oct 05 12:10:47 Just test? Oct 05 12:11:07 Never mind.... Saw it above Oct 05 12:11:54 <-- __Gav_at_work has quit () Oct 05 12:13:36 I got all the way through the cocoon part but with lots of pipeline errors. Oct 05 12:13:47 Shall I post the whole thing? Oct 05 12:13:53 no Oct 05 12:14:01 post it to http://www.pastebin.com - easier to read Oct 05 12:14:16 k. Gimme a minute or so. Oct 05 12:14:53 I'm getting the feeling I've tried to be too clever all at once :-( Oct 05 12:15:11 http://pastebin.com/383376 Oct 05 12:15:12 I've been building up to resources.xmap in all the other simple sitemaps Oct 05 12:15:20 my confidence is too high :-( Oct 05 12:16:00 Paul, all the site: and ext: errors are because of the JXPath issue, so we can ignore them Oct 05 12:16:05 Understood... Oct 05 12:16:29 However, you are getting the same error as David basic.css is broken Oct 05 12:16:53 that is working for me Oct 05 12:17:23 Ross: Do 'svn status' perhaps. Oct 05 12:17:47 done, only locationmap.xml and resources.xml are touched Oct 05 12:17:55 I've not done svn up for a while though... Oct 05 12:18:06 i was just going to suggest that Oct 05 12:18:24 i have ... Last Changed Rev: 294916 Oct 05 12:18:28 by you Oct 05 12:18:58 forrest.xmap was touched in svn up Oct 05 12:19:06 shuoldn't affect it but testing anyway Oct 05 12:19:39 At revision 294964. Oct 05 12:19:48 Should I go back some? Oct 05 12:20:15 No I am 29464 Oct 05 12:20:22 Paul you are looking at the wrong revision number Oct 05 12:20:27 I just tested agin, still got basic.css Oct 05 12:20:32 here comes another patch Oct 05 12:20:39 Mine's SVN overall.. Oct 05 12:20:45 please revert to svn head in preparation Oct 05 12:20:48 that is the revision for the whole apapche.org Subversion Oct 05 12:21:40 Do, I'm blind!!! Oct 05 12:21:43 it changes very often becasue other projects are committing to theirs Oct 05 12:21:55 I missed locationmap-skins.xml Oct 05 12:22:00 cd forrest-trunk Oct 05 12:22:06 svn info Oct 05 12:22:21 and look at Oct 05 12:22:27 Last Changed Rev: Oct 05 12:23:04 Ok 294916, got it. Oct 05 12:24:40 http://issues.apache.org/jira/secure/attachment/12314735/resources.xmap.05_10_2005_01.22UTC.patch Oct 05 12:28:05 okay now 'build test' for me gives only profile.css (other css files are fine now) ... Oct 05 12:28:13 profile.css broken... Oct 05 12:28:22 OK, that's a start Oct 05 12:28:35 I've realised one of the problems here, it is built from an XSLT Oct 05 12:28:37 and i now get broken image urls which i didn't get with your last patch. Oct 05 12:28:56 II'll look at the images issue (they are a real mess) Oct 05 12:29:17 can you look at the profile.xslt issue Oct 05 12:29:29 Ask if you are confused by what I am trying to do Oct 05 12:29:37 will try Oct 05 12:30:03 The only things broken for me are: images.instruction_arrow.png, images.built-with-forrest-button.png and profile.css. Oct 05 12:30:10 original file is at Oct 05 12:30:11 http://svn.apache.org/viewcvs.cgi/forrest/trunk/main/webapp/resources.xmap?rev=294916&view=markup Oct 05 12:30:27 OK, you are the same as us Paul Oct 05 12:30:37 Thought so, just wanted to check. Oct 05 12:31:02 do you want to try and work out the XSLT resolution/rprocessing problem (assuming you understand the lm: resolution method) Oct 05 12:40:57 nailed the images - looking at XSLT now, Oct 05 12:41:12 actually - no, I'm lokking at other stuff that needs doing Oct 05 12:41:23 leaving profile.css to you folk Oct 05 12:41:39 thanks Oct 05 12:41:44 :-) Oct 05 12:41:59 My pleasure, I know you are always looking for something to do ;-) Oct 05 12:46:27 * diwaker is back. Oct 05 12:46:33 guys, I think I just spotted the problem with profile.css Oct 05 12:47:09 It is only looking in the project skin dir, failing to look in defined skin and common skin Oct 05 12:47:59 yes i saw that too and wondered. Oct 05 12:48:06 same issue as I am dealing with for svg.xslt files (testing now) Oct 05 12:48:10 Hi Diwaker Oct 05 12:48:21 hey david Oct 05 12:48:26 howz it goin? Oct 05 12:49:05 fine, i suppose Oct 05 12:49:38 hard to keep up, i still haven't read last night's IRC log Oct 05 12:50:38 i haven't been around at all [sigh] Oct 05 12:50:49 i did manage to catch most of the log up until a few hours back Oct 05 12:54:36 Hey, I'm confused again Oct 05 12:54:44 in resources.xmap we have... Oct 05 12:54:56 ... Oct 05 12:55:02 but where is that resource defined? Oct 05 12:56:32 It is in my resources.xmap Oct 05 12:56:45 line 118 Oct 05 12:56:58 sorry rong Oct 05 12:57:01 So why isn;t it in http://svn.apache.org/viewcvs.cgi/forrest/trunk/main/webapp/resources.xmap?rev=294916&view=markup? Oct 05 12:57:02 wrong Oct 05 13:00:04 it isn't defined! Oct 05 13:00:26 the changes in the lm exposed a bit of the sitemap that was never being reached before Oct 05 13:00:34 since it was surronded by a select Oct 05 13:00:43 I just removed it and it works OK. Oct 05 13:01:24 pesky round corners Oct 05 13:01:32 :-) Oct 05 13:20:50 Hurrah! Oct 05 13:20:58 Found it? Oct 05 13:20:59 I've cracked it , build test passes Oct 05 13:21:04 uploading new patch Oct 05 13:22:50 http://issues.apache.org/jira/secure/attachment/12314736/resources.xmap.05_10_2005_02.21UTC.patch Oct 05 13:23:06 I'm going to bed now, it is way to late for me to be trying to think Oct 05 13:23:22 I'm impressed you're functioning so well as it is. Good night! Oct 05 13:23:31 If anyone can test the resources.xmap patch against a more compelte site than fresh-site Oct 05 13:23:34 that would be greazt Oct 05 13:23:47 thanks for you efforts Ross. Oct 05 13:24:13 good night all, feel free to apply the patch if it works Oct 05 13:24:23 and there are still plenty of xmaps needing work ;-) Oct 05 13:24:54 http://issues.apache.org/jira/browse/FOR-200 has a breakdown of tasks Oct 05 13:32:27 no, 'build test' doesn't work for me. Now all the "round corner" images are not being generated. Oct 05 13:32:41 After build clean ; build. No errors on CSS files. Oct 05 13:32:57 Before clean it generated BROKEN on ALL css files... Oct 05 13:33:14 yes Ross solved that CSS problem a while ago Oct 05 13:34:18 Previous patch gave me BROKEN on some of the CSS files... Oct 05 13:34:42 (The 01_22UTC patch) Oct 05 13:35:37 The round corners are not working... Oct 05 13:36:03 yes, i see, you are just confirming that Ross did fix the CSS issue. Oct 05 13:36:36 Correct. Oct 05 13:37:30 with round corners, Ross must have missed something in giving us that patch. He said it worked for him. Oct 05 13:40:13 Is it possible that, with the advancing hour, he wasn't concentrating on the corners? Oct 05 13:42:02 He said that 'build test' works. Oct 05 13:44:01 Hmmmm... just checking before actually going to bed Oct 05 13:44:09 I can confirm build test works for me Oct 05 13:44:18 I'm checking the patch again Oct 05 13:45:04 can you check your locationmap-skin.xml for Oct 05 13:45:36 hang on no, scratch that Oct 05 13:45:54 ... Oct 05 13:49:19 ...If we see lots of random characters from rgardler, he's fallen asleep at the keyboard... Oct 05 13:49:26 :-) Oct 05 13:49:28 can you check your locationmap-skin.xml for Oct 05 13:49:30 Oct 05 13:50:21 got that Oct 05 13:50:29 hmmmm... Oct 05 13:50:37 It's there... Oct 05 13:50:41 id instruction.png broken as well? Oct 05 13:52:14 it is okay Oct 05 13:52:32 confusing... Oct 05 13:53:01 I think it must be a thing to do with the patches Oct 05 13:53:12 how about I commit and you test there Oct 05 13:53:17 if it fails you can always revert Oct 05 13:53:50 yes do that. As committer, you passed the 'build test'. Oct 05 13:54:05 just checking on site-author Oct 05 13:54:32 forresbot kicks in about 1 minutes' time. Oct 05 13:54:50 we have broken images in site-author Oct 05 13:54:58 so there is obviouly a problem Oct 05 13:55:13 I'll not commit now, I'll look at it tomorrow Oct 05 13:55:20 okay Oct 05 13:55:22 there are other xmaps to be done anyway Oct 05 13:55:48 good night all Oct 05 13:55:54 see ya Oct 05 13:56:05 <-- rgardler (n=chatzill) has left #for-oct Oct 05 14:24:12 * diwaker is away: for dinner Oct 05 14:53:47 --> dwalker (n=diwaker) has joined #for-oct Oct 05 14:55:20 Hi i was just looking at the Jira Roadmap Oct 05 14:55:25 and wondering about tidying up, renaming, classifying some issues Oct 05 14:55:33 ... big job. Oct 05 14:56:32 <-- diwaker has quit (Read error: 104 (Connection reset by peer)) Oct 05 14:56:56 David - I'm about to sign off for the night. Thanks for your help! Oct 05 14:57:09 I rebuilt the views version of the website I maintain and it looks ok. Oct 05 14:57:12 really nice to meet you paul Oct 05 14:57:26 Same here. I hope to be able to partcipate more in the future. Oct 05 14:57:41 that's great, but be prepared for continual change Oct 05 14:58:08 That's why I'm here... I'm trying to catch up on views. Oct 05 14:58:12 any participation is good, do what we can Oct 05 14:58:32 "catch up" ... hah, funny. Oct 05 14:58:36 :) Oct 05 14:59:07 i think that we have your site linked Oct 05 14:59:19 from the Example page, dont't we. Oct 05 14:59:54 I think I have the pelt site in the examples (http://www.cfas.org) Oct 05 15:00:13 The one for views I keep at home (http://cfas-new.kronenpj.dyndns.org) Oct 05 15:00:15 On my quip regarding cocoon caching... That site took 50 minutes to build using 0.8 & views. Oct 05 15:00:21 Ah yeah. Oct 05 15:00:34 With essentially the same content the pelt site takes much less. Oct 05 15:00:35 good success story Oct 05 15:00:40 :) Oct 05 15:01:12 hey this is my worry with current views development ... Oct 05 15:01:37 i get a sneaking feeling that it doesn't use cocoon ideally. Oct 05 15:01:58 I think you're probably right. Oct 05 15:02:24 have you tried the Cocoon Profiler as described for forrest Oct 05 15:02:31 Although I don't think it's specific to views... Oct 05 15:02:33 No, I haven't. Oct 05 15:02:57 getting ref ... Oct 05 15:03:20 http://forrest.apache.org/docs_0_80/howto/howto-dev.html#debug Oct 05 15:04:43 Is this the stuff Ron was working on? Oct 05 15:05:18 (The pelt version takes 5.5 minutes!) Oct 05 15:05:21 Ron started on a grand effort to do other stuff ... Oct 05 15:05:50 this was one of the pieces that we worked out along the way Oct 05 15:06:10 obviously much more to profiling. Oct 05 15:06:54 Cool. Oct 05 15:07:03 I'll take a look at it tomorrow and see what I can find out from it. Oct 05 15:07:08 what do you mean above "my quip regarding cocoon caching" Oct 05 15:07:24 I made a remark about cocoon caching earlier in the day. Oct 05 15:07:37 Mostly that I'd like to see it work. Oct 05 15:07:52 It seems like with views cocoon needs to process every page in the site about 4 times... Oct 05 15:08:32 you'll love cprfile.html then. Oct 05 15:08:36 :> Oct 05 15:08:39 I need to run. Oct 05 15:08:57 Thanks again! Good night. Oct 05 15:09:07 bye Oct 05 15:26:03 * dwalker is back. Oct 05 15:26:18 --- dwalker is now known as diwaker Oct 05 16:52:21 <_Gav_> hello Oct 05 16:52:52 hi Gavin Oct 05 16:53:23 <_Gav_> hows things Oct 05 16:55:17 i am trying to do one of the locationmap tidy up of sitemaps like Ross has been doing Oct 05 16:55:53 <_Gav_> ok, I started one last night , Im doing linkmap.xmap Oct 05 16:56:06 <_Gav_> hopefully not the same one your doing :) Oct 05 16:56:13 i am doing search.xmap Oct 05 16:57:10 <_Gav_> I started making changes, will let you know when I test it if it works or not, in fact I should not see a difference, it should just work shouldn't it ? Oct 05 16:57:38 yep, i have been testing mine each step of the way Oct 05 16:57:51 <_Gav_> ok Oct 05 16:58:10 <_Gav_> in site-author ? Oct 05 16:59:14 will do that later, for now i am doing in a fresh "seed-sample" ... Oct 05 16:59:27 <_Gav_> ok Oct 05 16:59:37 just because it is quicker to index for the Lucene search. Oct 05 17:22:33 <-- kronenpj has quit ("Hey! Where'd my controlling terminal go?") Oct 05 17:44:02 <_Gav_> xley: you there Oct 05 17:44:18 yeah Oct 05 17:44:24 <_Gav_> Does : Oct 05 17:44:40 <_Gav_> become : Oct 05 17:45:37 Ross has already done some of those ... Oct 05 17:45:45 cd main/webapp Oct 05 17:45:58 grep absolutize *.* Oct 05 17:46:36 <_Gav_> ok thanks, will look, the one Im in is linkmap.xmap he left me to play with. Oct 05 17:48:11 The search.xmap is done ... was reasonably easy. Oct 05 17:48:33 but still takes lots time. Oct 05 17:49:34 <_Gav_> ok good, yes it does. Ross used linkmap-absolutised for absolutize-linkmap , is he Irish Oct 05 17:50:24 I gather that that was so that it followed the naming convention. Oct 05 17:50:57 i.e. transform from this to that. Oct 05 17:51:24 <_Gav_> ok Oct 05 17:51:50 <_Gav_> so I see document would be wrong then Oct 05 17:52:54 ??? Oct 05 17:52:56 <_Gav_> but the same match pattern applies : Oct 05 17:54:01 <_Gav_> I originally put linkmap to document when it should of been linkmap to absolutize-linkmap Oct 05 18:00:34 I found a bug with our lucene search while testing that search.xmap so i am adding that to Jira Oct 05 18:00:55 then probably call it quits for today Oct 05 18:07:48 <_Gav_> Ok, I have made changes to linkmap.xmap Oct 05 18:08:07 <_Gav_> It still works ok, but I can not see what else I need to do ( http://pastebin.com/383553 ) Oct 05 18:09:06 Would you add a patch to Jira please. Ross already has an issue open there to receive it. Oct 05 18:09:26 <_Gav_> Ok, will do Oct 05 18:14:06 <_Gav_> ok done. Oct 05 18:15:21 <_Gav_> kick me when your ready :) Oct 05 18:17:12 <-- diwaker has quit (Read error: 110 (Connection timed out)) Oct 05 18:18:35 thanks, will apply your patch. Thanks for your help today. Hope it was fun. Oct 05 18:19:13 <_Gav_> Very informative thanks. I know a little bit more each time and hopefully pennies will drop from time to time. Oct 05 18:19:57 until it changes again :-) But yes each of us learns more. Oct 05 18:20:08 I did. Oct 05 18:20:31 <_Gav_> I think the current changes are definatly for the better. Oct 05 18:21:09 some of the proposed changes i am not so sure. Oct 05 18:21:28 anyway we will need to discuss those on list Oct 05 18:21:37 <_Gav_> I will carry on after this session ends and keep looking to see what else there is Oct 05 18:21:38 <_Gav_> yes Oct 05 18:22:56 your locationmap.xmap changes were too easy. The revisions.xmap should be okay. Oct 05 18:22:58 <_Gav_> Im going to do FOR:682 now anyway, that looks at my current level :) Oct 05 18:23:04 i mean easy. Oct 05 18:23:15 Ross has a handle on the rest i think. Oct 05 18:23:35 <_Gav_> Ok, I'll leave the rest then. Oct 05 18:24:15 good luck with FOR-682 Oct 05 18:24:43 revisions.xmap if you want to Oct 05 18:25:00 <_Gav_> ok, I'll take a look at that now then. Oct 05 18:25:14 <_Gav_> and do for-682 later Oct 05 18:25:34 all good. Oct 05 18:27:25 <_Gav_> I got an error on svn up - tells me to do svn cleanup - is that my end Oct 05 18:27:59 yes, i think so Oct 05 18:28:16 do 'svn status' first Oct 05 18:28:32 <_Gav_> ok, wouldnt want to add any breakages Oct 05 18:30:00 <_Gav_> tools\targets is locked Oct 05 18:30:31 <_Gav_> what status is ? Oct 05 18:30:59 that is strange ... locked ... you were not touching that. Oct 05 18:31:24 The "?" means extra file which is not under revision control Oct 05 18:31:24 <_Gav_> Thats what I thought, never been there in my life Oct 05 18:31:34 <_Gav_> ok Oct 05 18:31:46 Just do 'svn cleanup' Oct 05 18:34:14 woops your patch to linkmap.xmap was broken (missing trailing "}") ... Oct 05 18:34:28 you didn't do 'build test' eh. Oct 05 18:36:13 i fixed it this side and doing some more tests on it now. Oct 05 18:37:05 <_Gav_> sorry Oct 05 18:37:30 no worries. I did the same error stacks of times myself. Oct 05 18:37:40 <_Gav_> I did forrest seed-sample then forrest run on it though Oct 05 18:37:59 write it out 100 times, "always do build test" :-) Oct 05 18:38:06 <_Gav_> hehe :) Oct 05 18:38:45 Hmmm, but it shouldn't have worked on anything, with that error. Oct 05 18:39:42 <_Gav_> actually, I got that error then fixed it then ran it ok, diff got mixed up somehow I dunno Oct 05 18:39:47 <_Gav_> no excuses Oct 05 18:39:52 :-) Oct 05 18:40:14 How did your 'svn cleanup' go? Oct 05 18:40:41 <_Gav_> just finished ok (took a while) , svn up now on 295011 Oct 05 18:40:58 <_Gav_> can continue on now Oct 05 18:43:58 okay, usually SVN is extremely reliable, must have been something your side, dunno what. Oct 05 18:44:19 applied your patch thanks. Oct 05 18:44:38 <_Gav_> I still get asked to set FORREST_HOME each time I build, it is in path already Oct 05 18:45:24 FORREST_HOME and PATH are two different things. Oct 05 18:46:43 <_Gav_> I did both, warning says to add %FORREST_HOME%\bin to path Oct 05 18:47:18 Ah, Windoze ... i cannot help sorry. Oct 05 18:48:01 <_Gav_> No probs, I add FORREST_HOME=c:\apache2\forrest to user variables, I'll try adding it to system variables Oct 05 18:48:16 <_Gav_> it works either way Oct 05 18:49:15 okay time to shut down the IRC. See you on list sometime. Oct 05 18:49:55 <_Gav_> ok, bye Oct 05 18:50:00 bye Oct 05 18:50:13 <-- _Gav_ has quit () Oct 05 18:51:17 <-- xley has quit () Oct 05 18:51:41 <-- xley2 has kicked cheche from #for-oct (closing channel) Oct 05 18:52:14 Bad arguments for user command. Oct 05 18:52:26 --- You have left channel #for-oct ("Client Exiting") **** ENDING LOGGING AT Wed Oct 5 18:52:26 2005