Wednesday September 03, 2003
|
Vanity Foul Dedicated to the wanderings of an egotistical mind. |
|
Circle Blog Cameron posts on Anthony blogging about Cedric's response to Joe's column. And now I will blog about Cameron, by pointing to my comment at Cedric's:
Every time someone has pointed at something wrong with Roller, we've said "yup, you may be right". We looked at it some more and said "yup, you were right". And then we did our level best to fix it. I'm sorry FreeRoller's performance has been so disappointing. The chief underlying cause has been addressed, and Roller 0.9.8 performs much better (under artificial load conditions anyway). We continue working to make Roller faster, and to make Roller better. Some people see Roller as a product, me, I see it as a hobby. Everybody wants to be better at their hobby. Forgotten Blogiversary Thanks to Niel Eyde for reminding me, I'd forgotten my anniversary as well. Maybe not forgotten, just had more pressing concerns (I have thought about posting this post before - just not when I was near a computer). In any case, blogging and Roller saved my sanity one year and 13 days ago.
Atom 4 U, Atom 4 Me, Atom4J Released I've joined the OSJava.org group (thanks guys) and released the first version of Atom4J. The "spec" is nowhere near complete, so expect a great deal of flux in Atom4j as well. As a part of my efforts, Atom4J includes an abstract Servlet upon which specific implementations can be built upon (thus enforcing the API). I should have a version for Roller checked in tonight.
Atom notes I'm nearly done with Atom4J, just doing some final cleanup and reviewing the existing documentation to make sure I'm in compliance with the latest 'spec'. A couple things I've noticed: 1) You must carefully read all the sources of information. The main wiki[1] contains some information (but this tends to be a moving target), Joe Gregorio has a "Busy Developers Guide"[2] and a full API spec[3], there is a new wiki for the API[4], and there is the Feed spec(s) located at Mark's place[5]. 2) I've had to "extend" the API a bit. The Introspection results don't specify how to retrieve the FindTemplates results - rather the edit-template element appears to specify the URI for editing an individual template. In Atom4J the edit-template URI now points to the FindTemplates service, where each template's URI is presented for editing. I've also adding "template" to the DELETE verb to allow for deleting templates. Still missing is a facility for creating templates (the POST verb). 3) Finally, it's hard to find the right place to bring up issues or point out inconsistencies. A wiki can be a wild-and-woolly place, and for a relative newcomer like myself (long-time lurker) it is intimidating. And so I've placed my thoughts here, in hopes they will filter into the main discussion in true Weblog style. Here are the URLs mentioned above ( Aug 30 2003, 08:18:51 AM ) Atom Permalink [Trackback] Re: Atom and Digester In between trips to the hospital, and sleeping, I had a little time to work on resolving my issue with Digester parsing Atom XML. As suggested in my comments and in the Digester-user list, I implemented a custom Rule and Handler. It took quite a bit of trial-and-error, but I got it working. Look for real results soon. It's a boy Baby Boy Lavandowska was born at 7:29 PM Central Time. He weighed 7 lbs, and is 21 inches long. After nearly 12 hours of contractions, it only took 20 minutes of pushing for him to make his appearance (he beat both his doctor and the in-house doctor). His nurse had only just come on duty, and she did a great job! Mother and Son are doing well, and no, we haven't settled on a name yet. ( Aug 26 2003, 10:06:16 PM ) Personal Permalink Comments [11] [Trackback] Atom and XOM Digester wasn't cutting it (haven't received that custom Rule yet), so I thought I'd try XOM. Well, XOM thinks the example Entry isn't WellFormed: I don't know what isn't well-formed about that, but since <content> could contain "nasty" non-X HTML, wellformedness shouldn't be a requirement (damn users, can never count on them to do the right thing!). My search for another solution will continue tonight (Jason, wanna send me that package your mentioned in my comments?). Good night, and thanks for your support. Atom and Digester As I mentioned in my previous post, I'm lazy. So I'm using Digester to parse the AtomAPI XML - but its falling down in one regard: <content>can contain <html></content>. And Digester sees this as just more XML, so the contents on the content tag only get assigned if there is no html. Some general websearching, and searching on the Digester-user list, revealed one other person with a similar concern earlier this year. I've sent him an email, asking if he can supply more information on his solution. If not, or if that doesn't solve the issue, I'm afraid I'll have to resort to "hand-parsing" the AtomAPI. I'm lazy, I don't wanna do that. But again, it may become necessary as the AtomAPI allows for adding arbitrary vendor-specific tags - <mt:fooBar> for example - and my Digester rules don't cover that at the moment. I do plan on re-expressing the rules via Digester's rules.xml capability though; that may allow for the flexibility necessary.... |
|
||||||||