Hopbot log for 2009-03-02 - Helma IRC channel: #helma on irc.freenode.net

2009-03-02:

[13:51] <simono> hi hannes, i'm unhappy about core/JSON . it doesn't behave as it should when handling undefined / NaN values. I'll file a bug report, but I'm note sure this is something you want to change - lots of apps, ours including, rely on the current, wrong behaviour
[13:55] <simono> no wait I'm confused, seems jquery's json-plugin doesn't do what i think it should.. i'll investigate
[14:03] <hannesw> simono: is this in helma 1 or ng?
[14:04] <simono> helma1 ... but it's not a helma problem, sorry. seems jquery is doing smth stupid. but while I'm at it: i've seen there are a new json-module out from crockford.. handling unicode and dates in particular, any plans on including one of those new versions in helma1?
[14:40] <simono> nvm, i got it working. the json-jquery-plugin returns "undefined" (string) for undefined-values instead of dropping the attribute completly and helma1 can (correctly) not parse that non-JSON
[15:08] <zumbrunn> philipp: are you the philipp that added the for/in loop and global.path related info to the Reference Errata page?
[15:09] <hannesw> simono: if this is json2.js, it's already in helma ng
[15:10] * zumbrunn doesn't see how the example using a for loop is wrong
[15:10] <hannesw> not sure it does something special about unicode, though
[15:10] <hannesw> unicode shouldn't be a problem on the jvm anyway
[15:11] <zumbrunn> and if it is wrong, I would like to add language that warns about using for loops in this context
[15:11] <zumbrunn> instead of just replacing it with an example that uses for/in
[15:12] <philipp> no I don't think so
[15:14] <philipp> I just joined to see the irc chat the first time ;)
[15:14] <zumbrunn> ok, welcome :-)
[15:15] <philipp> hi
[15:15] <hannesw> hi philipp!
[15:15] <philipp> where is this page located?
[15:15] <philipp> hi hannes
[15:15] <zumbrunn> for what its worth, this is the errata page I was talking about:
[15:15] <zumbrunn> http://dev.helma.org/wiki/Helma+Reference+Errata/
[15:15] <hannesw> you were jabbering me about some java related problem?
[15:17] <philipp> yes I was wondering if this is the new helma-dev group
[15:18] <hannesw> what is the new helma-dev group?
[15:18] <philipp> where I can post Problems that I encounter using helma in eclipse
[15:18] <philipp> so I give it a try and to see whats discussed here
[15:19] <philipp> I thought you said sometimes to me that the old mailman groups are dead
[15:20] <philipp> writing english is harder than I thougt ;)
[15:23] <hannesw> the helma 1 list is at http://groups.google.com/group/helma
[15:24] <philipp> ok for dev and user questions
[15:24] <philipp> maybe I am a user
[15:24] <hannesw> anything helma 1 related
[15:25] <philipp> thanks, than I will post it there so I don't disturb things here
[15:25] <philipp> or per direct chat with you hannes
[15:26] <philipp> so just a follower now
[15:30] <simono> hannesw, @unicode: javascript's charCodeAt() only works as expected for basic multilanguage plane, mozilla-wiki told me so. this can lead to JS dropping characters or interpreting them as linefeeds.
[15:32] <hannesw> simono: i don't know about this, but unless it's spec'ed that way, it maight be a spidermonkey problem only
[15:34] * simono checks the spec to see if crockford is wrong :)
[15:35] <hannesw> no, i don't mean anybody is wrong, i just mean it may be an implementation issue.
[15:36] <hannesw> but if crockford says it's intrinsic to js, it may very well be
[15:43] <simono> actually found it in the spec: charCodeAt must return non-neg number < 2^`6. higher code points are encoded with surrogates. been reading up on those, since we want to make the utf8-switch next month
[15:45] <hannesw> what are you switching to utf-8? your db?
[15:46] <simono> ah yes DB
[15:46] <hannesw> inside helma/jvm, everything is unicode, so you shouln't not any difference
[15:48] <simono> thanks for the reassuring words :)
[15:49] <hannesw> well, it's true. helma never notices what encoding your db uses. the only part that may notice is the jdbc driver
[15:49] <simono> yes, that's how i understand it. there are other things one can run into... like this charCodeAt, we had that in a forum-check, or how string.length doesn't tell you the byte-length anymore :)
[15:55] <simono> also now that we are talking char encodings, some time ago we had a problem with utf8 encoded post-requests not setting the req.getServletRequest().getCharacterEncoding() when behind ajp. i think we worked around it like that http://helma.pastebin.com/d4af189e6 and never told u about it :)
[15:55] <simono> well we definitly worked around it like that, not sure wether someone ever told u
[15:57] <hannesw> simono: could you provide a diff for that?
[15:58] <hannesw> oh, it's in your app code, right?
[15:58] <simono> we solved it in the app, didn't change helma. when there is no charset in req, we assume it's utf8
[15:58] <hannesw> ok
[15:59] <simono> obscure bug, but I should put it in tracker
[16:06] * zumbrunn sees it was maks that added the for/in global.path related errata, not phillipp
[16:20] <simono> zumbrunn, do you know what he meant? both - his and the one in docs - work and i favor the doc-version
[16:21] <zumbrunn> no, I don't, that's why I asked
[16:21] <zumbrunn> I get integers for keys both ways in my tests
[16:22] <zumbrunn> maybe it somehow depends on the type.properties settings
[16:23] <simono> hm, maybe b/c of path["story"] ...
[16:23] <simono> i mean path[PROTONAME] works and path[path.length-1] works too, so that might be confusing
[16:58] <botic> Hi! I've the following scenario: 2 skins, one is the embedded one and the other is the basic html-skin. at the bottom of the surrounding skin there is something like <% request.title %>, which is by by default "Home", if there is no get-param set. if the embedded skin is used, i want to modify the get-param "title" inside of the skin. Is that possible?
[17:01] <simono> hm, can u show us the code that renders the skins?
[17:08] <botic> maybe i should post that to the mailing list to provide you more info. ;-)

 

 

In the channel now:

Logs by date: