You are viewing a plain text version of this content. The canonical link for it is here.
Posted to java-dev@axis.apache.org by Aleksander Slominski <as...@cs.indiana.edu> on 2004/10/06 16:29:27 UTC
[Axis2] IRC chat log 2004-10-06
summary: today discussion continued on hot deployment (checking
how tomcat is doing it and working on service.xml schema)
then looking on current OM API prototypes, differences between them,
and next steps.
thanks,
alek
[10/6/2004 8:02 AM] =-= Topic for #apache-axis is ``apache-axis''
[10/6/2004 8:02 AM] =-= Topic for #apache-axis was set by Srinath_ on Wednesday, September 29, 2004
9:06:50 AM
[10/6/2004 8:02 AM] <Deepal> Hi all
[10/6/2004 8:03 AM] =-= YOU are now known as alek_logging
[10/6/2004 8:03 AM] <alek_logging> hello
[10/6/2004 8:03 AM] <Deepal> it is time to start
[10/6/2004 8:04 AM] <alek_logging> what is today agenda?
[10/6/2004 8:04 AM] -->| pzf (~paul@blueice2n1.uk.ibm.com) has joined #apache-axis
[10/6/2004 8:04 AM] <pzf> hi folks
[10/6/2004 8:04 AM] -->| Jaliya (~Miranda@220.247.240.27) has joined #apache-axis
[10/6/2004 8:04 AM] <pzf> i hear its going well
[10/6/2004 8:05 AM] <Ajith> hi all
[10/6/2004 8:05 AM] <Jaliya> Hi All
[10/6/2004 8:05 AM] <Srinath> hi all :)
[10/6/2004 8:05 AM] -->| chathura (~chathurah@203.94.95.24) has joined #apache-axis
[10/6/2004 8:06 AM] <vakYpollo> hi
[10/6/2004 8:06 AM] <Srinath> where are we from deployment stuff
[10/6/2004 8:06 AM] <Srinath> I think that is where we stop
[10/6/2004 8:06 AM] <chathura> hi guy
[10/6/2004 8:06 AM] =-= Srinath has changed the topic to ``axis2 deployment''
[10/6/2004 8:08 AM] -->| Chinthaka (~chinthaka@220.247.240.27) has joined #apache-axis
[10/6/2004 8:08 AM] <Chinthaka> hi all
[10/6/2004 8:09 AM] <Srinath> BTW Ias what happen to ur name :D
[10/6/2004 8:10 AM] -->| sanjiva_away (~sanjiva@203.94.84.117) has joined #apache-axis
[10/6/2004 8:10 AM] <sanjiva_away> hi guys sorry I'm late
[10/6/2004 8:10 AM] =-= sanjiva_away is now known as sanjiva
[10/6/2004 8:10 AM] <sanjiva> I was spsed to have sorted out the details of service.xml by now .. but I
messed up :-( ...
[10/6/2004 8:11 AM] <Deepal> I developed a prototye for hot deployemnt
[10/6/2004 8:11 AM] <sanjiva> cool!
[10/6/2004 8:11 AM] <Deepal> and sSrinath will commit it soon
[10/6/2004 8:11 AM] <Srinath> I going check it in to scratch
[10/6/2004 8:11 AM] <pzf> excellent
[10/6/2004 8:12 AM] <Deepal> BTW someone know how tomact handle hat deployement ?
[10/6/2004 8:12 AM] <sanjiva> hat?
[10/6/2004 8:12 AM] -->| Harsha (Harsha@CPE-147-10-37-44.vic.bigpond.net.au) has joined #apache-axis
[10/6/2004 8:12 AM] <Deepal> sorry HOt
[10/6/2004 8:12 AM] <pzf> i know how topcat manages hat deployment :-)
[10/6/2004 8:13 AM] <Deepal> so tell me
[10/6/2004 8:13 AM] <pzf> (that was a joke... sorry)
[10/6/2004 8:13 AM] <Chinthaka> :)
[10/6/2004 8:13 AM] <Srinath> Hi Harsha
[10/6/2004 8:13 AM] <pzf> http://www.dfcom.freeserve.co.uk/hbw/topcat/
[10/6/2004 8:14 AM] <chathura> by the way sir the wsdl object model code is also done
[10/6/2004 8:14 AM] <Harsha> Hi Everybody
[10/6/2004 8:14 AM] <chathura> srinath may be you can check in my wsdl object model too
[10/6/2004 8:14 AM] <Deepal> Hi Harsha
[10/6/2004 8:14 AM] <pzf> excellent
[10/6/2004 8:15 AM] <Harsha> I got another person here as well. He is Asanka jayawardena.
[10/6/2004 8:15 AM] <pzf> btw way we are working on how to move forward with JWSDL2.0
[10/6/2004 8:15 AM] <sanjiva> Chathura: great!
[10/6/2004 8:16 AM] <sanjiva> Harsha and Asanka - welcome
[10/6/2004 8:16 AM] <sanjiva> Guys these two folks are from Melbourne Australia .. they want to join
this project and contribute!
[10/6/2004 8:16 AM] <Srinath> :)
[10/6/2004 8:16 AM] <pzf> cool
[10/6/2004 8:16 AM] <chathura> welcome
[10/6/2004 8:16 AM] <Harsha> Thanks for welcoming us.
[10/6/2004 8:17 AM] <Ajith> always welcome
[10/6/2004 8:17 AM] <sanjiva> So Deepal will you give a summary of the hot deployment stuff please?
[10/6/2004 8:17 AM] <Deepal> Sir there is a thread runninf forever
[10/6/2004 8:17 AM] <Ajith> BTW most of the Axis Table Model is done and I checked it in to the SVN 30
mins ago
[10/6/2004 8:18 AM] <Deepal> and checkig whether a spcific folder is update
[10/6/2004 8:18 AM] <Jaliya> How you compaire
[10/6/2004 8:18 AM] <Deepal> meaning whether new jars added to taht folder
[10/6/2004 8:18 AM] <chathura> based on time stamp is it??
[10/6/2004 8:18 AM] <Ajith> right now I am writing an XML bench for that one
[10/6/2004 8:18 AM] <Chinthaka> hope u all had time to check the OM api we proposed
[10/6/2004 8:19 AM] <Deepal> yep Chathura
[10/6/2004 8:19 AM] <samisa> time stamp of what?
[10/6/2004 8:19 AM] <Chinthaka> and myself and Ajith implemented two prototypes of those APIs
[10/6/2004 8:20 AM] <Deepal> Im checking modified date of the jar files
[10/6/2004 8:20 AM] <samisa> ok
[10/6/2004 8:20 AM] <Deepal> not the content of that jar
[10/6/2004 8:20 AM] <Deepal> if it is modified only I will check for its content
[10/6/2004 8:20 AM] <Jaliya> How about the new additions>
[10/6/2004 8:20 AM] <chathura> so this is extensible for hot update too is it??(or am i too ambitious)
[10/6/2004 8:21 AM] <Deepal> yep
[10/6/2004 8:21 AM] <Deepal> in this prototype it will only check newlly added jars
[10/6/2004 8:21 AM] <Srinath> I think time stamp is enough
[10/6/2004 8:21 AM] <sanjiva> guys can we do one topic at a time please
[10/6/2004 8:21 AM] <sanjiva> let's finish the hot deployment stuff first
[10/6/2004 8:22 AM] <Deepal> I developed the thread as a scheduler
[10/6/2004 8:23 AM] <Deepal> and it will do it task for ever 10 s
[10/6/2004 8:23 AM] <samisa> Is it the same strategy used by Tomcat, JBoss etc?
[10/6/2004 8:23 AM] <Deepal> that is the cycle is 10s
[10/6/2004 8:23 AM] <alek_logging> Chinthaka: please od not put Word documents into SVN
[10/6/2004 8:23 AM] <alek_logging> Chinthaka: i think this kind of discussion/documentaiton belongs in
Wiki ...
[10/6/2004 8:23 AM] <Chinthaka> ok
[10/6/2004 8:23 AM] <Chinthaka> sorry abt that
[10/6/2004 8:24 AM] <Srinath> BTW code of deepal at
http://svn.apache.org/repos/asf/webservices/axis/trunk/java/dev/scratch/deepal/deployment/
[10/6/2004 8:24 AM] <Deepal> Samias :taht is whay I asked how tomcat handle hot deployemnt
[10/6/2004 8:24 AM] <Deepal> and I send a mail to tomcat mailing list today
[10/6/2004 8:24 AM] <Deepal> to know abt that
[10/6/2004 8:24 AM] <samisa> Ok great
[10/6/2004 8:24 AM] <alek_logging> Chinthaka: when you major chnages in Wiki please post short email
toindicate where is Wiki page to read ...
[10/6/2004 8:25 AM] <Chinthaka> alek : I think I posted that to dev list, last time
[10/6/2004 8:26 AM] <Deepal> to go a head with this I need to finalize the folder structure
[10/6/2004 8:26 AM] <Deepal> and the schemas
[10/6/2004 8:27 AM] <Srinath> shall we create schemas ASAP and put them to SVN so others can comment
[10/6/2004 8:27 AM] <Deepal> the folder structure that is shown in the wiki
[10/6/2004 8:27 AM] <chathura> think the rest of the deployment guys hav the same problem
[10/6/2004 8:27 AM] <Deepal> is correct ?
[10/6/2004 8:27 AM] <alek_logging> Ajith: when you have benchmark code written please post link to SVN
and maybe wiki page descirbing how to run it ...
[10/6/2004 8:27 AM] <Ajith> ok
[10/6/2004 8:28 AM] <Ajith> I was just listening since the "hot" topic is hot deployment :)
[10/6/2004 8:28 AM] <Deepal> http://wiki.apache.org/ws/FrontPage/Architecture/Deployment
[10/6/2004 8:28 AM] <sanjiva> Yes .. I'll work on service.xml schema tomorrow and we can discuss it by
email .. if we agree on a version of that then that can be the starting point for hot deployment right?
[10/6/2004 8:29 AM] <Deepal> not only hot deployemnt
[10/6/2004 8:29 AM] <alek_logging> Chinthaka: i am sorry but i have hard time commenting on Word
document - i think the content of this Word DOC should have been posted as email - i see no reason why
it was not done ...
[10/6/2004 8:29 AM] <Deepal> bt also for Deployemnt :)
[10/6/2004 8:30 AM] <Chinthaka> alek : I put the word doc, as it is , in the wiki
[10/6/2004 8:30 AM] <sanjiva> Alek: Wiki is a pain in the ass to keep editing and iterating on
[10/6/2004 8:30 AM] <sanjiva> Please read it and send comments to the list or wiki!
[10/6/2004 8:30 AM] <alek_logging> Sanjiva: do you sugest we use Word DOC to document java code and to
do discussion?
[10/6/2004 8:30 AM] <Chinthaka> alek : and I put that word doc in the svn, for some one to easily look
in to that, once he checks out the code from that place
[10/6/2004 8:31 AM] <sanjiva> Chintaka - can u PDF it and put it? Or better yet .. commit an html
version of it into the SVN
[10/6/2004 8:31 AM] <alek_logging> Sanjiva: if you looked on the mentioned Word DOC you can see there
is no reason why Word was used - it is just text ...
[10/6/2004 8:31 AM] <Ajith> hmmmmm
[10/6/2004 8:31 AM] <Chinthaka> ok, consider it done
[10/6/2004 8:31 AM] <Ajith> or just use deepals doc to wiki converter :)))
[10/6/2004 8:31 AM] <Srinath> html version is better
[10/6/2004 8:31 AM] <Srinath> :D
[10/6/2004 8:31 AM] <chathura> :d
[10/6/2004 8:32 AM] <Deepal> its not mine
[10/6/2004 8:32 AM] <sanjiva> Its just an editor dude ... and it gives room to insert a figure if one
wants to. Anyway we can convert to HTML and post it - that works for all right?
[10/6/2004 8:32 AM] <alek_logging> Sanjiva: check
http://svn.apache.org/repos/asf/webservices/axis/trunk/java/dev/scratch/chinthaka_ajith/OmtableModelAPI/
[10/6/2004 8:32 AM] <Deepal> its propert of LSF :)
[10/6/2004 8:32 AM] <Ajith> :)
[10/6/2004 8:32 AM] <alek_logging> i think that HTML version is btter too
[10/6/2004 8:32 AM] <samisa> +1 for HTML
[10/6/2004 8:33 AM] <Srinath> does all happy with html?
[10/6/2004 8:33 AM] <Ajith> ok may it be HTML then
[10/6/2004 8:33 AM] <sanjiva> Alek: saw it .. that doc is just a report right? the real meat is in the code
[10/6/2004 8:33 AM] <alek_logging> anyway for discussion of such important topic as "decoupling of DOM
API from OM Model" i would prefere to use email thread ...
[10/6/2004 8:34 AM] <sanjiva> So shall we switch topics to OM now?
[10/6/2004 8:34 AM] <alek_logging> i am not sure what it is - it looks to me like status update /
starting point for discscussion
[10/6/2004 8:34 AM] <sanjiva> Are we done with deployment?
[10/6/2004 8:35 AM] <sanjiva> I guess until we have the schemas we can't do a whole lot more on
deployment .. my apologies for not delivering this week
[10/6/2004 8:35 AM] <alek_logging> Chinthaka/Ajith brought the topic so i had to respond
[10/6/2004 8:35 AM] <Srinath> Shall we switch to OM :)
[10/6/2004 8:36 AM] <Srinath> we hav three Impl I belive ..
[10/6/2004 8:36 AM] <Deepal> its ok sir
[10/6/2004 8:36 AM] <sanjiva> OK let's talk about OM :)
[10/6/2004 8:36 AM] <sanjiva> What are the 3 impls?
[10/6/2004 8:37 AM] <Ajith> 1. the table model with rows
[10/6/2004 8:37 AM] <Ajith> 2. the link list model
[10/6/2004 8:37 AM] <Ajith> 3. the table model with columns
[10/6/2004 8:38 AM] <Srinath> the diff #1 #3 is based on how table is implemented
[10/6/2004 8:38 AM] <Srinath> what is diff #1 vs #2
[10/6/2004 8:38 AM] <Ajith> 1st one is kind of abandoned now
[10/6/2004 8:38 AM] <Ajith> since the 1st one actually converted to the 2nd while implementing
[10/6/2004 8:38 AM] <Srinath> shall we drow the concepts without saying 1, 2
[10/6/2004 8:39 AM] <Ajith> :)
[10/6/2004 8:39 AM] <Srinath> now everybody get it that way I think
[10/6/2004 8:39 AM] <Ajith> ok here are the conceptual differences
[10/6/2004 8:39 AM] <Srinath> :)
[10/6/2004 8:40 AM] <Chinthaka> I just posted that doc file in html format to, axis-dev
[10/6/2004 8:40 AM] <Ajith> the link list model has objects (plain old "shim dom" objects) linked by
obejct refeences in the fashion of a link list
[10/6/2004 8:40 AM] <Chinthaka> can someone please apply that patch for me
[10/6/2004 8:41 AM] <Srinath> sure about patch.. back to OM :)
[10/6/2004 8:41 AM] <Ajith> the column table model stroes everything in either integer or string mode
in arrays and gives the object by filling in the values when needed
[10/6/2004 8:42 AM] <Ajith> more like the XALAN DTM i guess
[10/6/2004 8:42 AM] <sanjiva> So can both of these implement the API that Chinthaka posted?
[10/6/2004 8:42 AM] <Chinthaka> both those models have the common api
[10/6/2004 8:42 AM] <Chinthaka> yeah, they are already being implemeted
[10/6/2004 8:43 AM] <Ajith> note however that both support incremental building through pull parsing
[10/6/2004 8:43 AM] <Ajith> yeah
[10/6/2004 8:43 AM] <Chinthaka> I will commit that linked list model to svn tomorrow
[10/6/2004 8:43 AM] <sanjiva> So maybe we should examine the interfaces and not the impls right now then?
[10/6/2004 8:43 AM] <Ajith> so that we can run the same perf test on both
[10/6/2004 8:43 AM] <Chinthaka> thats correct
[10/6/2004 8:43 AM] <Ajith> yeah
[10/6/2004 8:45 AM] <pzf> doesn't the table model require the schema in advance?
[10/6/2004 8:45 AM] <alek_logging> so here my two cents about API: why not use String directly as TEXT
nodes? this wil make API smaller and more lightweight
[10/6/2004 8:45 AM] <sanjiva> ok
[10/6/2004 8:45 AM] <alek_logging> XML Infoset suggest to have [element content whitespace] but i do
not think we need this info for SOAP ...
[10/6/2004 8:45 AM] <alek_logging> http://www.w3.org/TR/xml-infoset/#infoitem.character
[10/6/2004 8:46 AM] <alek_logging> and for sure we can live without backlink to [parent] ...
[10/6/2004 8:46 AM] <alek_logging> when DOM API is needed it can wrap String into TEXT nodes - so
priice for using DOM is explicit
[10/6/2004 8:47 AM] <Chinthaka> the link list has backlink to parent
[10/6/2004 8:47 AM] <Chinthaka> every node knows its parent
[10/6/2004 8:47 AM] <Srinath> pzf: no the om is just like dom with streaming
[10/6/2004 8:47 AM] <alek_logging> i think that for SOAP infoset API (OM) in case of Character
Information Item [parent[ is not needed ...
[10/6/2004 8:48 AM] <Ajith> yeah linking can be aprob if the text nodes are just String
[10/6/2004 8:48 AM] <alek_logging> i think using Strings directly will make API impls much smaller on
memory footprint ...
[10/6/2004 8:48 AM] <Srinath> Alex:but I belive we decided we need full infoset two weeks ago
[10/6/2004 8:49 AM] <alek_logging> we take and pick things from XML Infoset
[10/6/2004 8:49 AM] -->| Essington (~Essington@essington.user) has joined #apache-axis
[10/6/2004 8:50 AM] <Srinath> alek do not get what you mean ?
[10/6/2004 8:50 AM] <Deepal> i didnt get u alex
[10/6/2004 8:50 AM] <alek_logging> i do not think we have in API Document Type Notatoin, Unparsed
Entity Information, or Declaration Information Items ...
[10/6/2004 8:50 AM] <Chinthaka> alek : no we don't have
[10/6/2004 8:51 AM] <Srinath> yes that three is out
[10/6/2004 8:51 AM] <Deepal> for the sopa we dont need Doc tyep notation
[10/6/2004 8:51 AM] <Srinath> but whitespaces we need 4 secuirty
[10/6/2004 8:51 AM] <alek_logging> currently proposed OM API does not support full XML Infoset !
[10/6/2004 8:52 AM] <alek_logging> you are ocnfusing whitespace characters with boolean flag [element
content whitespace]
[10/6/2004 8:52 AM] <Deepal> do we need to support fulinforset ?
[10/6/2004 8:52 AM] <alek_logging> (actually it is not boolelan it is tr-state)
[10/6/2004 8:52 AM] <Srinath> Chinthka: what does OM API support
[10/6/2004 8:52 AM] <Srinath> can you explain bit
[10/6/2004 8:52 AM] <Chinthaka> OM has the infoset that is required by SOAP only
[10/6/2004 8:52 AM] <Ajith> basically elements, attribs ,text, comment and cdata
[10/6/2004 8:53 AM] <alek_logging> (tri-state: true, false, or "unknown" - or maybe more states ...)
[10/6/2004 8:53 AM] <Srinath> what is it:)
[10/6/2004 8:53 AM] <Jaliya> Alek please explian a bit
[10/6/2004 8:53 AM] <alek_logging> CDATA is not in XML Infoset ...
[10/6/2004 8:53 AM] <Deepal> I think it is
[10/6/2004 8:53 AM] <alek_logging> see Appendix D list item 19: http://www.w3.org/TR/xml-infoset/#omitted
[10/6/2004 8:54 AM] <alek_logging> CDATA content is just Character Information Items
[10/6/2004 8:54 AM] <alek_logging> i think you are ocnfusing DOM API and XML Infoset ...
[10/6/2004 8:54 AM] <Srinath> oops I thought we decided on what is in OM two weeks ago
[10/6/2004 8:55 AM] <alek_logging> DOM API pre-dates XML Infoset
[10/6/2004 8:55 AM] <Deepal> I agree with Srinath
[10/6/2004 8:55 AM] <pzf> srinath: I have seen a model where XML is stored as tables but the schema is
known... each node in the tree is prenumbered
[10/6/2004 8:55 AM] <alek_logging> Srinath: what do you mean by decided? you mean API?
[10/6/2004 8:55 AM] <pzf> but i guess this is different
[10/6/2004 8:55 AM] <sanjiva> alek: if u lose cdata doesn't canonicalization fail?
[10/6/2004 8:56 AM] -->| Farhaan (~farhaan@dhcp-128-83-88-185.commons.utexas.edu) has joined #apache-axis
[10/6/2004 8:56 AM] <Chinthaka> pzf : what we implemented was a model to hold elements, text,
attributes, comments and texts
[10/6/2004 8:56 AM] <Deepal> hi Farahan
[10/6/2004 8:56 AM] <Deepal> :)
[10/6/2004 8:56 AM] <Farhaan> Hi Deepal
[10/6/2004 8:56 AM] <alek_logging> Sanjiva: we do not keep CDATA in XSUL and have no problems with
WS-Security*
[10/6/2004 8:56 AM] <Srinath> the problem is does secuirty need them or not
[10/6/2004 8:56 AM] <Chinthaka> pzf : this is all we require in SOAP
[10/6/2004 8:56 AM] <alek_logging> but i am not security expert - maybe we should aks Dims?
[10/6/2004 8:57 AM] <Chinthaka> pzf : when we say that, that implies that we have considered the schema
[10/6/2004 8:57 AM] <alek_logging> looking on
http://webservices.xml.com/pub/a/ws/2002/10/09/canonicalization.html
[10/6/2004 8:58 AM] <alek_logging> suggests that during canonicalization CDATA sections are removed
[10/6/2004 8:58 AM] <pzf> chinthaka: I see.... the only benefit of the pre-numbered approach is direct
access to anywhere in the tree... I see this is a different design point
[10/6/2004 8:58 AM] <pzf> on the linked list approach, I guess that is almost directly the same as XBIS
and cbXML binary encodings of XML
[10/6/2004 8:58 AM] <Chinthaka> pzf : I didn't get the last part of what u said :(
[10/6/2004 8:59 AM] <Chinthaka> pzf : do we need the direct access to somepoint ?
[10/6/2004 8:59 AM] <pzf> the numbered/schema object model is about trying to allow me to directly
access say envelope/body/quote/symbol[3]
[10/6/2004 8:59 AM] <pzf> (xpath)
[10/6/2004 9:00 AM] <pzf> but it requires the schema to be preprocessed
[10/6/2004 9:00 AM] <Chinthaka> xpath will be used when u use security, am I correct ?
[10/6/2004 9:00 AM] <pzf> this model doesn't give that but only an efficient representation of th tree
[10/6/2004 9:00 AM] <Srinath> alek_logging: where that we leve with that
[10/6/2004 9:00 AM] <sanjiva> btw guys pzf = Paul Fremantle .. in case he didn't say :)
[10/6/2004 9:00 AM] <alek_logging> leave? with what?
[10/6/2004 9:01 AM] <Chinthaka> we are making this model efficient for sequential access
[10/6/2004 9:01 AM] <Srinath> what happen to xmlinfoset discussion
[10/6/2004 9:01 AM] <Srinath> :)
[10/6/2004 9:02 AM] <chathura> think we havent try to implement xpath on OM right
[10/6/2004 9:02 AM] <alek_logging> maybe we canmove discussion to mailing list?
[10/6/2004 9:02 AM] <Ajith> hopefully we can extend jaxen to do that
[10/6/2004 9:02 AM] <Chinthaka> alek : if I say , we only need to support elements, comments, texts,
attributes, namespace and CDATA
[10/6/2004 9:02 AM] <Chinthaka> am i correct ?
[10/6/2004 9:03 AM] <Chinthaka> I mean for OM
[10/6/2004 9:03 AM] <alek_logging> it is relatively easy to use jaxen to understand any kind of XM
tree API
[10/6/2004 9:03 AM] <Srinath> alek_logging: ok let us do it mailing list
[10/6/2004 9:03 AM] <Ajith> i guess waht alek is saying is we can drop CDATA
[10/6/2004 9:03 AM] <Srinath> to new topic
[10/6/2004 9:03 AM] <Ajith> is it?
[10/6/2004 9:03 AM] <alek_logging> i think CDATA is needed only for DOM completeness and not needed in
security ...
[10/6/2004 9:04 AM] <Chinthaka> oh, is it ? then its a good thing :)
[10/6/2004 9:04 AM] <sanjiva> so if we don't need cdata that makes it easy .. what about entity refs
Alek? Can those be removed?
[10/6/2004 9:04 AM] <Ajith> entity refs not in OM yet I guess
[10/6/2004 9:05 AM] <alek_logging> i think character ii are better modeled as String and DOM:Text class
should be used to wrap String whn DOM API is required by some AXIS2 handler
[10/6/2004 9:05 AM] <alek_logging> i do not have definitive answers :) let have this (and benchmark)
dicussed in mailing list?
[10/6/2004 9:05 AM] <Chinthaka> we have a OMText class which has a String to handle
[10/6/2004 9:05 AM] <Ajith> ok
[10/6/2004 9:06 AM] <Chinthaka> comments, texts and CDATA
[10/6/2004 9:06 AM] <Chinthaka> ok, then lets do this in a mailing list
[10/6/2004 9:06 AM] <sanjiva> I guess the point is we may be able to do without OMText and use String
directly ...
[10/6/2004 9:06 AM] <Chinthaka> and , IMO, I think if we use STaX
[10/6/2004 9:07 AM] <Chinthaka> nextToken and next methods have no definite meanings
[10/6/2004 9:07 AM] <alek_logging> Chinthaka: i have seen it and i think it is waste of memory to
create wrapper around String if it does nto provide any real functionality and maybe is not needed ...
[10/6/2004 9:07 AM] <Chinthaka> alek : ok
[10/6/2004 9:07 AM] <alek_logging> definite? meanings?
[10/6/2004 9:08 AM] <Chinthaka> yeah, in XPP there is a difference between next and next token, right ?
[10/6/2004 9:08 AM] <alek_logging> Chinthaka: did yo say you posted email with Word DOC conveted to
text? i did not think it made it to axis-dev [Axis2]?
[10/6/2004 9:08 AM] <alek_logging> StAX does not have tokens so round trip is not possible
[10/6/2004 9:09 AM] <alek_logging> however for XML Infoset it should not make much of difference
[10/6/2004 9:09 AM] <Chinthaka> (I was refering to those methods )
[10/6/2004 9:10 AM] <Chinthaka> that email has gone to axis-dev :( :(
[10/6/2004 9:10 AM] <Chinthaka> btw : I need to clarify on that next and nextToken thingy
[10/6/2004 9:10 AM] <alek_logging> it seems it got lost? did you put [Axis2] in subject?
[10/6/2004 9:11 AM] <Chinthaka> yeah
[10/6/2004 9:11 AM] <Chinthaka> I have recd that email, from axis-dev :(
[10/6/2004 9:11 AM] <Deepal> i got the mail with the subject [Axis2]OM Prototype Implementation
Progress html file
[10/6/2004 9:11 AM] <Chinthaka> yeah thats the email
[10/6/2004 9:11 AM] <Srinath> chintakas mail is on the dev
[10/6/2004 9:11 AM] <Srinath> it reach me
[10/6/2004 9:11 AM] <Chinthaka> ok, back to STaX thing
[10/6/2004 9:12 AM] <alek_logging> Chinthaka: if you send patch file gice it .txt or .diff extension so
it is easier on Windowze folks that have fear of attachments ...
[10/6/2004 9:13 AM] <Chinthaka> ok
[10/6/2004 9:13 AM] <alek_logging> ok i got thisone but this is not readable in email client - it
loooks like cvs diff patch and binary files are mssing anyway ...
[10/6/2004 9:13 AM] <alek_logging> also it uses lot of Micorsoft extra markup making it unreadable as
text ...
[10/6/2004 9:14 AM] <alek_logging> example: <p class=MsoNormal
style='margin-left:.25in;mso-pagination:none;mso-layout-grid-align:
[10/6/2004 9:14 AM] <alek_logging> none;text-autospace:none'>This extends <span
class=SpellE>Dasarath’s</span>
[10/6/2004 9:14 AM] <Chinthaka> I converted the doc to html using MS word :)
[10/6/2004 9:14 AM] <Chinthaka> forget it, I will mail it again tomorrow
[10/6/2004 9:14 AM] <alek_logging> did you take a look on output?
[10/6/2004 9:15 AM] <Chinthaka> shall we resume the OM discussion
[10/6/2004 9:15 AM] <Chinthaka> yeah, sure
[10/6/2004 9:15 AM] <Chinthaka> I checked and sent :(
[10/6/2004 9:16 AM] <Chinthaka> I need to have some more info about STaX implementation
[10/6/2004 9:16 AM] <Chinthaka> I will post those questions to dev list soon
[10/6/2004 9:17 AM] <Chinthaka> the current OM API, proposed doesn't say anything about the parser
being used
[10/6/2004 9:17 AM] <Chinthaka> so one can use xpp, stax or whatever
[10/6/2004 9:18 AM] <alek_logging> let move discussion to mailing list
[10/6/2004 9:19 AM] <alek_logging> i have sent "OM Prototype Implementation Progress.htm" as email that
is readable as text ...
[10/6/2004 9:19 AM] <alek_logging> i think OM should allow multiple parsers/OM builders
[10/6/2004 9:20 AM] <alek_logging> so we continue on mailing list?
[10/6/2004 9:20 AM] <Chinthaka> ok
[10/6/2004 9:20 AM] <Ajith> ok
[10/6/2004 9:20 AM] <samisa> Yes please; it is 8.20 pm in SL :)
[10/6/2004 9:20 AM] <sanjiva> Sorry for going missing; on a call too :(
[10/6/2004 9:20 AM] <sanjiva> ok guys let's call it a nite ... thanks
[