Hi Thiwanka, I've just reviewed your proposal. Looks good overall, but I have a few suggestions for improvement:
For the asynchronous LSParser it would be good if you mentioned something about how the main thread and the 2nd thread that the parser runs on interact, specifically your thoughts on updating the LSParser's 'busy' state and how a call to abort() would terminate the parsing thread if it hadn't completed yet. The return value of dispatchEvent() is determined to be true/false based on whether preventDefault() was called on the Event by an EventListener, not on whether an EventListener could "handle" the event. It would be good if you would correct that. There's a typo in the parseWithContext() section. I'm sure you meant " <ns2:c/><ns1:d/>" (two empty-tags) instead of "</ns2:c></ns1:d>" (two unpaired end-tags). In the last section on parseWithContext() you seem to be implying that a Document node (as the context node) with an action of ACTION_REPLACE_CHILDREN is an error condition. When I said it was an exception, I meant it was an exceptional case that needed to be handled differently than the other inputs. It's not an error, but is the one case where a whole XML document is expected instead of a fragment. More detail in my previous reply below. Once you're ready feel free to enter your proposal into the GSoC tool. Great work! Thanks. Michael Glavassevich XML Parser Development IBM Toronto Lab E-mail: mrgla...@ca.ibm.com E-mail: mrgla...@apache.org Thiwanka Somasiri <asthiwa...@gmail.com> wrote on 04/05/2011 12:43:32 AM: > Hi Michael and devs, > > As per discussion with Michael Glavassevich about the > "Asynchronous LSParser and parseWithContext()" GSoC project under XERCESJ-1429 > , I have prepared my proposal and attached it to this mail. I would > like to know whether I should improve the proposal and if so which > areas should be improved in it. Hope you feedbacks soon since the > deadline is on 8th of April. > > P.S. - Michael, I have included you as the mentor for the project. > > Thanks. > > On Fri, Apr 1, 2011 at 9:58 AM, Michael Glavassevich <mrgla...@ca.ibm.com > > wrote: > Hi Thiwanka, > > Thiwanka Somasiri <asthiwa...@gmail.com> wrote on 03/31/2011 08:21:55 AM: > > > Hi Michael, > > > > > By using a synthesized wrapper XML document, does > > it mean to convert the fragment to a temporary complete XML document > > (using a dummy root - so it can be parsed like a normal XML > > document), parse it and merge the nodes created with this to the > > existing document? If so, whenever the fragment contains an > > erroneous point, an exception will be thrown. Otherwise we can > > extract the relevant nodes to the XML fragment (removing dummy root > > nodes) and add them to the existing document. > Right, that's what I meant, though you're not really 'converting' > the fragment, but including it by reference in a dummy wrapper XML > document so that it can be parsed like any other XML document. The > one exception is when the context node is a Document node and the > action is ACTION_REPLACE_CHILDREN. You would process the input as a > whole XML document, very similar to if the input was parsed using > the regular LSParser.parse() method. > > > > I have attached an image(http://img145.imageshack.us/i/ > > xmlfragmentinsertion.png/) describing what I understood from you > > previous reply. If I have misunderstood that, please correct me. > Your diagram illustrates the idea quite well. > > > > I am looking for your advices to start with my proposal. I want to > > know what are the key point that should be included and is it enough > > to have high level implementation details in the proposal? > Your proposal should include a high level design for the > asynchronous mode for the LSParser and parseWithContext(), as well > as details on your thoughts for implementation. It should include a > plan/timeline for the work you would be doing, some details about > yourself and your development experience, details about the > interaction that you've had with the community and also say > something about the time you would be able to commit to the project > over the summer. Proposals from 2010 are up on the Xerces Wiki ( > http://wiki.apache.org/xerces/FrontPage). You may want to take a > look at them to get a better idea of the structure and level of > detail in a successful GSoC proposal. > > > > Thanks. > > > > On Thu, Mar 31, 2011 at 10:40 AM, Michael Glavassevich <mrgla...@ca.ibm.com > > > wrote: > > Hi Thiwanka, > > > > Aside from the previous discussion on the mailing list [1] and the > > specification I'm not aware of any good reading material on > > parseWithContext(). > > > > I have given the design more thought since last year and think it > > should be possible to implement this method (in the case where the > > context node is not a Document node) by synthesizing a wrapper XML > > document which contains a reference to the fragment and the > > necessary context (e.g. namespace declarations) required to parse > > the fragment. You can then transfer the nodes created for the > > fragment into the existing DOM. > > > > For example, if your existing document looked like this: > > > > <ns1:a xmlns:ns1="http://ns1";> > > <ns2:b xmlns:ns2="http://ns2"/> > > </ns1:a> > > > > and you wanted to insert an XML fragment like: > > > > <ns2:c/><ns1:d/> > > > > as a child of 'ns2:b', you could generate a wrapper document: > > > > <!DOCTYPE DUMMY_ROOT [ > > <!ENTITY fragment PUBLIC "***" "***"> > > ]> > > <DUMMY_ROOT xmlns:ns1="http://ns1"; xmlns:ns2="http://ns2 > > ">&fragment;</DUMMY_ROOT> > > > > where the 'fragment' entity points to the XML fragment the user > > provided, parse the document you synthesized and then move the nodes > > created underneath the entity reference into the existing document, > > resulting in: > > > > <ns1:a xmlns:ns1="http://ns1";> > > <ns2:b xmlns:ns2="http://ns2";><ns2:c/><ns1:d/></ns2:b> > > </ns1:a> > > > > Hope that made sense. > > > > Thanks. > > > > [1] http://markmail.org/thread/ool7dyyzffor6o7b > > > > > > Michael Glavassevich > > XML Parser Development > > IBM Toronto Lab > > E-mail: mrgla...@ca.ibm.com > > E-mail: mrgla...@apache.org > > > Thiwanka Somasiri <asthiwa...@gmail.com> wrote on 03/28/2011 02:05:24 PM: > > > > > Hi Michael, > > > > > > > > As I understood, the second part of the project is also a critical > > > implementation, which is known as parseWithContext(). According to > > > the W3C specification[1], this method is used to insert a fragment > > > of and XML document to an existing document. The parameters in the > > > method specify the position where the new fragment should be > > > attached and which action should be triggered between the existing > > > document and the new fragment. > > > > > > The implementation also include an error handling part (in a case > > > where the inserting fragment is erroneous,etc) to notify the caller > > > through an ErrorHandler object when an unexpected scenario occur. > > > > > > I need to understand the background for parseWithContext() some more > > > and can you tell me what are the useful resources in this regard. > > > (Understanding the overall idea of this will help me to allocate > > > approximate time in the proposal) > > > > > > I would like to start writing a proposal for the project and I need > > > a big help from you. I am happy if you can give me an starting point > > > and guide lines to write the proposal. > > > > > > [1] http://www.w3.org/TR/2004/REC-DOM-Level-3-LS-20040407/load- > > > save.html#LS-LSParser-parseWithContext > > > > > > Thanks. > > > > > > On Sat, Mar 19, 2011 at 11:41 PM, Michael Glavassevich < > mrgla...@ca.ibm.com > > > > wrote: > > > Hi Thiwanka, > > > > > > Thiwanka Somasiri <asthiwa...@gmail.com> wrote on 03/19/2011 02:25:03 AM: > > > > > > > Hi Michael, > > > > > > > > > > > Thanks for your reply. I will study more on the project within > > > > next week and come up with both my issues and ideas. Since I was > > > > having my exams, I could not focus much on the project and will be > > > > free after few more days. I hope the support of you all to step forward. > > > > > That's fine. It's still early on in the process so still quite a bit > > > of time to put together and discuss a proposal. > > > > > > > > > > Thanks. > > > > > > Thanks. > > > > > > Michael Glavassevich > > > XML Parser Development > > > IBM Toronto Lab > > > E-mail: mrgla...@ca.ibm.com > > > E-mail: mrgla...@apache.org > > > > > > > > > > > > -- > > > > > > Regards > > > > > > A.S.Thiwanka Somasiri > > > > > > Skype : executionerwild > > > MSN : thi...@ymail.com > > > > > > -- > > > > Regards > > > > A.S.Thiwanka Somasiri > > > > Skype : executionerwild > > MSN : thi...@ymail.com > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: j-dev-unsubscr...@xerces.apache.org > > For additional commands, e-mail: j-dev-h...@xerces.apache.org > > > Thanks. > > Michael Glavassevich > XML Parser Development > IBM Toronto Lab > E-mail: mrgla...@ca.ibm.com > E-mail: mrgla...@apache.org > > > > -- > > Regards > > A.S.Thiwanka Somasiri > > Skype : executionerwild > MSN : thi...@ymail.com > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: j-dev-unsubscr...@xerces.apache.org > For additional commands, e-mail: j-dev-h...@xerces.apache.org
