Menu
▾
▴
Re: [ jEdit-users ] Sidekick file size setting
|
From: Dale A. <da...@gr...> - 2011-06-25 15:39:21
|
On Fri, Jun 24, 2011 at 11:24 PM, Shlomy Reinstein <sre...@gm...>wrote: > For small buffers, it is very strange that the stop button suddenly appears > and then disappears instantly (affecting the layout). I prefer Alan's > suggestion of using the same button, just change icon and tooltip while the > parsing takes place. > I like that suggestion too. > > For large buffers, the SideKick tree is large, and the creation of the tree > in the GUI (which must be done in the EDT) takes a long time. This can also > be optimized by deferring the creation of tree nodes until they have to > become visible, I did that in some of my plugins (e.g. CodeHelper). > I'll take a look at what you did in CodeHelper. I spent some time looking around for ways to optimize the tree loading and rendering, which is a problem when there are a lot of tree nodes that need to be created. I didn't find anything that gave a significant improvement, so hopefully your CodeHelper technique will help. > > Shlomy > > > 2011/6/25 Alan Ezust <ala...@gm...> > >> I tried opening a 4gb xml file and it still blocks the GUI for quite some >> time while it is building its data structures. It has to wait for the >> initial IO request to complete before it can even start parsing, and >> once it starts, it can't show the tree for quite some time. Then it shows >> the tree, but it appears to still be building some other structures. >> >> Perhaps it should not show the tree until all the structures are built, >> and do all this in a Task, one that can be viewed in the Task Monitor and >> cancelled that way? >> >> I don't like that the sidekick has a cancel parsing button. It's taking up >> too much space and it is almost never needed. If anything, the "parse" >> button should CHANGE into a cancel button while it is parsing and change >> back to a parse button when it is done. >> >> >> 2011/6/24 Dale Anson <da...@gr...> >> >>> I have checked in my changes for SideKick. I'm still seeing issues in the >>> XML plugin, but the others I've tested all seem fine. Using the 1 MB movie >>> xml file as a test, it loads very fast, like in 2 or 3 seconds. I did a few >>> copy and pastes and made it a 10 MB file, it still loads and parses in XML >>> SideKick in less than a minute on my laptop. The problems I'm seeing are a >>> combination of out of memory errors from testing with large files and some >>> things in XML sidekick getting run on the EVT that shouldn't be. Eric, if >>> you don't mind, I'll look into those next. >>> >>> Dale >>> >>> >>> >>> On Fri, Jun 24, 2011 at 10:32 AM, Dale Anson <da...@gr...>wrote: >>> >>>> Oops. Apparently I didn't actually check in my changes in SideKick, so >>>> getting the latest won't let you see what I've done so far. >>>> >>>> Shlomy, yes, the threads terminate cleanly -- mostly. I'm using >>>> SwingWorker, which has a cancel method. In the case of a cancel, 'stop' is >>>> called on the parser. I did notice that the XML plugin throws an interrupted >>>> exception when it is cancelled and I haven't looked into it yet. It could be >>>> there are other SideKick plugins with similar issues. >>>> >>>> Dale >>>> >>>> >>>> >>>> On Fri, Jun 24, 2011 at 3:09 AM, Shlomy Reinstein <sre...@gm...>wrote: >>>> >>>>> Thanks, Dale, for doing this change in SideKick. Do the parser threads >>>>> have a way of terminating cleanly? (e.g. by catching an interrupted >>>>> exception) >>>>> >>>>> Shlomy >>>>> >>>>> 2011/6/24 Dale Anson <da...@gr...> >>>>> >>>>>> Hi Eric, >>>>>> >>>>>> I updated XML plugin to the latest from svn. It looks like there is an >>>>>> issue at XMLParsedData.java in the sort() method. I think the contents of >>>>>> this method needs wrapped in a SwingWorker. I was testing some changes I >>>>>> made in SideKick and was using the big movie xml file that Eric Berry has >>>>>> posted a link to a few times. I wanted the file to be bigger since it was >>>>>> parsing fairly fast for me, so I copied all of the "movie" nodes and tried >>>>>> to paste them at the bottom of the file after the last "movie" node. This >>>>>> causes jEdit to hang. A thread dump shows it's stuck in the sort() method. >>>>>> >>>>>> Also, I've done some work on the threading in SideKick, so the "Stop" >>>>>> button that Shlomy added works well -- or at least, it seems to work well >>>>>> for me. I'd appreciate if you could grab the latest SideKick code and test >>>>>> it out. >>>>>> >>>>>> Thanks, >>>>>> >>>>>> Dale >>>>>> >>>>>> >>>>>> >>>>>> On Mon, Jun 20, 2011 at 2:37 PM, Eric Le Lay < >>>>>> ker...@us...> wrote: >>>>>> >>>>>>> Le 04/06/2011 21:00, Mike Maxwell a écrit : >>>>>>> > On 6/3/2011 12:25 PM, Eric Le Lay wrote: >>>>>>> >> I've implemented a version of tag-matching based on the already >>>>>>> >> available sidekick information. It's performing better for big XML >>>>>>> >> files but it's less accurate than the original version when you >>>>>>> are >>>>>>> >> editing the buffer, since the sidekick tree becomes outdated so I >>>>>>> >> haven't committed it yet. >>>>>>> > >>>>>>> > Clearly there's a trade-off, but for me at least it would be an >>>>>>> > acceptable trade-off when editing large files. Especially if there >>>>>>> were >>>>>>> > a way to tell it to re-parse the file while I made my coffee :-) >>>>>>> > >>>>>>> Hi Mike, >>>>>>> >>>>>>> you can try the latest (2011-06-21 >>>>>>> >>>>>>> http://www.tellurianring.com/projects/jedit-daily/index.php?dir=XML%2F2011-06-20_12-02-05%2F >>>>>>> ) >>>>>>> XMLPlugin build : >>>>>>> I hope it will be much faster at navigating through the file, once >>>>>>> parsing has been done, >>>>>>> since has the new tag-matching mechanism on by default. >>>>>>> >>>>>>> To get the old tag-matching back, go to Utilities > Evaluate >>>>>>> Beanshell >>>>>>> and type >>>>>>> jEdit.setProperty("xml.structure-matcher","old"). It should also be a >>>>>>> bit faster... >>>>>>> >>>>>>> Cheers, >>>>>>> Eric >>>>>>> >>>>>>> >>>>>>> ------------------------------------------------------------------------------ >>>>>>> EditLive Enterprise is the world's most technically advanced content >>>>>>> authoring tool. Experience the power of Track Changes, Inline Image >>>>>>> Editing and ensure content is compliant with Accessibility Checking. >>>>>>> http://p.sf.net/sfu/ephox-dev2dev >>>>>>> -- >>>>>>> ----------------------------------------------- >>>>>>> jEdit Users' List >>>>>>> jEd...@li... >>>>>>> https://lists.sourceforge.net/lists/listinfo/jedit-users >>>>>>> >>>>>> >>>>>> >>>>>> >>>>>> ------------------------------------------------------------------------------ >>>>>> All the data continuously generated in your IT infrastructure contains >>>>>> a >>>>>> definitive record of customers, application performance, security >>>>>> threats, fraudulent activity and more. Splunk takes this data and >>>>>> makes >>>>>> sense of it. Business sense. IT sense. Common sense.. >>>>>> http://p.sf.net/sfu/splunk-d2d-c1 >>>>>> >>>>>> -- >>>>>> ----------------------------------------------- >>>>>> jEdit Users' List >>>>>> jEd...@li... >>>>>> https://lists.sourceforge.net/lists/listinfo/jedit-users >>>>>> >>>>>> >>>>> >>>> >>> >>> >>> ------------------------------------------------------------------------------ >>> All the data continuously generated in your IT infrastructure contains a >>> definitive record of customers, application performance, security >>> threats, fraudulent activity and more. Splunk takes this data and makes >>> sense of it. Business sense. IT sense. Common sense.. >>> http://p.sf.net/sfu/splunk-d2d-c1 >>> -- >>> ----------------------------------------------- >>> jEdit Users' List >>> jEd...@li... >>> https://lists.sourceforge.net/lists/listinfo/jedit-users >>> >>> >> >> >> ------------------------------------------------------------------------------ >> All the data continuously generated in your IT infrastructure contains a >> definitive record of customers, application performance, security >> threats, fraudulent activity and more. Splunk takes this data and makes >> sense of it. Business sense. IT sense. Common sense.. >> http://p.sf.net/sfu/splunk-d2d-c1 >> -- >> ----------------------------------------------- >> jEdit Users' List >> jEd...@li... >> https://lists.sourceforge.net/lists/listinfo/jedit-users >> >> > |