How many rows should be in the (main) buffer of a virtual Listview control?
I am witting an application in pure 'c' to the Win32 API. There is an ODBC connection to a database which will retrieve the items (actually rows).
The MSDN sample code implies a fixed size buffer of 30 for the end cache (Which would almost certainly not be optimal). I think the end cache and the main cache should be the same size.
My thinking is that the buffer should be more than the maximum number of items that could be displayed by the list view at one time. I guess this could be re-calculated each time the Listivew was resized?
Or, is it just better to go with a large fixed value. If so what is that value?
Use the ListView_ApproximateViewRect (or the LVM_APPROXIMATEVIEWRECT message) to get the view rect height.
Use the ListView_GetItemRect (or the LVM_GETITEMRECT message) to get the height of an item.
Divide the view rect height by the height of an item to get the number of items that can fit in your view.
Do this calculation on each size event.
Then create your buffer accordingly.
The LVN_ODCACHEHINT notification message will let you know how many items it is going to ask. This could help you in deciding how big your cache should be.
#Brian R. Bondy Thanks for the explicit help for how to do get the number of items. In fact I was all ready working my way to understanding that it could be done (for list or report view) with ListView_GetCountPerPage, and I would use you way to get it for for the others, though I don't need the ListView_ApproximateViewRect since I will all ready know the new size of the ListView.
#Lars Truijens I am already using the LVN_ODCACHEHINT and had though about using that to set the buffer size, however I need to read to the end of the SQL data to find the last item to get the number of rows returned from ODBC. Since that would be the best time to fill the 'end cache' I think I have to set up the number of items (and therefore fill the buffer) before we get a call to LVN_ODCACHEHIN.
I guess my real question is one of optimization for which I think Brian hinted at the answer. The amount of overhead in trashing your buffer and reallocating memory is smaller than the overhead of going out to the network and doing an ODBC read, some make the buffer fairly small and change it often rather.
Is this right?
I have done a little more playing around, and it seems that think that LVN_ODCACHEHINT generally fills the main buffer correctly, and only misses if a row (in report mode) is partially visible.
So I think the answer for the size of the cache is: The total number of displayed items, plus one row of displayed items (since in the icons views you have multiple items per row).
You would then re-read the cache with each WM_SIZE and LVN_ODCACHEHINT if either had a different begin and end item number.
The answer would seem to be: (Or a random collection of notes as I fiddle around with ideas)
As a general answer for buffers:
Start with some amount, in this case a screen full (I add an extra row in case the next is partially uncovered), and then every time the screen is scrolled, double the buffer size (up to the point before you run out of memory).
Which would seem to be wrong. As it turns out, most ways of loading data are all ready buffered. ODBC calls of File I/O. Pretty much anything that isn't that I can think off is either in memory or is recalculated on the fly. This means the answer really is: take the values provided in LVN_ODCACHEHINT (and add 1 either side - this just seems to work faster if you don't have an integral height).
Related
I have a list of anything from 10-500 records appearing on screen as a graphical array using standard NG-REPEAT, visualisation is a row of photographs
What I would like to achieve, but cant quite get my head around if its even possible is this...
A 'window' on screen that will only show 10 records - Done
You can scroll left or right through the entire list - Done
HOWEVER
What i would LIKE to achieve is that when the end of the list is reached, the list starts again from the beginning but in an infinite style loop. Both directions
I have never done anything like this and not quite sure how I would achieve this of my ng-repeat (item in ListofItems)
So I assume that the best solution is
when the user gets 40 records through 100 records, i push() the first 20 records from my $scope.ListofItems to the end of the array, and remove the equivalent first 20 records from the front.. this keeps a constant list of 100 records, no massive memory usage.. i can even do this on an individual record basis.. remove one record from end or beginning and add at the beginning or end of the list.
The user experience is an infinite scroll, but I suspect the browser experience could be slow and stuttering due to the processing going on
Push(), Pull(), Tracking and indexing would play a part but any suggestions would be appreciated whether this is even technically possible
EDIT : After some research. maybe
$scope.ListofItems.push($scope.ListofItems.shift());
could be a solution to move beginning to end, but not sure how to trigger this or go the other way (pull?)
EDIT2 : Just did a manually called function for the above and it shifts front item to end of list, though i would have no clue how to read the screen position to know when to fire the function
Going the other way around could be
$scope.ListofItems.unshift($scope.ListofItems.pop());
As a side note, I think it's important to keep in mind for large arrays that shift() and unshift() naturally cause a full reindexing and have a time complexity of O(n) where n is the length of the array, as opposed to push() and pop() which have a time complexity of O(1)
I'm trying to get pixel data from my X11 instance, I've seen this thread (How do take a screenshot correctly with xlib?) and the double for loop is taking just too long for me (over a million loops, as the system I'm building requires the highest amount of efficiency possible, sitting around for 600 miliseconds is just not an option). Is there no way to just get a raw array of pixels to avoid the for loop? I know the XImage class has a "data" member which is supposed to contain all of the pixels, but the organizational system that it uses is foreign to me. Any help would be greatly appreciated! The best end game here would be for me to just be able to have X11 write directly to /dev/fb0
I am trying to print a report where we have several different components within the xaml.
By what I`ve found, when printing, you have to treat every UIelement as a single one, thus if the desiredSize is bigger than the AvailableSize you have to activate the flag HasMorePages.
But here comes the problem.
My user can write as much text as he/she wants on the grid, therefore, depending on the amount, the row expands and goes off the printable area, as you can see on the picture below.
I thought about giving a whole page to the grid, but it was to big still, which got me into a loop where the DesizedSize was always bigger than the PrintableArea.
My code is not very different from any source you find on internet when searching for Multiple Page printing.
It is based on this http://eswarbandaru.blogspot.com.au/2011/02/print-mulitple-pages-using-silverlight.html , but using Stackpanels instead of Textboxes.
Any idea?
Thank you in advance.
First you need to work out how many pages are needed
Dim pagesNeeded As Integer = Math.Ceiling(gridHeight / pageHeight) 'gets number of pages needed
Then once the first page has been sent to the printer, you need to move that data out of view and bring the new data into view ready to print. I do this by converting the whole dataset into an image/UI element, i can then adjust Y value accordingly to bring the next set of required data on screen.
transformGroup.Children.Add(New TranslateTransform() With {.Y = -(pageIndex * pageHeight)})
Then once the number of needed pages is reached, tell the printer to stop
If pagesLeft <= 0 Then
e.HasMorePages = False
Exit Sub
Else
e.HasMorePages = True
End If
Or if this is too much work, you can simply just scale all the notes to fit onto screen. Again probably by converting to UI element.
Check out this link for converting to a UI element.
http://www.codeproject.com/Tips/248553/Silverlight-converting-to-image-and-printing-an-UI
Hope this helps
I have this code
n_userobject inv_userobject[]
For i = 1 to dw_1.Rowcount()
inv_userobject[i] = create n_userobject
.
.
.
NEXT
dw_1.rowcount() returns only 210 rows. Its so odd that in the range of 170 up, the application stop and crashes on inv_userobject[i] = create n_userobject.
My question, is there any limit on array or userobject declaration using arrays?
I already try destroying it after the loop so as to check if that will be a possible solution, but it is still crashing.
Or how can i be able to somehow refresh the userobject?
Or is there anyone out there encounter this?
Thanks for all your help.
First, your memory problem. You're definitely not running into an array limit. If I was to take a guess, one of the instance variables in n_userobject isn't being cleaned up properly (i.e. pointing to a class that isn't being destroyed when the parent class is destroyed) or pointing to a class that similarly doesn't clean itself up. If you've got PB Enterprise, I'd do a profiling trace with a smaller loop and see what is being garbage collected (there's a utility called CDMatch that really helps this process).
Secondly, let's face it, you're just doing this to avoid writing a reset method. Even if you get this functional, it will never be as efficient as writing your own reset method and reusing the same instance over again. Yes, it's another method you'll have to maintain whenever the instance variable list changes or the defaults change, but you'll easily gain that back in performance.
Good luck,
Terry.
I'm assuming the crash you're facing is at the PBVM level, and not a regular PB exception (which you can catch in your code). If I'm wrong, please add the exception details.
A loop of 170-210 iterations really isn't a large one. However, crashes within loops are usually the result of resource exhaustion. What we usually do in long loops is call GarbageCollect() occasionally. How often should it be called depends on what your code does - using it frequently could allow the use of less memory, but it will slow down the run. Read this for more.
If this doesn't help, make sure the error does not come from some non-PB code (imported DLL or so). You can check the stack trace during the crash to see the exception's origin.
Lastly, if you're supported by Sybase (or a local representative), you can send them a crash dump. They can analyze it, and see if it's a bug in PB, and if so, let you know when it was (or will be) fixed.
What I would normally do with a DataWindow is to create an object that processes the data in a row and call it for each row.
the only suggestion i have for this is to remove the rowcount from the for (For i = 1 to dw_1.Rowcount()) this will cause the code to recount the rows every time it uses one. get the count into a variable and then use the variable. it should run a bit better and be far more easy to debug.
I'm currently writing an eBook reader for Windows Phone Seven, and I'm trying to style it like the Kindle reader. In order to do so, I need to split my books up into pages, and this is going to get a lot more complex when variable font sizes are added.
To do this at the moment, I just add a word at a time into the textblock until it becomes higher than its container. As you can imagine though, with a document of over 120,000 words, this takes an unacceptable period of time.
Is there a way I can find out when the text would exceed the bounds (logically dividing it into pages), without having to actually render it? That way I'd be able to run it in a background thread so the user can keep reading in the meantime.
So far, the only idea that has occurred to me is to find out how the textblock decides its bounds (in the measure call?), but I have no idea how to find that code, because reflector didn't show anything.
Thanks in advance!
From what I can see the Kindle app appears to use a similar algorithm to the one you suggest. Note that:
it generally shows the % position through the book - it doesn't show total number of pages.
if you change the font size, then the first word on the page remains the same (so that's where the % comes from) - so the Kindle app just does one page worth of repagination assuming the first word of the page stays the same.
if you change the font size and then scroll back to the first page, then actually there is a discontinuity - they pull content forwards again in order to fill the first page.
Based on this, I would suggest you do not index the whole book. Instead just concentrate on the current page based on a "position" of some kind (e.g. character count - displayed as a percentage). If you have to do something on a background thread, then just look at the next page (and maybe the prev page) in order that scrolling can be more responsive.
Further to optimise your experience, there are a couple of changes you could make to your current algorithm that you could try:
try a different starting point and search increment for your algorithm - no need to start at one word and to then only add one word at a time.
assuming most of your books are ASCII, try caching the width of the common characters, and then work out the width of textblocks yourself.
Beyond that, I'd also quite like to try using <Run> blocks within your TextBlock - it may be possible to get the relative position of each Run within the TextBlock - although I've not managed to do this yet.
I do something similar to adjust font size for individual textboxes (to ensure they all fit). Basically, I create a TextBlock in code, set all my properties and check the ActualWidth and ActualHeight properties. Here is some pseudo code to help with your problem:
public static String PageText(TextBlock txtPage, String BookText)
{
TextBlock t = new TextBlock();
t.FontFamily = txtPage.FontFamily;
t.FontStyle = txtPage.FontStyle;
t.FontWeight = txtPage.FontWeight;
t.FontSize = txtPage.FontSize;
t.Text = BookText;
Size Actual = new Size();
Actual.Width = t.ActualWidth;
Actual.Height = t.ActualHeight;
if(Actual.Height <= txtPage.ActualHeight)
return BookText;
Double hRatio = txtPage.ActualHeight / Actual.Height;
return s.Substring((int)((s.Length - 1) * hRatio));
}
The above is untested code, but hopefully can get you started. Basically it sees if the text can fit in the box, if so you're good to go. If not, it finds out what percentage of the text can fit and returns it. This does not take word breaks into account, and may not be a perfect match, but should get you close.
You could alter this code to return the length rather than the actual substring and use that as your page size. Creating the textblock in code (with no display) actually performs pretty well (I do it in some table views with no noticeable lag). I wouldn't send all 120,000 words to this function, but a reasonable subset of some sort.
Once you have the ideal length you can use a RegEx to split the book into pages. There are examples on this site of RegEx that break on word boundaries after a specific length.
Another option, is to calculate page size ahead of time for each potential fontsize (and hardcode it with a switch statement). This could easily get crazy if you are allowing any font and any size combinations, and would be awful if you allowed mixed fonts/sizes, but would perform very well. Most likely you have a particular range of readable sizes, and just a few fonts. Creating a test app to calculate the text length of a page for each of these combinations wouldn't be that hard and would probably make your life easier - even if it doesn't "feel" right as a programmer :)
I didn't find any reference to this example from Microsoft called: "Principles of Pagination".
It has some interesting sample code running in Windows Phone.
http://msdn.microsoft.com/en-us/magazine/hh205757.aspx
You can also look this article about Page Transitions in Windows Phone and this other about the final touches in the E-Book project.
The code is downloadable: http://archive.msdn.microsoft.com/mag201111UIFrontiers/Release/ProjectReleases.aspx?ReleaseId=5776
You can query the FormattedText class that is used AFAIK inside textBlock. since this is the class being used to format text in preparation for Rendering, this is the most lower-level class available, and should be fast.