So is it me, my screen, an optical resolution, or how _putimage works that leads to me not actually getting square pixels with this little bit of code?   The red dots created at 15,15 and 16,16 in the sample image just aren't the same size when I run this code:

Ok, so this is probably quite simple, but it's frustrating the heck out of me! 
 This is A.jpg

(Please try to keep response simple).

This is a program I created when very bored to take a list of 19-digit integers from a text file and figure out if they're prime or not. I was trying to get a palindrome out of one of them but failed so far LOL.

The program is listed below but is worthless without the data files which are part of the attachment.

A for the giggles mod of @bplus ' https://staging.qb64phoenix.com/showthre...2#pid14122

After you open the link below to run the program:

• to pause the action, click and hold (or touch and hold for touch device) on the program's window
  
• to have a new random colour scheme, refresh the page
  

When using graphics, especially in games, it's often necessary to clear an image from a surface before placing another image in its place because of alpha transparency issues. The method I usually use to do this was way to slow for the current project I'm working on. So I started playing with and timing alternative methods.

Thought I would pass this tidbit of information along.

Using a 512x736 32bit image and each method repeated 10000 times I get the following.

My usual method: 0.644 seconds

Odest = _DEST
_DEST Image
CLS
_DEST Odest


Using LINE BF method: 0.620 seconds (curious that LINE is faster than CLS)

Odest = _DEST
_DEST Image
LINE(0, 0)-(_WIDTH(Image) - 1, _HEIGHT(Image) -1), _RGB32(0, 0, 0), BF
_DEST Odest


Freeing and recreating: 0.042 to 0.061 seconds (no idea why it fluctuates so much)

Iwidth = _WIDTH(Image)
Iheight = _HEIGHT(Image)
_FREEIMAGE Image
Image = _NEWIMAGE(Iwidth, Iheight, 32)

Well, as you can see, the 3rd method is superior. I created a simple subroutine to handle this for me now:

SUB RENEW_IMAGE (Image AS LONG)

    DIM Iwidth AS INTEGER '  width of image
    DIM Iheight AS INTEGER ' height of image

    Iwidth = _WIDTH(Image) '                get width of image
    Iheight = _HEIGHT(Image) '              get height of image
    _FREEIMAGE Image '                      remove the image from RAM
    Image = _NEWIMAGE(Iwidth, Iheight, 32) ' recreate a blank surface

END SUB

To add more files to EXE you will need the programs listed in this thread, but first also the PMF2 compressor program which is in the previous thread:  https://staging.qb64phoenix.com/showthre...5#pid14125

Be careful what you do, I always recommend COPYING the program to another folder and putting the files there. I've tried it all and it's a working procedure. The described procedure assumes that you have already created a PMF2 file that contains all the files you need to include in the EXE file.


STEP 1:

Let me introduce the PMF2 (that's my own file format) archiver. It is a compression program, the main advantage of which is that files can be compressed and decompressed using only QB64 directly from the source code. It allows you to view packed files and can unpack just one specific file or all files from the archive. It has a built-in feature that makes sure that if the unzipped file already exists on the hard drive, the new unzipped file will have a parenthesis with a number after the name, just like Windows does.

So you get the option to have many files in a single file with the option to extract just one specific one or all files.
Higher operations, such as updating a specific file in the archive, or deleting one file in the archive are not yet supported, I also saved something for another time, that's clear after all

Well, it's probably dawning on someone. So who asked how to add more files to the EXE? I'll add a new thread directly to that in a moment. 



rewrite in source code file names for your correct files in array FL for trying!

Piral:

Wondering if any of you guys have done any coding with Blockchain. As I understand it, there is basically a record which is not written until all records are searched and the new record to be written is deemed valid in someway before it can be written to the chain of records. The two main/key points I take away from this is that VALIDATION before writing the record can be very complex and secondly, the chain must be sequential or there has to be an index table that allows the records to be found in the order in which they were written. So in pseudocode the layout would be something like this:


- compose a validation routine which comprises decisions on what is accurate and acceptable

- Input a record

- Review the present chain
  a:for duplication of inputted record
  b:for a location in chain where the inputted record will be written
  c: subject the inputted record to the validation routine
  d: write the inputted record or reject the record.
  e: keep an index as to where this inputted record can be found

The blockchain that I'm trying to work on is not for public input but rather something which raises the accuracy of the data in the data base and improves retrieval of any given small piece of data stored. In this regard, what I don't have in the above pseudocode layout is a search routine. It would seem I need 2 search routines, one associated with the Review of the chain before writing a record and one associated with a search of data where No record is intended on being written. A number of years ago I started on a blockchain program but found it was going to be extremely large and abandoned it. So I'm looking at it again but trying to conceptualize a smaller program.

Here is a small program for solving a Sudoku.
Start the program, press "C" to clear the field. Then fill in the sudoku that you can find in newspapers or other media and finally press "S" to solve the sudoku.
You cannot solve it yourself, the computer does all the work for you.
Comments are welcome 


  sudoku.bas (Size: 10.37 KB / Downloads: 83)

Attached Files

Thumbnail(s)