Recent

Recent Posts

Pages: [1] 2 3 ... 10
1
Cocoa / WKWebView takeSnapshot
« Last post by MISV on Today at 04:53:41 pm »
The WKWebView has a separate function for taking screenshots since

Code: Pascal  [Select][+][-]
  1.           TmpNSBitmapRep := FWebBrowser.FWebView_CocoaWKWebView.bitmapImageRepForCachingDisplayInRect(TmpNSRect);
  2.           TmpNSBitmapRep.setSize(TmpNSSize);
  3.           FWebBrowser.FWebView_CocoaWKWebView.cacheDisplayInRect_toBitmapImageRep(TmpNSRect, TmpNSBitmapRep);
  4.  

is not working on this control


Has anyone successfully implemented all headers and got it working?


https://stackoverflow.com/questions/31969950/wkwebview-cachedisplayinrect-screenshot-in-os-x
2
Other / Re: Question for people who have built commercial apps
« Last post by VisualLab on Today at 04:46:24 pm »
Typically, the software that the average Smith might need and be able to use (at least on a basic level) comes in commercial and free versions. For example:

  N1. content browsers: WWW, PDF, photos,
  N2. players: audio, video,
  N3. editors: text, presentation, raster and vector graphics, audio and video editing,
  N4. computational: spreadsheet, etc.

In this category (N), there are actually a lot of free programs. This is an obvious issue, as previous speakers have already described.

The second category (S) are programs that the average Smith does not need (e.g. at home, for his own needs)1, because they are usually of no use to him. These are programs like:

  S1. databases for servicing companies/institutions (i.e. for "bureaucracy" of various types): customers, patients, accounting, etc.
  S2. engineering - various specialized CADs, CAMs, etc. (mechanics, chemistry, construction, electronics, etc.),
  S3. laboratories, hospitals (quality control, forensics, etc.): operation of laboratory and medical equipment and specific data processing (numerical, statistical),
  S4. scientific – highly specialized programs with a narrow subject, designed to simulate phenomena, processes, etc. (physics, chemistry, biology, etc.).

In the S1 category it should still be profitable to create this type of software, but there is huge competition (these programs are relatively easy to create compared to the next 3 groups in this category).

In the S2 and S3 categories, software development should be most profitable. But there are much fewer recipients of such programs than in the S1 category. Additionally, developing software in S2 and S3 categories is much more difficult and time-consuming.

Category S4 is basically unprofitable. The number of recipients is relatively small, very specific, and the production costs are very high. Moreover, in this group there is an increasing expectation that the program will be free (especially those written for biologists dealing with the processing of genetic or protein sequence data). These people have heard somewhere that there are modules or libraries developed in Python. Many of these biologists, who have quickly learned a bit of scripting "programming", have created "applications" and make them available for free, without any guarantees. Basically, it is a plague of "IT crap". I criticize this, not because they are free, but because they are most often "monsters", full of omissions and errors.

To sum up: large engineering, laboratory and medical programs can only be created by large companies, because it requires many experienced programmers and large funds. But small groups of programmers (companies) can also create useful software, for example for smaller companies (hospitals are usually excluded due to the need to provide various certificates, etc. documents). It remains a matter of deciding how to choose a sales model. For a small software company, subscription sales can be the final nail in the coffin. And here I share the opinion of marcov.

---
1) I am leaving out the rather small percentage of people in society, such as technology enthusiasts and hobbyists, who need (want) to use engineering programs (e.g. electronics, microcontrollers, etc.) and are able to use them.
3
General / Re: How to: create DLL file for Windows 10 64-Bit Pro
« Last post by paule32 on Today at 04:46:23 pm »
Thank You for your feedback.

The comparison itself is not the problem - I fix this.
The problem is, that the "write" give Windows access denied error $c0005 - see attachment.
I log the Events of Windows, and the Application seems to raise a silent exception.

$c0005 mean, that the application tries to access memory, where is protected.

This come to me in mind, that something else is going on there.
A virus scanner ? - I have to check...
4
General / Re: Slow copying of small structures
« Last post by Nitorami on Today at 04:41:52 pm »
@Alpine: yes, understood.
@Marcov: I do not know if it is theoretically possible to optimise out the temporaries and the REP MOVSL. The compiler currently does not do it, even with inlining. Only for structures up to 8 bytes in size, VMOVSD is used instead of REP MOVSL. Operations with complex number of type single are therefore significantly faster than when using double.
5
General / Re: How to: create DLL file for Windows 10 64-Bit Pro
« Last post by marcov on Today at 04:41:49 pm »
Afaik it was always key to do as little as possible in the dll main event.

If you move the code to a separate procedure, does it work then?
6
General / Re: How to: create DLL file for Windows 10 64-Bit Pro
« Last post by KodeZwerg on Today at 04:25:09 pm »
Code: Pascal  [Select][+][-]
  1.         if ((dummy = INVALID_SET_FILE_POINTER) and (error = NO_ERROR)) then begin
Code: Pascal  [Select][+][-]
  1.       if ((dummy <> INVALID_SET_FILE_POINTER) and (error = NO_ERROR)) then
Why in the world you change it like that? Do you even understand the meaning of that comparision?
I do not check for more modifications made by you but a simple question remain, lets assume you truly copy and paste my example, letting each single char untouched, no "paule32 optimizations", just like I showed it, will it run successful or not?
7
General / Re: location of non visual component
« Last post by gerardus on Today at 04:16:44 pm »
Crude solution:

Open the .lfm file with a text editor and search for your TImageList

  object ImageList1: TImageList
    Left = 148
    Top = 60
  end

and modify the Left and Top.
And yes, it would nice to have an IDE command to center "lost" components in the form.
8
General / Re: Slow copying of small structures
« Last post by alpine on Today at 04:15:44 pm »
Ok the temporaries are necessary but this would not be of concern if copying them wasn't so slow. The reason why the compiler inserts no REP MOVSL in your code is, I believe, because you use single. When changing to double, the REP MOVSL are back.
True, that's why I put the float_type - to experiment with different types.

My point is - while the small structure copying can be optimized, and that will be great by itself, the operators will stay suboptimal compared to the procedures, since the procedures won't require copying at all. They'll evaluate the calculations in correct order using references, probably without the need for intermediates. 
9
FPC development / Re: what to do if my target MIPS cpu has no FPU
« Last post by Laksen on Today at 04:14:52 pm »
Check how the embedded target does it. Most of those don't have FPU support

But beware that it's not just a one liner. It likely requires lots of code to get completely right
10
General / Re: Slow copying of small structures
« Last post by marcov on Today at 03:53:00 pm »

By crafting the expression (see my last 4 lines) you can actually remove the REPs, but that's not the point. Generally, the infix operator notation will always require temporaries, hence operators are suboptimal. 

For simple types those temporaries could be optimized out when inlined
Pages: [1] 2 3 ... 10

TinyPortal © 2005-2018