inmarket

The trouble here is that you get any two devices from different sources and they will possibly have different alignments between the touch screen and the display screen. To make things worse very few devices are well aligned physically. Also the resolution of the display seldom matches the resolution of the touch panel. What this means is that floating (or fixed) point multiplication is needed in 99% of cases whether or not that is via fixed calibration settings. Despit this we have implemented fixed alignment code as an option for some drivers but that doesn't work for all devices (as you have already discovered) due to the reasons above. In V3 we are looking at providing better fixed alignment code that caters for more orientations but for now the correct solution is either a custom driver hack or, preferably, the fixed calibration matrix.

Have a look in the example makefiles. It should be obvious.

This has now been updated into the repository. Thanks for finding it.

Use our file instead of the firmware file in your build. STM keep changing their firmware in incompatible ways which is why we include a particular version in our repository. @cpu20 is the expert in getting ide's to include the correct files. He has published a number of guides that may help. The other way is to not rely on the ide's build generator and to instead create a true makefile project in eclipse. It is by far the easiest way to build uGFX and it is where we as developers started. If you are nterested in this have a look at some of the prepared makefiles in the various board example directories.

Thanks. I will look closely at that asap and update the repository accordingly.

If the line specified is right at the top of the function my guess is that it will work. If it is not at the top then you are running into a common C, C++ language misconception. C does not allow variable definitions except at the top of a block. Once a non-variable definition statement has occurred variable definitions are not allowed in the block. Whilst the C language does not allow this, C++ does. As many C compilers have associated C++ compilers many C compilers relax the rules and allow such definitions eg gcc. My guess is that the IAR C compiler follows the language standard and doesn't allow this. In general in C it is a good idea to seperate variable definitions from code to set the variables.

It is possible to use RTL languages with ugfx but it's a little more work for the programmer. 1. Use right justification instead of left justification for all your text boxes. 2. Reverse the characters in any string you want to print so they appear in the correct order when displayed. This can either be done programmatically (something like the gtrans module) or simply writing any fixed string with the characters reversed when programming. Eg. "abc def" would be written as "fed cba" and then right justified.

Yes

I presume you are using an M3 with the internal ltdc video controller. If that is not the case the below discussion may help even if it not directly applicable to your circumstances. The problem is likely to be related to your external sdram and it's bandwidth. The proof of this is the slow performance (noticeable redraw speed) of the gdispClear loop.The image drawing tests are therefore meaningless until that problem is solved. The first thing to try is putting the framebuffer into internal ram as a test. If that operates quickly the problem is your bus interface to the external sdram. Note that if your framebuffer is in external ram do not put any other objects, heaps, stacks, code or anything else in that external sdram. Framebuffers are extremely bandwidth intensive just with video refreshing. Other things to check are your CPU clock is running at full speed, wait states and speed and width of the sdram bus and also look for any other high speed requirement that may affect available bandwidth or speed.

Have a look at the CMD_SET_REMAP. The value 0x14 may not match your display.

It looks like the column order on your display might be a bit different to the standard arrangement. Check the initialisation sequence particularly around the column order settings. The other place to look is around the spi data format eg if you are outputting 9 bit spi bit the controller expects 8 bit spi (or visa versa) you will get similar symptoms.

What type of physical interface do you have to your display? Eg i2c, spirit, 8 bit parallel.

More information... Aquire_bus and release_bus give you an opportunity to use a shared bus. Post_init is called once the controller has been initialised. Many controllers need to be initialised at a slower bus speed. Post_unit gives you the opportunity to increase the bus speed after the controller is initialised.

This has now been fixed in the repository. Thanks for finding it!

Unfortunately there are only two mechanisms, off-screen drawing (pixmaps) and double buffering. Double buffering is a big nasty topic in its own right - search the forum for some of the discussions on that subject. It is probably also not viable for you given your controller and bus interface. As far as pixmaps are concerned v3.0 will fix the incompatibility with single file make. The filled string drawing algorithm will be updated at some time in the future too but that is not a simple task as their are lots of complexities with fonts.

I have seen the post and have left it open to sort out when I get time but haven't had that time yet due to various project deadlines. Thank you for finding it and I will report back when I have had time to properly investigate and correct the master sources.

Have a look at the dial devices. A rotary encoder fits fairly nicely with the dial device.

You need to create a gadc driver for OSX or a dummy driver that will work cross-platform. As an example look at the existing gadc driver in the drivers directory.

Note that reparenting a widget is a real hack and is officially not supported by uGFX. The reason is that there are dependencies that may stop things working eg... 1. A child object must always be completely enclosed within its parent container 2. A child object must always have a higher z order than its parent container. 3. There are visibility dependencies. I am sure there are others too but those three alone are enough to make this a complex operation requiring deep understanding of the internal workings of uGFX to get right. That complexity is why it is not a standard part of the uGFX API. It is much better that you arrange your containers and widgets as non-overlapping, or duplicate such controls across the pages. As an example see the console object on the widgets demo.

Unlikely the compile will affect anything. Try just outputting directly a string with Unicode characters in it without using snprintf. If that doesn't work there can only be one of a few problems... 1. Utf8 support is not turned on 2. Your editor is not putting the Utf8 characters into the string correctly 3. Your compiler is stripping them out of the string or not encoding them correctly. 4. The font you are using doesn't map the particular characters used. In your case I think 1 or 2 would be the problem as you are seeing 'extra' characters.

It may be asking the obvious but have you got unicode support turned on in your gfxconf.h?

Great to hear you got it working.

This may indicate a unicode bug in snprintf. I will look later. It could also indicate that your editor is not in utf8 character format. Perhaps it is using extended ASCII or one of the other unicode formats eg utf16le

You are mixing up compiling and linking. uGFX must be compiled with gcc as it is written in C. It will quite happily link with C++ and the header file ugfx.h is fully C++ compatible. If you are using make this is really simple... .c files get compiled with gcc .c++ files get compiled with g++ Link with g++ While C and C++ are similar languages they are not totally the same.

Your memory allocation scheme sounds appropriate to me too. Good thinking.