Phase 1/2 - Help wanted - Please help me test a fix to ENIGMA's font problem

[edsquare]

Ok on a ATI [Radeon X1200] no problem with or without fix.

[Darkstar2]

Right ok appreciate the results and feedback.  So one thing that is evident, those who see the problem with AMD also see it properly with the FIX.  And those who don't have the issue with AMD see the same proper display fix or no fix.  This is what I'm aiming - this is what I will focus on. I'm not a hardware engineer and I honestly don't  care who's fault is it.   These kinds of problems are not specific to ENIGMA only as many commercial games have issues of their own and release patches to fix things.  Look at the history of software releases and updates to address specific graphic cards, chipsets, etc.   Clearly this is not just limited to fonts as I noticed that even without applying ANY offset ANYWHERE, there was glitching with vertical scrolling, that I got fixed too, so it's more than just fixing GSFont.cpp and GLMatrix.

And Harri, yes if not mistaken I saw these issues with GL3 as well, but one thing at a time. 

[Goombert]

Now, anyway, this was a good job Darkstar, we now have a better idea of where the issue occurs and where it does not. Is that AMD card it occurred on x86 or x64?

[Darkstar2]

Now, anyway, this was a good job Darkstar, we now have a better idea of where the issue occurs and where it does not. Is that AMD card it occurred on x84 or x64?

I think you mean x86 or x64

Job is only half done.

I have been extremely busy and everything is crazy around me lately so did not have time to complete things and things trailed a bit.......... Today I will conclude this with a new test I will submit, there needs to be one more change to the fix as I discovered 1px gaps on seams in vertical scrolling.   Interestingly this was NOT caused by the fix itself and even happened without any fix or any offsets at all in the CPPs, so I added another element to patch this.
According to my testing here, everything is now 100% - will submit later on for testing by everyone.
This 2nd project I made is more complete, tests scrolling H/V, border issues, and I cherry picked the most problematic fonts, sizes, character combos where artifacts could be found.  Hopefully this will be all behind us soon.

Currently the latest build of ENIGMA includes erroneous offsets that were not removed, these offsets present in the GLMatrix.cpp add + 0.375/0.375 to the d3d projection ortho and + 0.175/0.175 to the GL MODEL VIEW, these fixes were deemed ineffective for certain people.  I don't know why they were left there as I was testing the whole time without any offsets at all + the new updated fixes.

So please don't compile from my source EGM ! I will remove this for the time being.  I will submit another EXE, this was compiled from a source with all previous offsets removed and the new ones applied.

In theory, since the glitches people notice are 1px, even a tiny offset like the one tested might be helping to round things whilst at the same time not causing any adverse effects on GPUs that did not have this problem. 

[Darkstar2]

testing phase 1/2 complete, thanks to all who contributed and your feedback, the results are positive.  The tests passed on AMD, INTEL, NVIDIA, I have older boards here I tested with onboard video all worked as well.

A new up to date topic will be opened later with a final testing I will submit, which includes a revised fix that should work for everyone.  I have tested it extensively offline and it proved successful.

I will be locking this topic once I submit the new EXE in its new topic. 

The first fix was positive for all, however I discovered gaps in vertical scrolling when I was working on my 2nd test program.  Oddly, this vertical glitch had NOTHING to do with the fix as it was present even without any offsets, so I added an element to the fix to fix all these things.

In another topic later today I will submit a new EXE for you to try.  If it has positive results as I'm sure it will, then it will be up to the devs to merge it.

For anybody who still has not run this EXE, please wait for the 2nd updated version.

I will also be submitting later on a fix for the DX9 issues with display clearing when swtching FS/window during a game or losing focus and back. This will work for projects that have resources loaded in RAM.  For those using *add and external resources, I know how to fix this but it's more complex, can be done through scripts and is more advanced so for now I will focus on the 1st option since it is what people use the most.
Until this is integrated into CPP I have made it into a script, uses C++, so this putting ENIGMA's ability to use C++/EDL/GML in the same project to good use.

I will submit the idea for approval as well to see what the devs think and whether it should be integrated into the DX9 CPPs.  So far it made DX9 fully functional, switching full screen, window, going in and out of focus, and it's as if the problem was never there, seamless!  and no it does not involve room restart or game restart,

Both fonts/scrolling and DX9 fixes have been tried with 100% success on my end as well.

I can now confirm that the problems concerning fonts/artifacts, scrolling, etc occur in OpenGL1/3,
DX9 does not exhibit these strange behaviours and so require no offset fixing.

More on this later.

Meanwhile I am locking this topic,
please refer to this new topic which is up to date:
http://enigma-dev.org/forums/index.php?topic=2022.0