Overview
Examples
Screenshots
Comparisons
Applications
Download
Documentation
Tutorials
Bazaar
Status & Roadmap
FAQ
Authors & License
Forums
Funding Ultimate++
Search on this site
Search in forums












SourceForge.net Logo
Home » U++ Library support » Draw, Display, Images, Bitmaps, Icons » Strange issue with text in Painter
Re: Strange issue with text in Painter [message #51060 is a reply to message #51058] Thu, 17 January 2019 14:30 Go to previous messageGo to previous message
Tom1
Messages: 698
Registered: March 2007
Contributor
mirek wrote on Thu, 17 January 2019 15:16
Tom1 wrote on Thu, 17 January 2019 14:12
mirek wrote on Thu, 17 January 2019 14:58
...
It starts to look like Win32 does not like the sharing of HDC... (despite of what the documentation says).

What about this:

void RenderCharacterSys(FontGlyphConsumer& sw, double x, double y, int ch, Font fnt)
{
	RLOG(CoWork::GetWorkerIndex() << " RenderCharSys " << fnt << " " << ch << " " << (char)ch);
	static Atomic h;
	h++;
	HFONT hfont = GetWin32Font(fnt, 0);
	VERIFY(hfont);
	if(hfont) {
		HDC hdc = CreateIC("DISPLAY", NULL, NULL, NULL);
		HFONT ohfont = (HFONT) ::SelectObject(hdc, hfont);
		GLYPHMETRICS gm;
		MAT2 m_matrix;
		memset(&m_matrix, 0, sizeof(m_matrix));
		m_matrix.eM11.value = 1;
		m_matrix.eM22.value = 1;
		int gsz = GetGlyphOutlineW(hdc, ch, GGO_NATIVE|GGO_UNHINTED, &gm, 0, NULL, &m_matrix);
		if(gsz < 0) {
			LOGFONT lf;
			int ret = GetObject(hfont, sizeof(lf), &lf);
			RLOG("Failed " << CoWork::GetWorkerIndex() << " RenderCharSys " << fnt << " " << ch << " " << (char)ch
			               << ", real: "  << " " << lf.lfFaceName << " " << lf.lfHeight);
			RDUMP((void *)hdc);
			RDUMP((void *)hfont);
			RDUMP(GetLastErrorMessage());
		}
		VERIFY(gsz >= 0);
		if(gsz < 0)
			return;
		StringBuffer gb(gsz);
		gsz = GetGlyphOutlineW(hdc, ch, GGO_NATIVE|GGO_UNHINTED, &gm, gsz, ~gb, &m_matrix);
		if(gsz < 0) {
			LOGFONT lf;
			int ret = GetObject(hfont, sizeof(lf), &lf);
			RLOG("Failed2 " << CoWork::GetWorkerIndex() << " RenderCharSys " << fnt << " " << ch << " " << (char)ch
			               << ", real: "  << " " << lf.lfFaceName << " " << lf.lfHeight);
			RDUMP((void *)hdc);
			RDUMP((void *)hfont);
			RDUMP(GetLastErrorMessage());
		}
		VERIFY(gsz >= 0);
		if(gsz < 0)
			return;
		RenderCharPath(~gb, gsz, sw, x, y + fnt.GetAscent());
		::SelectObject(hdc, ohfont);
		::DeleteDC(hdc);
	}
	h--;
	VERIFY(h == 0);
}



As far as I can see, not a single bad or missing letter in 8 minutes of systematic intensive scanning of names. No VERIFYs/ASSERTs either. The log is attached.

BR, Tom


Well, so be it, lets call it a fix. I will remove shared information context from the trunk... There will be small performance penalty, but it should be negligible (as there are caches on many levels anyway).

Hopefully my interpretation is right and we are not just hiding some other bug...


OK, I hope so too. Using Windows can be expensive sometimes...

As for the 'rounding of tolerance' we discussed about one hundred (or so) postings ago, should it be brought in rather sooner or later? I would like to test this current solution without other changes for a while before taking in an optimization that will decrease likelihood for error. (Actually, I believe this will solve it for good, but it's better keep an eye on it for a while... This does not 'add to cost' as I will anyway look at the map during testing of new features I'm building on top of this.)

Of course I can take this fix in now, and then postpone updating the tolerance rounding, until a week or so have passed...

In any case, Thank you Mirek for all your help and support!

Best regards,

Tom
 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic: MT + Subpixel appear incompatible in Painter
Next Topic: Painter: Excessive memory usage in PainterExamples moving window to edges of screen.
Goto Forum:
  


Current Time: Wed Apr 01 00:30:09 CEST 2020

Total time taken to generate the page: 0.01751 seconds