Hi Phil,
> Greg, can I ask what the specific call order to plplot is and how do
> you end up making use of the buffer in the first place?
I can't identify where the plot buffer is used, I thought it
was used routinely now in 5.11.
> For example do
> you plot with gcc then the error occurs during a resize?
The previous message should clarify this. The window is holding
a mix of painted btmaps and line drawings. somehow the bitmap gets buried.
First the line drawings are made, then Saturn image is put on via PaintImage
which calls directly SetDIBitsToDevice(hdc , ...).
> Has plinit
> already been called (directly or indirectly)? Do you do something a
> bit more unusual such as initially set the buffer to null, do the
> plotting to generate the buffer, then use plcpstrm() to copy to the
> gcc driver?
there is no plcpstrm call in the code at all. Images are put up directly
via SetDIBitsToDevice(hdc, ...) from an internally maintained buffer tv_buf
through GDLWINSTREAM::PaintImage and it is redrawn from ::RedrawTV()
These functionalities have only recently been adapted in the gdlwin driver.
I think I need to expose the GDL/plplot connection, which starts at the
general
class for all GDL devices
>
>> class GDLGStream: public plstream
>
> {
>
> // protected:
>
> // void init(); // prevent plstream::init from being called directly
>
> private:
>
> gdlpage pageLayout;
>
> gdlbox boxLayout;
>
>
>
> protected:
>
> bool valid;
>
> gdlCharInfo theCurrentChar;
>
> gdlCharInfo theDefaultChar;
>
> int gdlDefaultCharInitialized;
>
> gdlbox theBox;
>
> gdlpage thePage;
>
> PLStream* pls;
>
> DFloat thickFactor;
>
>
>
> public:
>
>
>> GDLGStream( int nx, int ny, const char *driver, const char *file=NULL)
>
> : plstream( nx, ny, driver, file), valid( true), thickFactor(1.0)
>
> {
>
> if (!checkPlplotDriver(driver))
>
> ThrowGDLException(std::string("PLplot installation lacks the
>> requested driver: ") + driver);
>
> gdlDefaultCharInitialized=0;
>
> thePage.nbPages=0;
>
> theBox.initialized=false;
>
> plgpls( &pls);
>
> if (GDL_DEBUG_PLSTREAM) printf(" new GDLGstream( %d , %d , %s ):pls=%p
>> \n", nx, ny, driver, (void *)pls);
>
>
>> }
>
>
>> any supported device has a specifically programmed GraphicsDevice which
for windows:
>
>> class DeviceWIN : public GraphicsDevice
>
> {
>
> private:
>
> static LRESULT CALLBACK _CallWndProc(int nCode, WPARAM wParam, LPARAM
>> lParam);
>
> LRESULT CallWndProc(int nCode, WPARAM wParam, LPARAM lParam);
>
> static LRESULT CALLBACK _GetMsgProc(int nCode, WPARAM wParam, LPARAM
>> lParam);
>
> LRESULT GetMsgProc(int nCode, WPARAM wParam, LPARAM lParam);
>
>
>> std::vector<GDLGStream*> winList;
>
> std::vector<long> oList;
>
> long oIx;
>
> int actWin;
>
> int decomposed; // false -> use color table
>
> .. etc...
>
> a window needs a specifically programmed stream class which for windows:
>
>> class GDLWINStream : public GDLGStream
>
> {
>
> private:
>
> //Atom wm_protocols;
>
> //Atom wm_delete_window;
>
> HWND refocus;
>
>
>> PLStream* pls;
>
> plstream *plst;
>
>
>> tv_buf_t tv_buf;
>
> int _mode;
>
> PLGraphicsIn *_gin;
>
> POINT GinPoint;
>
> HCURSOR CrosshairCursor;
>
> bool rbutton, xbutton, mbutton, buttonpressed;
>
> public:
>
> std::map<UINT, void (CALLBACK GDLWINStream::*)(UINT, WPARAM, LPARAM)>
>> msghooks;
>
>
>> GDLWINStream(int nx, int ny) :
>
> #ifdef USE_WINGDI_NOT_WINGCC
>
> GDLGStream(nx, ny, "wingdi")
>
> #else
>
> GDLGStream(nx, ny, "wingcc")
>
> #endif
>
> {
>
> pls = 0;
>
> // get the command interpreter window's handle
>
> // plsetopt("drvopt","nofocus"); // avoid stealing focus on window
>> creation
>
> // plsetopt("drvopt","text"); // use freetype fonts
>
> refocus = GetForegroundWindow();
>
> }
>
>
>> ~GDLWINStream();
>
> void Init();
>
> void EventHandler();
>
>
> so this is happening:
>
> void GDLWINStream::Init()
{
this->plstream::init();
plgpls(&pls);
#ifdef USE_WINGDI_NOT_WINGCC
#else
wingcc_Dev* dev = (wingcc_Dev *)pls->dev;
dev->waiting = 1;
#endif
UnsetFocus();
tv_buf.has_data = false;
> HINSTANCE hInstance = (HINSTANCE)GetModuleHandle(NULL);
CrosshairCursor = CreateCursor(hInstance, 15, 14, 32, 32,
> ANDmaskCursor, XORmaskCursor);
}
>
--------------------
So the action goes as follows, for the WINDOW,# command;
> bool success = actDevice->SetBackingStore(retainType);
> success = actDevice->WOpen( wIx, title, xSize, ySize, xPos, yPos);
bool DeviceWIN::WOpen( int wIx, const std::string& title,
>
> int xSize, int ySize, int xPos, int yPos)
>
> {
>
> winList[ wIx] = new GDLWINStream( nx, ny);
>
// set initial window size
>
PLFLT xp; PLFLT yp;
>
PLINT xleng; PLINT yleng;
>
PLINT xoff; PLINT yoff;
>
winList[ wIx]->plstream::gpage( xp, yp, xleng, yleng, xoff, yoff);
>
...
winList[ wIx]->spage( xp, yp, xleng, yleng, xoff, yoff);
> // no pause on win destruction
winList[ wIx]->spause( false);
if(debug) cout << " WOpen: ->fontld( 1) WOpen: ->scolor( ";
> // extended fonts
winList[ wIx]->fontld( 1);
> // we want color
winList[ wIx]->scolor( 1);
>
PLINT r[256], g[256], b[256];
actCT.Get( r, g, b);
// winList[ wIx]->scmap0( r, g, b, actCT.size());
winList[ wIx]->scmap1( r, g, b, ctSize);
> if(debug) cout << "; WOpen:winList[ wIx]->Init(";
> winList[ wIx]->Init();
...
winList[ wIx]->ssub(1,1);
>
> winList[ wIx]->adv(0);
>
> // load font
>
> winList[ wIx]->font( 1);
>
> winList[ wIx]->vpor(0,1,0,1);
>
> winList[ wIx]->wind(0,1,0,1);
>
> winList[ wIx]->DefaultCharSize();
>
>
>
On Sun, Jan 31, 2016 at 6:51 AM, Phil Rosenberg <p.d.rosenb...@gmail.com>
wrote:
> Hi Greg
> I appreciate your frustrations. There is clearly a bug somewhere. t is
> just a case of finding it and deterimining which side of the
> GDL/plplot boundary it lies on - or if it is a Plplot documentation or
> error checking issue. Clearly any call to plplot which results in a
> segfault/crash of any sort is bad.
>
> For some reason I can't build the gcc driver on my system (Windows
> 10), so I am doing this with the wxWidgets driver
> >>
> >> Ok I can simply build plplot without the bop() in rdbuf_bop and now I
> ask,
> >> which example does it affect in linux?
>
> I have just stepped through the function calls with the wxWidgets
> driver in the Visual Studio debugger.The only thing that the plbop()
> call at the end of rdbuf_bop() does is call plP_subpInit(), this is
> because plsc->page_status is set to AT_BOP already. In this case all
> that is actually done is to check that the number of sub pages in each
> direction is > 0, set the current subpage to 0 and set the text and
> tick mark sizes. Actually it turns out there is a bug here and the
> text size is not set correctly for x03 because the orientation change
> is not saved which affects the scale. I am looking to fix that now.
>
> Regarding the specific bug with GDL and gcc driver. The gcc driver
> should not actually called by the plbop() call. It would only get
> called if plsc->page_status was not already AT_BOP.
>
> Greg, can I ask what the specific call order to plplot is and how do
> you end up making use of the buffer in the first place? For example do
> you plot with gcc then the error occurs during a resize? Has plinit
> already been called (directly or indirectly)? Do you do something a
> bit more unusual such as initially set the buffer to null, do the
> plotting to generate the buffer, then use plcpstrm() to copy to the
> gcc driver?
>
> Phil
>
------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140
_______________________________________________
Plplot-devel mailing list
Plplot-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/plplot-devel
