At first, bug of screen.inc is bogus even in fpc 1.0.7. Anyway, it is a very bad idea to place const parameter in interface and non-const in the implementation (or vice vercia).

Right. Where are they?

Lazarus for old stable fpc 1.1 have some more non-fatal bugs (I have fixed them in 2 min). There are: TRtlCriticalSection is not in standard unit, but in somewhat like SysThrds.

I hope, this was answered by Peter.

And second, I had to replace Index to AIndex in property declaration.

Sounds like objfpc. The fpc object pascal mode is far less ambigious than Delphi/Kylix. That's why it is preferred in lazarus.

The attachment was probably too big for the list. Can you send it to me personally?

Mattias

At first, bug of screen.inc is bogus even in fpc 1.0.7. Anyway, it is a very bad idea to place const parameter in interface and non-const in the implementation (or vice vercia).
Lazarus for old stable fpc 1.1 have some more non-fatal bugs (I have fixed them in 2 min). There are: TRtlCriticalSection is not in standard unit, but in somewhat like SysThrds.
And second, I had to replace Index to AIndex in property declaration. As for my project, as I see, I have attached it to the file. It was for fpc-1.1 only (translated from kylix 2 oe), but I have made it compartible to fpc 1.0.7, but the diskette was bad :(. Errors are not reproducable. FPC 1.0.7 can result in program sudden dearth (I can't locate a reason with gdb). May be, it can result in strange Round error, but I have found it only with fpc-1.1 (I haven't played with 1.0.7 enough).
More problems is that the project is Russian (I remembered this fact right now, when I am on non-linux University PC). So, click top left of four buttons (������������ ������������ means experiment planning). Then you will need to enter "2" in spinedit (besides, floats can't be edited correctly in OI), and to click OK twice . This is a probable place of error (may not occur), otherwise click right top button.

At first, there is a bug in forms.pp/screen.inc - header of somewhat like SetCorsors method doesn't match (const/not const parameter).

Sounds like an 1.1 bug. I will check it, as soon as 1.1. compiles again.

Second, TTreeView doesn't work with font - so charset is poor.

That's probably something for the bugtracker.

Hmm, looks like the old float conversion bug of fpc. I thought, this was fixed. Apparently not. It appears on some extrem values, which means, you found also a bug in the LCL.

Can you send me a failing program or describe how I can reproduce the bug?

Mattias

At first, there is a bug in forms.pp/screen.inc - header of somewhat like SetCorsors method doesn't match (const/not const parameter). Second, TTreeView doesn't work with font - so charset is poor. And, at last, my project (attached) have resulted in a very strange bug. Dump is following
#0 0x0810c77b in _SYSTEM_ROUND$EXTENDED ()
#1 0x080aec7d in TGTKOBJECT__SETSCROLLINFO (HANDLE=143870872, SBSTYLE=1,
valSCROLLINFO=
{CBSIZE = 28, FMASK = 103, NMIN = 0, NMAX = 0, NPAGE = 321, NPOS = 0, NTRACKPOS = 0}, BREDRAW=false, this=0x815b384) at gtkwinapi.inc:6786 #2 0x080dd812 in SETSCROLLINFO (HANDLE=143870872, SBSTYLE=1, valSCROLLINFO=
{CBSIZE = 28, FMASK = 103, NMIN = 0, NMAX = 0, NPAGE = 321, NPOS = 0, NTRACKPOS = 0}, REDRAW=false) at winapi.inc:923 #3 0x08105a4e in UPDATEVSCROLL () at scrollingwincontrol.inc:193 #4 0x08105c16 in TCONTROLSCROLLBAR__UPDATESCROLLBAR (this=0x8163724) at scrollingwincontrol.inc:235
#5 0x081061fb in TSCROLLINGWINCONTROL__UPDATESCROLLBARS (this=0x8163014)
at scrollingwincontrol.inc:421
#6 0x081058d4 in TCONTROLSCROLLBAR__AUTOCALCRANGE (this=0x8163754) at scrollingwincontrol.inc:161
#7 0x08105ef7 in TSCROLLINGWINCONTROL__ALIGNCONTROLS (ACONTROL=0x8163a64,
ARECT=
{LEFT = 0, TOP = 0, RIGHT = 673, BOTTOM = 320, TOPLEFT = {X = 0, Y = 0}, BOTTOMRIGHT = {X = 673, Y = 320}}, this=0x8163014) at scrollingwincontrol.inc:341
#8 0x080ed841 in TWINCONTROL__ALIGNCONTROL (ACONTROL=0x8163a64, this=0x8163014) at wincontrol.inc:1706
#9 0x080ed704 in TWINCONTROL__INSERTCONTROL (ACONTROL=0x8163a64, ---Type <return> to continue, or q <return> to quit--- this=0x8163014) at wincontrol.inc:1666
#10 0x080f0d85 in TCONTROL__SETPARENT (APARENT=0x8163014, this=0x8163a64)
at control.inc:1631
#11 0x0805ee0c in TCUSTOMGRID__SETEDITOR (AVALUE=0x8163a64, this=0x8163014)
at grids.pas:876
#12 0x08068b6d in TSTRINGGRID__SELECTEDITOR (this=0x8163014) at grids.pas:4236
#13 0x080632d0 in TCUSTOMGRID__MOVEEXTEND (RELATIVE=false, DCOL=1, DROW=1,
this=0x8163014) at grids.pas:2602
#14 0x08062509 in TCUSTOMGRID__MOUSEDOWN (BUTTON=MBLEFT, SHIFT=Invalid C/C++ type code 10 in symbol table.
)
at grids.pas:2257
#15 0x080f0039 in TCONTROL__DOMOUSEDOWN (MESSAGE=
{MSG = 513, KEYS = 1, XPOS = 229, YPOS = 41, POS = {X = 229, Y = 41}, RESULT = 0}, BUTTON=MBLEFT, SHIFT=Invalid C/C++ type code 10 in symbol table.
) at control.inc:950
#16 0x080f00c8 in TCONTROL__WMLBUTTONDOWN (MESSAGE= {MSG = 513, KEYS = 1, XPOS = 229, YPOS = 41, POS = {X = 229, Y = 41}, RESULT = 0}, this=0x8163014) at control.inc:979 #17 0x08112cc9 in _SYSTEM_TOBJECT_$_DISPATCH$formal () #18 0x080effc2 in TCONTROL__WNDPROC (THEMESSAGE=
{MSG = 513, WPARAM = 1, LPARAM = 2687205, RESULT = 0, WPARAMLO = 1, WPARAMHI = 0, LPARAMLO = 229, LPARAMHI = 41, RESULTLO = 0, RESULTHI = 0},
this=0x8163014) at control.inc:922
#19 0x080ecf62 in TWINCONTROL__WNDPROC (MESSAGE=
{MSG = 513, WPARAM = 1, LPARAM = 2687205, RESULT = 0, WPARAMLO = 1, WPARAM---Type <return> to continue, or q <return> to quit--- HI = 0, LPARAMLO = 229, LPARAMHI = 41, RESULTLO = 0, RESULTHI = 0}, this=0x8163014) at wincontrol.inc:1329
#20 0x0806142d in TCUSTOMGRID__WNDPROC (THEMESSAGE= {MSG = 513, WPARAM = 1, LPARAM = 2687205, RESULT = 0, WPARAMLO = 1, WPARAMHI = 0, LPARAMLO = 229, LPARAMHI = 41, RESULTLO = 0, RESULTHI = 0},
this=0x8163014) at grids.pas:1833
#21 0x0806c62a in DELIVERMESSAGE (TARGET=0x8163014, AMESSAGE=void) at gtkproc.inc:1080
#22 0x08071fd2 in DELIVERMOUSEDOWNMESSAGE (WIDGET=0x8934f80, EVENT=0x8941478,
AWINCONTROL=0x8163014) at gtkcallback.inc:1007
#23 0x0807206e in GTKMOUSEBTNPRESS (WIDGET=0x8934f80, EVENT=0x8941478, DATA=0x8163014) at gtkcallback.inc:1061
#24 0x40232b26 in gtk_marshal_BOOL__POINTER () from /usr/lib/libgtk-1.2.so.0
#25 0x40277e16 in gtk_signal_remove_emission_hook () from /usr/lib/libgtk-1.2.so.0
#26 0x40276f3b in gtk_signal_handlers_destroy () from /usr/lib/libgtk-1.2.so.0
#27 0x40274b8e in gtk_signal_emit () from /usr/lib/libgtk-1.2.so.0 #28 0x402c0ba4 in gtk_widget_event () from /usr/lib/libgtk-1.2.so.0 #29 0x40231a7f in gtk_propagate_event () from /usr/lib/libgtk-1.2.so.0 #30 0x40231224 in gtk_main_do_event () from /usr/lib/libgtk-1.2.so.0 #31 0x40108985 in gdk_flush () from /usr/lib/libgdk-1.2.so.0 #32 0x403093ee in g_idle_remove_by_data () from
/usr/lib/libglib-1.2.so.0
#33 0x40309199 in g_idle_remove_by_data () from
/usr/lib/libglib-1.2.so.0
---Type <return> to continue, or q <return> to quit--- #34 0x403080e3 in g_main_iteration () from /usr/lib/libglib-1.2.so.0 #35 0x40231f37 in gtk_main_iteration_do () from /usr/lib/libgtk-1.2.so.0

#36 0x08096ebf in TGTKOBJECT__HANDLEEVENTS (this=0x815b384) at gtkobject.inc:770
#37 0x08096ff6 in TGTKOBJECT__WAITMESSAGE (this=0x815b384) at gtkobject.inc:828
#38 0x08108c29 in TAPPLICATION__IDLE (this=0x815b274) at application.inc:278
#39 0x0810959c in TAPPLICATION__HANDLEMESSAGE (this=0x815b274) at application.inc:616
#40 0x0810962e in TAPPLICATION__RUN (this=0x815b274) at application.inc:661
#41 0x0811b81a in main () at planexp.dpr:23

-------------- next part --------------
A non-text attachment was scrubbed...
Name: bin00019.bin
Type: application/octet-stream
Size: 12522 bytes
Desc: "planexp.tgz"
Url : http://localhost/pipermail/lazarus/attachments/20030307/2d8fb026/bin00019.bin

forms.pp - TScreen - SetCursors method and screen.inc. May be, I have a bogus cvs?

So, line {$ifndef VER1_0}{$THREADING ON}{$endif} should be added, right?

> And second, I had to replace Index to AIndex in property declaration. Sounds like objfpc. The fpc object pascal mode is far less ambigious than Delphi/Kylix. That's why it is preferred in lazarus.

It's in cvs of Lazarus!

Course. It seems to be attached correctly, I have taken it from mailing list. I'll send it in a minute

Lazarus for old stable fpc 1.1 have some more non-fatal bugs (I have fixed them in 2 min). There are: TRtlCriticalSection is not in standard unit, but in somewhat like SysThrds.

The SysThrds unit is an internal unit that will be included by the compiler when using {$THREADING ON}

Without the {$THREADING ON} switch the threadvar keyword is also not available.