| Commit | Date | |
|---|---|---|
| 2009-08-23 05:58:31 | Tree | |
|
[r2952]
by
david_costanzo
Fix a regression where zooming to a very small number (1e-20) would cause the WM_PAINT handler to coerce a large number into an integer, resulting in an unhandled floating point exception. My fix was to keep the numbers as floating point values until they had been clipped into the proper range before converting them back into ints. |
2009-08-23 03:45:53 | Tree |
|
[r2951]
by
david_costanzo
Refactoring: replace code that called numeric_arg() followed by numeric_arg_to_flonum() with a single call to float_arg(). This was done to make the code more readable, but the resulting code should also be a little smaller/faster. |
2009-08-23 02:05:47 | Tree |
|
[r2950]
by
david_costanzo
Refactoring: rename pos_int_arg() to nonnegative_int_arg() to better reflect what it does. |
2009-08-23 01:44:55 | Tree |
| 2009-08-23 01:00:22 | Tree | |
|
[r2948]
by
david_costanzo
Extend test coverage of FORM, particularly of the bad inputs. This is being done in preparation for some refactoring involving the input parsing for FORM. |
2009-08-23 00:50:38 | Tree |
|
[r2947]
by
david_costanzo
Partially update with the results from the test pass for 6.25.0. FMSLogo crashed unusually early, which I need to investigate, but I have signed off on all of the changes before the point of failure. |
2009-08-22 23:44:35 | Tree |
|
[r2946]
by
david_costanzo
Fix bug #2842770; FMSLogo continues to draw properly after BITFIT is called with very large inputs. The fix was to clean up side-effects on an error path, instead of simply returning from lbitfit(). |
2009-08-22 22:13:54 | Tree |
| 2009-08-22 02:26:50 | Tree | |
| 2009-08-22 02:24:47 | Tree |