FlashFXP Forums - View Single Post

bigstar · 08-28-2011, 01:34 PM

FlashFXP 4.1 Build 1639 (RC 3)

This is a critical update, please install this update ASAP.

I discovered 3 separate cases where the application memory was corrupting itself.

A calculate folder size thread we use to calculate the size of a folders content for upload, when communicating with the main thread sometimes the data object being passed was invalid because the main thread had already freed it.

Remote server search, the sub-search is threaded and the thread was missing critical section around a data structure which could cause corruption.

The routine used for generating the information bar gradient background was incorrectly reading the size of the bitmap, corrupting the content of the memory following the bitmap.

In addition I discovered several variable data type overflows. Where the size of the data being assigned to the variable was larger than the maximum allowed by the variable type.

These defects are not typical and I was caught off guard when my regular debugging tools and techniques did not detect them.

08-28-2011, 01:34 PM
bigstar FlashFXP Developer Join Date: Oct 2001 Posts: 8,012	FlashFXP 4.1 RC 3 build 1639 FlashFXP 4.1 Build 1639 (RC 3) This is a critical update, please install this update ASAP. I discovered 3 separate cases where the application memory was corrupting itself. A calculate folder size thread we use to calculate the size of a folders content for upload, when communicating with the main thread sometimes the data object being passed was invalid because the main thread had already freed it. Remote server search, the sub-search is threaded and the thread was missing critical section around a data structure which could cause corruption. The routine used for generating the information bar gradient background was incorrectly reading the size of the bitmap, corrupting the content of the memory following the bitmap. In addition I discovered several variable data type overflows. Where the size of the data being assigned to the variable was larger than the maximum allowed by the variable type. These defects are not typical and I was caught off guard when my regular debugging tools and techniques did not detect them.