On Mon, 1 May 2017 11:53:55 +1200 Graham Ward via Lazarus <lazarus@lists.lazarus-ide.org> wrote:
>[...] > I recently had a software/compilation bug whereby no exception was > raised when my code was accessing an array outside of its range. Enable runtime range checks: Project / Project Options / Compiler Option / Debugging / Checks / Range Or have a "debug" and "release" mode: http://wiki.lazarus.freepascal.org/IDE_Window:_Compiler_Options#Adding_a_release_and_debug_build_modes Mattias Took > some detective work, as it was resulting in corruption at runtime with > completely unrelated code. This was caused by a bug in my own software > that tried to access index 72 for an array of range [0..71]. What was > more interesting was that if I expanded the array to 72 (or 73,74,75 for > that matter) and then tried to access range+1 (IE index 73 for the array > with an index of [0..72] this appeared to cause no ill effect to my > program once more, so I was extremely "lucky" to have such an obvious > bug appear in my code. Something more subtle would have been a lot more > difficult to locate. > > The offending code is as follows: > > setlength(isMerged, ParityCol); // fixed length for array keeping > track of merged cells. > setlength(ColIndex, ParityCol); // Define the length of the Column > Index > For J := 0 to ParityCol do > Begin > isMerged[J] := False; //This line was leading to the program > becoming unresponsive in another totally unrelated part of code > (AssignFile...) in another procedure > ColIndex[J] := 1000; // This line, although also accessing outside > of the legal range, did not appear to cause any problems!!! > end; > > Note I even tried the following line, which is obviously wrong (as the > array has a maximum index of 71) > > ColIndex[175] := 500; > > In other languages (that I am more familiar with) an exception (Error : > Array index outside of range) would have been raised. Why does the > Lazarus compiler not raise a similar error? > > I am using: > > Code Typhon Version 6.0 > FPC Version 3.1.1 > SVN Revision : 54036 > > > Assuming this is an actual bug in the compiler, then I hope this > information proves useful. If not, then please advise me accordingly, > thanks. > > Aside from this, let me say a very big thank you to you guys for all > your hard work in what is a very useful piece of software for me (FPC). > Your efforts are greatly appreciated. > > Kind Regards, > Graham > New Zealand -- _______________________________________________ Lazarus mailing list Lazarus@lists.lazarus-ide.org http://lists.lazarus-ide.org/listinfo/lazarus
