recompiled an 2012 appl gives runtime error lazarus / 13-General ---------------------------------------------------------------------- fikkies · 2019-07-20 19:28 I build a (rather complex) application in 2012 and I still use it. It was build in 32 bits and also workes fine on recent 64 bit (Linux Mint) Of course after 7 years some changes are nessecary and no problem compiling but when I run I get -- Acces Violation Press OK to ignore -- before the app itself starts, and the application stops. on Debugging I get - for me unknown -  assembler codes. Main question. Is there an old error  in the application or is the Lazarus 2012 version not (backward) compatible to the 1.8.2  version. Frits ---------------------------------------------------------------------- Thaddy · 2019-07-20 19:34 On some points, both the compiler and Lazarus are not fully backwards compatible. The compiler issues are usually minor and documented in the release notes, including work-arounds tips how to modify your code. They can have a major effect, though, if you are not aware of those changes. Lazarus is (was) less well documented on breaking changes. BTW: same issues exist with e.g. Delphi. The most likely is, however, that there was already a bug in the existing code that is now revealed because of the compiler being more strict. (e.g. less forgiving in the case of use-after-free) ---------------------------------------------------------------------- fikkies · 2019-07-20 19:49 thanks Any way, I will have to break up all the modules and compile - debug them one by one,  starting with the database. frits ---------------------------------------------------------------------- af0815 · 2019-07-21 07:02 You can still create an old compiler side by side and recompile with it. ---------------------------------------------------------------------- Martin_fr · 2019-07-21 09:31 When you get the asm code, have you tried the stack window? And the thread window (changing to each thread, watching the stack) ---------------------------------------------------------------------- marcov · 2019-07-21 11:57 I wouldn't really phrase it as the newer versions necessarily "more strict". That implies deliberation, which it is probably not. More a result of chance memory layouts. ---------------------------------------------------------------------- Thaddy · 2019-07-21 15:43 Quote from: marcov on July 21, 2019, 01:57:05 pmMore a result of chance memory layouts. With which you had a higher chance of succeeding on dangling pointers before 3.0.0. Hence my comment.