On 13 March 2012 09:07, Don Clugston <d...@nospam.com> wrote: > On 12/03/12 01:20, Walter Bright wrote: >> >> On 3/11/2012 2:57 PM, Caligo wrote: >>> >>> And just for the record, there are software projects that are millions >>> of lines of code in C/C++ and have ZERO workarounds. Also, I have >>> never encountered a bug in GCC when programming in C++, even when >>> trying out the latest C++11. >> >> >> GCC itself is fairly bug free, > > > I had used g++ for a grand total of three hours before I found a wrong-code > regression -- the absolute worst category of bug. I've never found a DMD bug > which was significantly worse than that one. > > GCC is much more mature than the DMD front-end, of course, but I don't think > it's any more bug-free than the DMD back-end. > > > but then again I don't push it that hard.
The only bug I've been bitten by is bug #323 - http://gcc.gnu.org/bugzilla/show_bug.cgi?id=323 Which has only recently been somewhat addressed in the latter versions of GCC (though requires frontend language support that I haven't decided whether it's worth adding - so is still technically a bug in GDC ;) Other than that - I've found that GCC is *very* good at finding bugs in the DMD Frontend codegen that I have both raised and fixed (or fixed by Walter and co) since I picked up the project. This is in part due to it's strict tree checking in development builds. Regards -- Iain Buclaw *(p < e ? p++ : p) = (c & 0x0f) + '0';