On Apr 26, 2008, at 8:11 AM, Daniel Kolbo wrote:
> The sh.exe in the PATH variable was indeed the culprit. Even though
> the path to sh.exe was after the g++ compiler in the PATH, it still
> was ostensibly causing problems.
MinGW's version of GNU make has special code in it to search for a
Bourne shell and use it for shell commands if it's available, falling
back on cmd.exe only as a last resort.
One can argue about how much sense this makes, but one will refrain. :)
> I don't quite understand the difference between a release version
> and a debug version.
You can "re-bake" Makefile.mingw with Bakefile to change certain
settings to get a release build. This will optimize the library code
for speed, and potentially remove some bits of debug-only code,
resulting in a smaller, faster library.
Not that it really matters, because MySQL++ is never the bottleneck.
Since MinGW make doesn't make building separate debug and release
versions the easy default -- as do VC++ and Xcode, by contrast -- why
bother?
