Eric Blake wrote: > Do we need to worry about deleting *.manifest files at the appropriate > 'make *clean' level?
Hmm. Yes, the one in the build directory (for the wrapper, or the "real" exe if building without dependencies on uninstalled DLLs) probably should be. The one in .libs/ not so much, because the catchall rm -rf .libs will take care of it. >> + # Create the UAC manifests first if necessary >> + case $output_name in >> + *instal*|*patch*|*setup*|*update*) >> + func_emit_exe_manifest > $cwrapper.manifest >> + func_emit_exe_manifest > >> $output_path/$objdir/$output_name.exe.manifest > > Would it be more efficient to use tee or cp, rather than two calls to > func_emit_exe_manifest? Well, you're not guaranteed to have tee on MSYS. The current code is two forks: each function call forks a 'cat' process. Either of your proposals is also two forks: one 'cat', and one 'tee' or 'cp'. So I don't think it makes much difference. > Does the .manifest need +x permissions? I seem to recall that it does, but I'm not sure. -- Chuck
