I think the logic is sound, and the hash is useful for truncating long
paths, but IMO you could skip the CM_<ID_TYPE><SCHEME_TYPE>_ part.
The SCHEME_TYPE is apparent from the presence of a hash, which anyway
makes the id non-deterministic and unusable from other code. What does
the CM_ prefix mean, btw?
The hashes are deterministic. You should always get the same hash for
the same path.
"CM_" (CMake) is intend for cmake reserved IDs which will leave IDs not
starting with CM_ to users.
The explicit SCHEME_TYPE prevents you from creating a filename based on
an existing hash which would conflict with a hash generated ID.
IMO, the prefix is redundant since the IDs will be used in context,
e.g.: <ComponentRef Id="share.cmake_2.8.Modules.BundleUtilities.cmake"
/> or <Property Id="CMAKE_BINARY" Value="[#bin.cmake.exe]" />.
Simple is better. Finally, I would not replace '-' in file names with
'_'.
<ID_TYPE> is primarily for me since I expect it makes IDs easier to
trace when things don't work as expected/intended (plus that one extra
character doesn't really hurt either).
Acording to the documentation on msdn "-" is not legal in identifiers.
Legal are A-Z a-z 0-9 . and _.
Nils
--
Powered by www.kitware.com
Visit other Kitware open-source projects at
http://www.kitware.com/opensource/opensource.html
Please keep messages on-topic and check the CMake FAQ at:
http://www.cmake.org/Wiki/CMake_FAQ
Follow this link to subscribe/unsubscribe:
http://public.kitware.com/cgi-bin/mailman/listinfo/cmake-developers
