Changes between Version 1 and Version 2 of CMakeTODO


Ignore:
Timestamp:
May 22, 2008, 12:07:09 PM (14 years ago)
Author:
Douglas Gregor
Comment:

Some more TODO items

Legend:

Unmodified
Added
Removed
Modified
  • CMakeTODO

    v1 v2  
    33These are some features that we would like to implement in the CMake-based build system for Boost (in no particular order).
    44
    5   * For each library, automatically introduce a test that uses the "inspect" program to verify that the library meets all of Boost's coding and licensing requirements (and anything else that inspect checks).
     5  * Regression testing:
     6    * For each library, automatically introduce a test that uses the "inspect" program to verify that the library meets all of Boost's coding and licensing requirements (and anything else that inspect checks).
     7  * Documentation generation:
     8    * Port the BoostBook toolchain over to CMake. Bonus points if it can use CMake to download the appropriate XSLT stylesheets and DocBook DTD using file(DOWNLOAD...), therefore taking the role of the BoostBook configuration scripts, too.
     9    * Installation of generated documentation.
     10  * Extend [wiki:CMakeLibraryProject boost_library_project] to accept optional description, author, and maintainer arguments, and possibly other descriptive arguments that might be helpful for libraries (main documentation URL?). Having this information within the CMake system should make it easier to provide this information to users in various ways, e.g., a per-module description in the Windows installer.
     11  * Determine how the "tools" directory should fit in with "libs". Right now, "tools" is disabled because it doesn't work well with modular libraries.
     12  * Building and testing against an installed tree:
     13    * Allow a single Boost component's sources to be downloaded and built against an installed Boost
     14    * Allow regression tests to be built and executed against an installed Boost
     15    * Update regression-testing scripts to simulate a user's usage: build a Boost tree, build its installer, execute the installer, build the regression tests against the installed version (separately!), run regression tests and report the results.
     16  * "Modularize" the rest of the Boost libraries
     17  * Testing on more platforms and tweaks for those platforms