Modularizing GCC Gets Discussed, Again
There's yet another conversation taking place within the GCC camp about modularizing the GNU Compiler Collection.
A student developer started the conversation about modularizing GCC on the mailing list after seeing it listed as a possible Google Summer of Code 2015 work item. There's been the Modular GCC Wiki Page for some time as it's long been a high goal to modularize GCC, which would effectively on an architectural level make it much more like LLVM. This discussion has happened many times before and was once considered a goal for GCC 5.0, but due to the significant investment needed to make this a reality, it hasn't materialized yet.
The reasoning for modularizing GCC is that it's grown to become too big as a monolithic compiler package and over its many years of development a lot of crusty code has built up and thus makes the barrier of entry to new developers very high -- certainly much higher than LLVM with its modular design and modern code-base. With being more modular and cleaning things up, it would be easier for new contributors and also open up the possibilities of GCC. It's only with the upcoming release of GCC 5 that GCC JIT is a reality.
Modularizing GCC in whole is far from a Google Summer of Code project that could be done in the course of a few months by a new contributor to the compiler, but the developers commenting on the discussion say it'd be more reasonable to just pick one sub-area of the massive project to focus on over the summer.
A student developer started the conversation about modularizing GCC on the mailing list after seeing it listed as a possible Google Summer of Code 2015 work item. There's been the Modular GCC Wiki Page for some time as it's long been a high goal to modularize GCC, which would effectively on an architectural level make it much more like LLVM. This discussion has happened many times before and was once considered a goal for GCC 5.0, but due to the significant investment needed to make this a reality, it hasn't materialized yet.
The reasoning for modularizing GCC is that it's grown to become too big as a monolithic compiler package and over its many years of development a lot of crusty code has built up and thus makes the barrier of entry to new developers very high -- certainly much higher than LLVM with its modular design and modern code-base. With being more modular and cleaning things up, it would be easier for new contributors and also open up the possibilities of GCC. It's only with the upcoming release of GCC 5 that GCC JIT is a reality.
Modularizing GCC in whole is far from a Google Summer of Code project that could be done in the course of a few months by a new contributor to the compiler, but the developers commenting on the discussion say it'd be more reasonable to just pick one sub-area of the massive project to focus on over the summer.
18 Comments