In your opinion, could it be of any use in the project if I tried to merge
ccache into gcc, so as to assess this "intelligent agent" approach ?
I can't really tell if, as for an underlying object-oriented structure that
is already present for serialization technique in gcc's source code, despite
it being written in C, gcc is "sort of" already build as an intelligent
agent or if it could be worth clearly showing it in the modular structure of
gcc.
It could lead to some comfort for the community in understanding and
thinking on the project, as introducing intermediate languages, like gimple,
RTL and such, did.
I haven't completely tested ccache yet, but I still expect big compilations
to perturb my PC (freezing, many disk accesses). If part-time contributors
to open-source projects are like me, they may also enjoy the opportunity to
have a very silent background - yet not too slow - compilation, while they
work on other tasks.
Stealth is also one of the reasons why intelligent agents are successful.
If I were to try this, should I work from a tagged version from the trunk,
or is there a branch I should ask to participate into ?
Thank you.
