This is the mail archive of the guile@cygnus.com mailing list for the guile project.
| Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
|---|---|---|
| Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
Chris.Bitmead@misys.com.au writes: > Is it possible perhaps that the interface to gc could be well enough > defined that the user > could choose which gc to use? Ideally you could say (use-modules (gc > incremental)) > and get the sort of gc you want. Such an interface would also make it easy > for people > to experiment with different sorts of gc. > > Knowing how hairy gc can be though, I don't know how feasable that would > be. > At the very least, it will be possible to select gc's at compile time. It might be a good idea to constrain any runtime selection to start up time, rather than any time, because different gc's require different information to be associated with objects in order to work, and may require other changes throughout the system (a generational gc not using memory protection would need other bits doing some additional work, for example). As it stands, the interface to the gc (from what I've seen so far, I'm pretty sure I'll find something soon that doesn't want to play nice) is fairly well defined. The hairiest bit is gc marks, but that should be mostly painless if code sticks to the currently provided macros & functions. -- Greg