[m-rev.] [reuse] diff: --promise-no-modified-source-files

Peter Ross peter.ross at miscrit.be
Sat Mar 24 00:05:34 AEDT 2001


On Fri, Mar 23, 2001 at 11:44:00PM +1100, Peter Schachte wrote:
> On Fri, Mar 23, 2001 at 10:30:41AM +0100, Peter Ross wrote:
> > > No source files have been modified?
> > > None, ever?
> > > That seems like a hard thing to promise.
> > > 
> > Good point, no source files have been modified since the last
> > compilation.  Here is the change 
> 
> How hard/expensive would it be to check the time stamp on the source file
> against the time stamp on the transopt file when doing this, detecting this
> condition automatically (avoiding the need for the switch)?  If it's somehow
> a bit performance hit, you could add the switch but at least verify as you
> go.
> 
Unfortunately mmake gets in the way of this optimization.  You change a
.m file and mmake will rebuild the .trans_opt file leading to the time
stamp being more recent.

What needs to happen at this point is to decide whether the trans_opt
file is being rebuilt because there is more accurate information
available in other trans_opt files or because the .m file has been
modified.  This information then needs to be propogated to the rest of
the invocations to mmc so that it knows which of the conditions we are
in.  Not so easy with the current build architecture, of course Simons
work on a new build architecture will avoid this problem.  This solution
is more stopgap so that we can get some structure reuse results where
we have as much information as available as possible.

Pete
--------------------------------------------------------------------------
mercury-reviews mailing list
post:  mercury-reviews at cs.mu.oz.au
administrative address: owner-mercury-reviews at cs.mu.oz.au
unsubscribe: Address: mercury-reviews-request at cs.mu.oz.au Message: unsubscribe
subscribe:   Address: mercury-reviews-request at cs.mu.oz.au Message: subscribe
--------------------------------------------------------------------------



More information about the reviews mailing list