On 24/03/2008, at 19:15, Simon Marlow wrote:
Ian Lynagh wrote:
On the second one whether it is the right fix. I could be missing
something.
I don't think that this will work if we do
:cd myDir
:l myFile.hs
Am I wrong?
Pepe - perhaps you could elaborate on exactly what is wrong and how
you're trying to fix it.
As I understand it, the problem we have right now is what happens
when the program itself calls System.Directory.setCurrentDirectory,
and when we stop at a breakpoint, :list doesn't work. I think the
right fix for this is to "virtualise" the current working directory
for the client program - that is, we save/restore the cwd whenever
we run client code. This is more robust than trying to use absolute
paths throughout GHC.
Cheers,
Simon
The problem I am trying to solve is the one you describe; in fact I
sent a companion patch for the testsuite including a test list002
which does exactly that. I have indeed encountered this problem in
daily use of the debugger, e.g. while debugging a patch for Cabal
Install.
On IRC discussion with Igloo some days ago we found that saving the
cwd right after :load is used is enough to solve this particular
issue, but I agree that your proposal is more robust. I will send a
patch as soon as I get some time for it.
Cheers,
pepe
_______________________________________________
Cvs-ghc mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/cvs-ghc
