On 08/03/2012 04:10 AM, Makarius wrote:
On Tue, 31 Jul 2012, Christian Sternagel wrote:
In the example below Abstract-Rewriting did initially not build (since
I forgot to update my local AFP clone). Instead of giving me an
error-message (which did work when I just put the single theory
Abstract_Rewriting under "theories"), however, the build process ran
for more than 20 minutes (using several GB of RAM) before I killed it
manually.
Did you see that problem again?
I am not able to reproduce this behavior (unfortunately I did not take
down the changeset identifier when it occurred the first time and I have
fetched several times since).
The process management on the JVM is a bit different than on the shell.
For example, there are separate JVM threads that consume the
stdout/stderr streams until completion -- this could run into pathologic
situations.
Anyway, which process was consuming the GBs of memory? Java or poly?
Sorry, I only had a look at the overall memory consumption, not at
specific processes.
cheers
chris
_______________________________________________
isabelle-dev mailing list
isabelle-...@in.tum.de
https://mailmanbroy.informatik.tu-muenchen.de/mailman/listinfo/isabelle-dev
