aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorletouzey2011-04-03 15:18:56 +0000
committerletouzey2011-04-03 15:18:56 +0000
commitf8e8ffa4b7c42f6c53126d284c0bfbf8a992bc89 (patch)
tree557bb27a8ba221cb0664aeab558eacaa83531d96
parenta9bf7d8e017488d94e8491c432a77a9a178be3a8 (diff)
Update documentation concerning proofs loading (cf last commit)
git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@13954 85f007b7-540e-0410-9357-904b9bb8a0f7
-rw-r--r--CHANGES3
-rw-r--r--doc/refman/RefMan-com.tex22
2 files changed, 22 insertions, 3 deletions
diff --git a/CHANGES b/CHANGES
index 0054e4cef6..19edf585e8 100644
--- a/CHANGES
+++ b/CHANGES
@@ -75,6 +75,9 @@ Libraries
Internal infrastructure
+- Opaque proofs are now loaded lazily by default. This allows to be almost as
+ fast as -dont-load-proofs, while being safer (no creation of axioms) and
+ avoiding feature restrictions (Print and Print Assumptions work ok).
- Experimental support added for camlp4 (the one provided alongside ocaml),
simply pass option -usecamlp4 to ./configure. By default camlp5 is used.
- Revised build system: no more stages in Makefile thanks to some recursive
diff --git a/doc/refman/RefMan-com.tex b/doc/refman/RefMan-com.tex
index f808047b12..aaa962570b 100644
--- a/doc/refman/RefMan-com.tex
+++ b/doc/refman/RefMan-com.tex
@@ -265,9 +265,25 @@ The following command-line options are recognized by the commands {\tt
\item[{\tt -dont-load-proofs}]\
- This avoids loading in memory the proofs of opaque theorems
- resulting in a smaller memory requirement and faster compilation;
- warning: this invalidates some features such as the extraction tool.
+ Warning: this is an unsafe mode.
+ Instead of loading in memory the proofs of opaque theorems, they are
+ treated as axioms. This results in smaller memory requirement and
+ faster compilation, but the behavior of the system might slightly change
+ (for instance during module subtyping), and some features won't be
+ available (for example {\tt Print Assumptions}).
+
+\item[{\tt -lazy-load-proofs}]\
+
+ This is the default behavior. Proofs of opaque theorems aren't
+ loaded immediately in memory, but only when necessary, for instance
+ during some module subtyping or {\tt Print Assumptions}. This should be
+ almost as fast and efficient as {\tt -dont-load-proofs}, with none
+ of its drawbacks.
+
+\item[{\tt -force-load-proofs}]\
+
+ Proofs of opaque theorems are loaded in memory as soon as the
+ corresponding {\tt Require} is done. This used to be Coq's default behavior.
\item[{\tt -vm}]\