actually, this isn't our fault, this is just generic 5.8 brokenness

[gitmo/Package-Stash-XS.git] / Changes

diff --git a/Changes b/Changes
index 14dfca8..f1a1b43 100644 (file)
--- a/Changes
+++ b/Changes
@@ -1,6 +1,52 @@
-Revision history for Package-Stash
+Revision history for Package-Stash-XS
 
 {{$NEXT}}
+
+0.25  2011-09-06
+      - fix compilation on perl 5.8 (RT#70762)
+
+0.24  2011-09-05
+      - invalid package names (for instance, Foo:Bar) are not allowed
+      - invalid stash entry names (anything containing ::) are not allowed
+
+0.23  2011-08-08
+      - fix the test for scalar values, again (t0m, rafl)
+      - disallow assigning globrefs to scalar glob slots (this doesn't actually
+        make any sense)
+
+0.22  2011-03-04
+      - make the namespace cache lazy and weak, in case the stash is deleted
+      - but, this doesn't work on 5.8, so disable the namespace caching
+        entirely there
+
+0.21  2011-01-25
+      - make the leak tests author-only, since some smokers run release tests
+      - fix some xs forward compat stuff (Dave Mitchell)
+
+0.20  2011-01-11
+      - lower perl prereq to 5.8.1 (ribasushi)
+      - make the leak tests release-only, since they keep randomly failing on
+        platforms i don't have access to. people are encouraged to submit
+        patches for these if they are affected.
+
+0.19  2011-01-05
+      - actually fix scalar initial value validity test
+
+0.18  2011-01-05
+      - more correct validity test for scalars (rafl, Roland van Ipenburg)
+
+0.17  2010-11-16
+      - bah, get rid of extra junk in the tarball
+
+0.16-TRIAL  2010-11-16
+      - fix some issues in the undocumented-but-still-maybe-used-some-places
+        hashref api - if you don't know what this is, good!
+
+0.15-TRIAL  2010-11-16
+      - split the XS implementation out into its own dist, which will be used
+        as a backend for Package::Stash if possible
+
+0.14-TRIAL  2010-11-14
       - complete rewrite in C, for speed (this includes the vivification
         changes from earlier). should be entirely backwards compatible
         otherwise (in terms of documented api anyway).