version bump -> 0.2.5 [skip ci]

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: dodgr
 Title: Distances on Directed Graphs
-Version: 0.2.4.001
+Version: 0.2.5
 Authors@R: c(
     person("Mark", "Padgham", email="[email protected]", role=c("aut", "cre")),
     person("Andreas", "Petutschnig", role="aut"),

NEWS.md

@@ -1,3 +1,7 @@
+# v0.2.5
+
+- bug fixes from previous versions
+
 # v0.2.4
 
 Major changes:

codemeta.json

@@ -10,7 +10,7 @@
   "codeRepository": "https://github.com/ATFutures/dodgr",
   "issueTracker": "https://github.com/ATFutures/dodgr/issues",
   "license": "https://spdx.org/licenses/GPL-3.0",
-  "version": "0.2.4.1",
+  "version": "0.2.5",
   "programmingLanguage": {
     "@type": "ComputerLanguage",
     "name": "R",
@@ -400,5 +400,5 @@
     }
   ],
   "relatedLink": "https://CRAN.R-project.org/package=dodgr",
-  "fileSize": "16795.838KB"
+  "fileSize": "371.252KB"
 }

cran-comments.md

@@ -1,25 +1,9 @@
-# CRAN notes for dodgr_0.2.4 submission
-
-A few submissions over the past week have failed autocheck, partly because of
-tests which have now been rectified, and presumably partly because of the
-problems with tlmgr and tikz on the Debian machine. This submission avoids the
-use of tikz, so hopefully overcoming those Debian issues.
-
-## ASAN/UBSAN notes on previous version
-
-This submission *should* rectify previous notes from address & undefined
-behaviour sanitizers. I have nevertheless been unable to confirm this as I have
-not been able to reproduce the result. I tried the rocker/r-devel-ubsan-clang
-container, but that failed to install package due to the valgrind memory leak
-from TBB via RcppParallel mentioned below. I also checked on r-hub's sanitizer
-system, but that failed for same reason, and not for reasons related to current
-CRAN failures. I could also not reproduce locally. One failing test I did
-manage to reproduce was the valgrind test, which this submission definitely
-fixes, reducing the possibly lost byte count from >200kB back to the "usual"
-2kB due to RcppParallel / TBB. I confidently presume that the fix also
-successfully addresses the ASAN/UBSAN issues. Sorry for any inconvenience.
-
-I have also rectified one previous, intermittently failing test.
+# CRAN notes for dodgr_0.2.5 submission
+
+I finally managed to reproduce the previous UBSCAN and valgrind errors observed
+by Brian Ripley on 20 Sept 2019. This submission definitely fixes, with both
+AddressSanitizer and valgrind on r-devel returning only the by-now usual loss of
+2-3000 bytes due to TBB code bundled in RcppParallel (see below).
 
 ## Notes