Fix a segfault on global pool cleanup after merge which has been
commit9cbd5e4ce9382f8f2fd1c23e679113af03376220
authorglasser <glasser@612f8ebc-c883-4be0-9ee0-a4e9ef946e3a>
Fri, 8 Feb 2008 00:13:39 +0000 (8 00:13 +0000)
committerglasser <glasser@612f8ebc-c883-4be0-9ee0-a4e9ef946e3a>
Fri, 8 Feb 2008 00:13:39 +0000 (8 00:13 +0000)
tree8a690c24a6f2e77c5c73933f3f6e8a6b16c843f3
parent3007386d5d8d5e979eb8ca2f9b288104a8018061
Fix a segfault on global pool cleanup after merge which has been
occurring on some platforms since r28442.  It appears that, due to the
ne_decompress destruction in the request cleanup handler,
svn_ra_neon__parsed_request must always be called with a pool whose
lifetime is no longer than the RA session pool's.  (This will be
documented in a future revision.  (Update: actually, r29231 fixes
this in a different way.))

* subversion/libsvn_ra_neon/get_location_segments.c
  (svn_ra_neon__get_location_segments): Run everything (including
   svn_ra_neon__parsed_request) in a subpool.

git-svn-id: http://svn.collab.net/repos/svn/trunk@29230 612f8ebc-c883-4be0-9ee0-a4e9ef946e3a
subversion/libsvn_ra_neon/get_location_segments.c