[stable-2.11] subversion - fix stack trace when getting repository information (#74405) (#74977)

* subversion - set LC_ALL for accurate command output parsing

When LC_ALL is not set, the output language of commands will differ based on locale. There
is a lot of history of trying to fix this. See the following pull requests:

https://github.com/ansible/ansible-modules-core/pull/4358
https://github.com/ansible/ansible-modules-core/pull/4358

This patch attempts to fix this my setting LC_ALL to a UTF-8 locale. Setting LC_ALL to C reintroduces this bug https://github.com/ansible/ansible-modules-core/issues/4178.

I'm sure there are some problems I am not seeing with setting this to en_US.UTF-8, but that is
the only way I could find to fix this bug without reintriducing the bug mentioned above.

* Rather than setting locale, just check for matches before trying to get groups

This is a pragmatic solution to avoid the stack trace since setting the locale correctly
to ensure message parsing is accurate is problematic.

* Improve regexps for finding revision and URL
(cherry picked from commit a8cf0196f7)

Co-authored-by: Sam Doran <sdoran@redhat.com>

2 files changed, 31 insertions, 4 deletions

diff --git a/changelogs/fragments/36498-subversion-fix-info-parsing.yml b/changelogs/fragments/36498-subversion-fix-info-parsing.yml
new file mode 100644
index 0000000000..d14605b6c0
--- /dev/null
+++ b/changelogs/fragments/36498-subversion-fix-info-parsing.yml
@@ -0,0 +1,2 @@
+bugfixes:
+  - subversion - fix stack trace when getting information about the repository (https://github.com/ansible/ansible/issues/36498)
diff --git a/lib/ansible/modules/subversion.py b/lib/ansible/modules/subversion.py
index c28471bf22..c62146af59 100644
--- a/lib/ansible/modules/subversion.py
+++ b/lib/ansible/modules/subversion.py
@@ -134,6 +134,13 @@ from ansible.module_utils.basic import AnsibleModule
 
 
 class Subversion(object):
+
+    # Example text matched by the regexp:
+    #  Révision : 1889134
+    #  版本: 1889134
+    #  Revision: 1889134
+    REVISION_RE = r'^\w+\s?:\s+\d+$'
+
     def __init__(self, module, dest, repo, revision, username, password, svn_path, validate_certs):
         self.module = module
         self.dest = dest
@@ -228,14 +235,28 @@ class Subversion(object):
     def get_revision(self):
         '''Revision and URL of subversion working directory.'''
         text = '\n'.join(self._exec(["info", self.dest]))
-        rev = re.search(r'^Revision:.*$', text, re.MULTILINE).group(0)
-        url = re.search(r'^URL:.*$', text, re.MULTILINE).group(0)
+        rev = re.search(self.REVISION_RE, text, re.MULTILINE)
+        if rev:
+            rev = rev.group(0)
+        else:
+            rev = 'Unable to get revision'
+
+        url = re.search(r'^URL\s?:.*$', text, re.MULTILINE)
+        if url:
+            url = url.group(0)
+        else:
+            url = 'Unable to get URL'
+
         return rev, url
 
     def get_remote_revision(self):
         '''Revision and URL of subversion working directory.'''
         text = '\n'.join(self._exec(["info", self.repo]))
-        rev = re.search(r'^Revision:.*$', text, re.MULTILINE).group(0)
+        rev = re.search(self.REVISION_RE, text, re.MULTILINE)
+        if rev:
+            rev = rev.group(0)
+        else:
+            rev = 'Unable to get remote revision'
         return rev
 
     def has_local_mods(self):
@@ -250,7 +271,11 @@ class Subversion(object):
     def needs_update(self):
         curr, url = self.get_revision()
         out2 = '\n'.join(self._exec(["info", "-r", self.revision, self.dest]))
-        head = re.search(r'^Revision:.*$', out2, re.MULTILINE).group(0)
+        head = re.search(self.REVISION_RE, out2, re.MULTILINE)
+        if head:
+            head = head.group(0)
+        else:
+            head = 'Unable to get revision'
         rev1 = int(curr.split(':')[1].strip())
         rev2 = int(head.split(':')[1].strip())
         change = False