Remaining log/data files generated by repo2html will now take into account path_root

This ensures that all logs are available in the directory in which repo2html is
invoked, which makes reference easier (this is also necessary if repo2html is
located in a directory that the user cannot write to, such as /usr/local/bin).
Furthermore, this makes more sense if repo2html is used with multiple
repositories on the machine.

.gitignore

@@ -1,3 +1,3 @@
 .cref-*
-.list
+.clist
 .hashcache

repo2html

@@ -167,14 +167,14 @@ echo "Clearing hashcache..." >&2
 # set the cref error log, which will allow us to re-process *only* those commits
 # that actually need to be reprocessed; this avoids the need to prime the
 # hashcache, saving cycles
-cref_errlog_first=.cref-errlog
+cref_errlog_first="$path_root/.cref-errlog"
 export cref_errlog="$cref_errlog_first"
->"$cref_errlog"
+>"$cref_errlog_first"
 
 # pass commit list to the HTML and RSS processors (cache list to disk for
 # reference and further processing); note that we clear the cref_errlog for the
 # RSS feed so that we do not get duplicates
-listcache=.list
+listcache="$path_root/.clist"
 "$repotype"/list | tee \
   >( cref_errlog= ./rss "$repotype" "$url_root" "$rss_count" \
      > "$path_out/rss.xml" \
@@ -186,7 +186,7 @@ listcache=.list
 # re-process cref errors (but only once; any errors at this point will be
 # considered to be problem refs) TODO: RSS; note that we redirect all output to
 # /dev/null so as not to re-output the index template
-export cref_errlog=.cref-bad
+export cref_errlog="$path_root/.cref-bad"
 >"$cref_errlog"
 ( grep -f<( sed 's/^/\^/' "$cref_errlog_first" ) "$listcache" \
   | ./processor "$repotype" "$path_out" 2> >( _reout ) \