On Tue, Mar 10, 2020 at 01:21:34PM -0400, Richard Hipp wrote: > On 3/10/20, Stefan Sperling <s...@apache.org> wrote: > > The query being executed is "STMT_INSERT_NODE": > > -- STMT_INSERT_NODE > > INSERT OR REPLACE INTO nodes ( > > wc_id, local_relpath, op_depth, parent_relpath, repos_id, repos_path, > > revision, presence, depth, kind, changed_revision, changed_date, > > changed_author, checksum, properties, translated_size, last_mod_time, > > dav_cache, symlink_target, file_external, moved_to, moved_here, > > inherited_props) > > VALUES (?1, ?2, ?3, ?4, ?5, ?6, ?7, ?8, ?9, ?10, ?11, ?12, ?13, ?14, > > ?15, ?16, ?17, ?18, ?19, ?20, ?21, ?22, ?23) > > It would be great if you could send us the database schema - or even > the whole database if it isn't too big. > > Probably if I have the schema I will be able to repro the problem.
The schema is public: https://svn.apache.org/repos/asf/subversion/branches/1.13.x/subversion/libsvn_wc/wc-metadata.sql > Is this only happening on Sparc, and not on x64 or ARM or PPC, etc? It looks like it is specific to sparc64 so far. I cannot reproduce it on amd64 with the same software versions. I do not have access to arm/ppc machines at present. > Does valgrind give any clues? Valgrind does not run on the OpenBSD/sparc64 platform, unfortunately. > Something else that might be helpful, if you have time: > > Set a breakpoint on the sqlite3VdbeExec() function that fires when the > statement that is crashing is first executed. (You can determine that > it is the correct statement by looking at the value of p->zSql.) Then > do: > > set p->db->flags = p->db->flags | ((0x0600000)<<32) > > That will turn on bytecode listing and tracing, and might provide > further clues. Please record and send in the trace. I will look into this. Thanks! _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
