On Wed, Mar 11, 2020 at 08:20:01AM -0400, Richard Hipp wrote:
> On 3/11/20, Stefan Sperling wrote:
> >
> > Does this help?
> >
>
> Studying the trace output makes me think this is the same problem that
> came up on the s390 hardware and was fixed here:
>
On 3/11/20, Stefan Sperling wrote:
>
> Does this help?
>
Studying the trace output makes me think this is the same problem that
came up on the s390 hardware and was fixed here:
https://www.sqlite.org/src/info/04885763c4cd00cb
Stefan: Can you verify that the patch above fixes the problem? If
On 3/11/20, Stefan Sperling wrote:
>
> Does this help?
>
It does help some. But I need a good copy of the database schema,
which I don't have.
Please bring up the database file in the "sqlite3" command-line shell,
then enter the command ".fullschema". Save off the results and send
them to me,
On Tue, Mar 10, 2020 at 06:42:18PM +0100, Stefan Sperling wrote:
> On Tue, Mar 10, 2020 at 01:21:34PM -0400, Richard Hipp wrote:
> > 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
Stefan Sperling wrote:
> > Does valgrind give any clues?
>
> Valgrind does not run on the OpenBSD/sparc64 platform, unfortunately.
Would the address sanitizer (i.e. gcc -fsanitize=address, or
clang -fsanitize=address) work on OpenBSD/sparc64?
Regards
Dominique
On Tue, Mar 10, 2020 at 01:21:34PM -0400, Richard Hipp wrote:
> On 3/10/20, Stefan Sperling 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,
> >
On 3/10/20, Stefan Sperling 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,
>
7 matches
Mail list logo