On Tue, Mar 13, 2018 at 2:29 AM, Dave Page <dpage@pgadmin.org> wrote:
So I was trying to test this, and every time I try to run a backup, I'm getting the following, with or without your patch:
(sqlite3.ProgrammingError) You must not use 8-bit bytestrings unless you use a text_factory that can interpret 8-bit bytestrings (like text_factory = str). It is highly recommended that you instead just switch your application to Unicode strings. [SQL: u'INSERT INTO process (pid, user_id, command, "desc", arguments, logdir, start_time, end_time, exit_code, acknowledge) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?)'] [parameters: (180312205250107339, 1, u'/Library/PostgreSQL/9.4/bin/pg_dump', 'ccopy_reg\n_reconstructor\np0\n(cpgadmin.tools.backup\nBackupMessage\np1\nc__builtin__\nobject\np2\nNtp3\nRp4\n(dp5\nS\'cmd\'\np6\nV --file "/Users/dpage/foo.dmp" --host "127.0.0.1" --port "5432" --username "postgres" --no-password --verbose --format=c --blobs "\xe9"\np7\nsS\'backup_type\'\np8\nI3\nsS\'database\'\np9\nV\xe9\np10\nsS\'bfile\'\np11\nS\'foo.dmp\'\np12\nsS\'sid\'\np13\nI1\nsb.', u'--file,/Users/dpage/foo.dmp,--host,127.0.0.1,--port,5432,--username,postgres,--no-password,--verbose,--format=c,--blobs,\xe9', '/var/lib/pgadmin/sessions/process_logs/180312205250107339', None, None, None, None)]
Any thoughts as to what's going on? I wasn't getting this on my other laptop, and I can't think what else we would have changed to cause this.
Deleting all the records from the process table from SQLITE will solve this problem.
There were few issues related to encoding-decoding in my old patches, you may have applied those.
I deleted the database entirely, and still saw the problem.
I have tried many things to reproduce this but couldn't. I faced this issue when I was fixing the issue but not now.
Please make sure to delete the old session and a process table (or database) and apply my latest patch. (of course you do this :) )
I have PY - 2.7.12, psycopg2 - 2.7.4 and SQLAlhemy -1.0.14 to be more specific.
Well I eventually got this to work (basically recreated most of my dev environment), so I committed the patch as it's clearly an improvement.
I can still reproduce the display issue I mentioned though - see the attached screenshot which shows é in a couple of places. I wonder if it's because the database name is just a single character in my tests, whilst you had some other unicode characters in the string?
