#177 UserPrefs in postgress bad SQL

[Grzegorz Kowalski]

I installed latest stable version of SM 1.4.23-svn (from 14 May 2020) and I found two bugs/problems.

1) SM generate not proper SQL to postgresql database to table userprefs.

postgres logs:
2020-05-14 15:56:58.629 CEST [7605] ERROR: zero-length delimited identifier at or near """" at character 74
2020-05-14 15:56:58.629 CEST [7605] STATEMENT: SELECT "prefkey" AS prefkey, "prefval" AS prefval FROM "userprefs" WHERE ""user"" = $1

problem: double "" on column user.

2) My old messages with iso-8859-2 encoding display wrong polish characters in headers and body.
My old SM 1.4.22 displays ok such messages.

Language preferences
1. Default Language : pl_PL
2. Default Charset : UTF-8
3. Enable lossy encoding : true

Enviroment:
System: CentosOS 8
PostgreSQL 10.1
PHP 7.2.11
Apache 2.4.37
Plugins: only one plugin: login_auth
IMAP: Dovecot 2.2.36
SMTP: Postfix 3.3.1

[Grzegorz Kowalski]

1)
In file /functions/db_prefs.php in function open I commented out

// if ($this->db_type == SMDB_PGSQL) {
// $this->user_field = '"' . $this->user_field . '"';
// }

and SM can read and write userprefs to postgresql 11 database.

2)
I'm still looking for a workaround for the problem with correct decoding
for headers and body with iso-8859-2 (polish characters).
The encoding for mails with iso-8859-2 works, but such mails are not
corectly displayed in SM.

[Paul Lesniewski]

Thank you for reporting these issues. The first has been fixed. The second is likely fixed by other commits since the time that you had originally submitted this ticket, so please try a new SVN snapshot from our downloads page.