- sections in my.cnf are called groups in MySQL documentation. So
let's be consistent and call them groups as well, and name the
options (both the innobackupex and the xtrabackup one)
--defaults-group, rather than --section and --mysqld-section
- it looks like get_option is always passed the same value as its
second argument, so I wonder if it really needs that argument, or
rather just use $option_mysqld_section (or whatever it will be
renamed to, i.e. $option_defaults_group) internally.
- s/laod/load/
- s/accetps/accepts/
- s/rstore_args/restore_args/. Though you don't really need that
function, as tests are executed in a separate shell process, so
modifications to variables have no effect on other tests anyway.
- I know that code in main() that scan options for "--mysqld-section"
was copy-pasted, but please format it according to InnoDB
style. Because currently it's a terrible mix of all possible
formatting styles.
- in the same code, I don't think strstr() is necessary, because you
already have the pointer to '=' (or terminating zero) in optend.
Sergei,
- sections in my.cnf are called groups in MySQL documentation. So defaults- group, rather than --section and --mysqld-section
let's be consistent and call them groups as well, and name the
options (both the innobackupex and the xtrabackup one)
--
- it looks like get_option is always passed the same value as its mysqld_ section (or whatever it will be defaults_ group) internally.
second argument, so I wonder if it really needs that argument, or
rather just use $option_
renamed to, i.e. $option_
- s/laod/load/ args/restore_ args/. Though you don't really need that
- s/accetps/accepts/
- s/rstore_
function, as tests are executed in a separate shell process, so
modifications to variables have no effect on other tests anyway.
- I know that code in main() that scan options for "--mysqld-section"
was copy-pasted, but please format it according to InnoDB
style. Because currently it's a terrible mix of all possible
formatting styles.
- in the same code, I don't think strstr() is necessary, because you
already have the pointer to '=' (or terminating zero) in optend.