ganeti.utils.text

Build a safe shell command line from the given arguments.

This function will check all arguments in the args list so that they are valid shell parameters (i.e. they don't contain shell metacharacters). If everything is ok, it will return the result of template % args.

Nicely join a set of identifiers.

Generates regular expression from DNS name globbing pattern.

A DNS name globbing pattern (e.g. *.site) is converted to a regular expression. Escape sequences or ranges (e.g. [a-z]) are not supported.

Matching always starts at the leftmost part. An asterisk (*) matches all characters except the dot (.) separating DNS name parts. A question mark (?) matches a single character except the dot (.).

Encode a list in a way parsable by UnescapeAndSplit.

Filters empty lines and comments from a line-based string.

Whitespace is also removed from the beginning and end of all lines.

Formats a dictionary as "key=value" parameters.

The keys are sorted to have a stable order.

Formats a number as an ordinal in the English language.

E.g. the number 1 becomes "1st", 22 becomes "22nd".

Formats seconds for easier reading.

Formats a time value.

Formats an incoming number of MiB with the appropriate unit.

Generates a random secret.

This will generate a pseudo-random secret returning an hex string (so that it can be used where an ASCII string is needed).

Verifies is the given word is safe from the shell's p.o.v.

This means that we can pass this to a command via the shell and be sure that it doesn't alter the command line and is passed as such to the actual command.

Note that we are overly restrictive here, in order to be on the safe side.

Try to match a name against a list.

This function will try to match a name like test1 against a list like ['test1.example.com', 'test2.example.com', ...]. Against this list, 'test1' as well as 'test1.example' will match, but not 'test1.ex'. A multiple match will be considered as no match at all (e.g. 'test1' against ['test1.example.com', 'test1.example.org']), except when the key fully matches an entry (e.g. 'test1' against ['test1', 'test1.example.com']).

Normalizes and check if a MAC address is valid and contains six octets.

Checks whether the supplied MAC address is formally correct. Accepts colon-separated format only. Normalize it to all lower case.

Normalizes a potential MAC address prefix (three octets).

Checks whether the supplied string is a valid MAC address prefix consisting of three colon-separated octets. The result is normalized to all lower case.

Tries to extract number and scale from the given string.

Input must be in the format NUMBER+ [DOT NUMBER+] SPACE* [UNIT]. If no unit is specified, it defaults to MiB. Return value is always an int in MiB.

Return a 'safe' version of a source string.

This function mangles the input string and returns a version that should be safe to display/encode as ASCII. To this end, we first convert it to ASCII using the 'backslashreplace' encoding which should get rid of any non-ASCII chars, and then we process it through a loop copied from the string repr sources in the python; we don't use string_escape anymore since that escape single quotes and backslashes too, and that is too much; and that escaping is not stable, i.e. string_escape(string_escape(x)) != string_escape(x).

Out of a list of shell comands construct a single one.

Quotes shell argument according to POSIX.

Quotes a list of shell arguments.

Truncate string and add ellipsis if needed.

Split and unescape a string based on a given separator.

This function splits a string based on a separator where the separator itself can be escape in order to be an element of the elements. The escaping rules are (assuming coma being the separator):

• a plain , separates the elements
• a sequence \\\\, (double backslash plus comma) is handled as a backslash plus a separator comma
• a sequence \, (backslash plus comma) is handled as a non-separator comma