| |
@@ -0,0 +1,211 @@
|
| |
+ package dnf;
|
| |
+
|
| |
+ use strict;
|
| |
+
|
| |
+ use base 'Exporter';
|
| |
+ use Exporter;
|
| |
+ use lockapi;
|
| |
+ use testapi;
|
| |
+ use utils;
|
| |
+
|
| |
+ our @EXPORT = qw(install check remove reinstall parse_list_output parse_history_info latest_item parse_info parse_package_name confirm_in_lines confirm_in_output);
|
| |
+
|
| |
+
|
| |
+ # This subroutine install a $package using DNF5.
|
| |
+ # It dies if the command returns a non-zero exit
|
| |
+ # code or if it times out.
|
| |
+ sub install {
|
| |
+ my $package = shift;
|
| |
+ assert_script_run("dnf5 install -y $package", timeout => 120);
|
| |
+ }
|
| |
+
|
| |
+ # This subroutine checks if the $package is installed
|
| |
+ # using the rpm command. It returns True if it is
|
| |
+ # installed.
|
| |
+ sub check {
|
| |
+ my $package = shift;
|
| |
+ if (script_run("rpm -q $package", timeout => 60)) {
|
| |
+ return 0;
|
| |
+ }
|
| |
+ else {
|
| |
+ return 1;
|
| |
+ }
|
| |
+ }
|
| |
+
|
| |
+ # This subroutine removes the package using DNF5.
|
| |
+ sub remove {
|
| |
+ my $package = shift;
|
| |
+ assert_script_run("dnf5 remove -y $package", timeout => 120);
|
| |
+ }
|
| |
+
|
| |
+ # This subroutine reinstalls the package using DNF5.
|
| |
+ sub reinstall {
|
| |
+ my $package = shift;
|
| |
+ assert_script_run("dnf5 reinstall -y $package", timeout => 120);
|
| |
+ }
|
| |
+
|
| |
+ # This subroutine will parse the DNF5 cli output
|
| |
+ # and return a hash with activities.
|
| |
+ # Parsing the output is quite fragile, because
|
| |
+ # the info is only divided by whitespaces and some
|
| |
+ # columns are empty, so we retain column information
|
| |
+ # in this case. Therefore, we will only consider
|
| |
+ # the following structure, that will do for simple
|
| |
+ # transactions:
|
| |
+ # number : dnf5 : command : value : date
|
| |
+ sub parse_list_output {
|
| |
+ my $dnf = shift;
|
| |
+ # Let us split the whole ouput into single lines
|
| |
+ my @lines = split("\n", $dnf);
|
| |
+ # The final hash with output information.
|
| |
+ my $parsed_output = {};
|
| |
+ # Let's iterate over the single lines and split
|
| |
+ # them into single columns and store in the
|
| |
+ # parsed output.
|
| |
+ my $values = [];
|
| |
+ for my $line (@lines) {
|
| |
+ # Split columns on a white space.
|
| |
+ my @columns = split(" ", $line);
|
| |
+ # Take the columns
|
| |
+ $values->[0] = $columns[1];
|
| |
+ $values->[1] = $columns[2];
|
| |
+ # When date comes on the third place, it means
|
| |
+ # that the command did not have any arguments.
|
| |
+ # We need to remember this and we will store
|
| |
+ # an empty string instead.
|
| |
+ if ($columns[3] =~ /\d\d\d\d-\d\d-\d\d/) {
|
| |
+ $values->[2] = "";
|
| |
+ $values->[3] = $columns[3];
|
| |
+ }
|
| |
+ # All DNF commands have the "-y" argument
|
| |
+ # which we also need to think about.
|
| |
+ elsif ($columns[3] eq "-y") {
|
| |
+ $values->[2] = $columns[4];
|
| |
+ $values->[3] = $columns[5];
|
| |
+ }
|
| |
+ else {
|
| |
+ $values->[2] = $columns[3];
|
| |
+ $values->[3] = $columns[4];
|
| |
+ }
|
| |
+ # The output also contains a description line that
|
| |
+ # we do not want to store. That line leaves the "ID"
|
| |
+ # string in the first place, so it is easy to find.
|
| |
+ unless ($columns[0] eq "ID") {
|
| |
+ $parsed_output->{$columns[0]} = $values;
|
| |
+ }
|
| |
+ }
|
| |
+ return $parsed_output;
|
| |
+ }
|
| |
+
|
| |
+ # This subroutine parses the DNF5 history info output. We will
|
| |
+ # only consider several fields in order not to make things overly
|
| |
+ # complicated, i.e. ID, user, status, releasever, and description.
|
| |
+ sub parse_history_info {
|
| |
+ my $dnf = shift;
|
| |
+ # First, split into single lines.
|
| |
+ my @lines = split("\n", $dnf);
|
| |
+ # The final hash with output info
|
| |
+ my $parsed_output = {};
|
| |
+ # These are columns we are interested in.
|
| |
+ my $selected = ["Transaction ID", "User", "Status", "Releasever", "Description"];
|
| |
+ # Let us iterate over the lines and only choose info we want
|
| |
+ for my $line (@lines) {
|
| |
+ my @columns = split(" : ", $line);
|
| |
+ # Trim the whitespaces from the column name
|
| |
+ my $name = $columns[0];
|
| |
+ $name =~ s/^\s+|\s+$//g;
|
| |
+ # Only remember lines where $name matches
|
| |
+ if ($name ~~ @$selected) {
|
| |
+ $parsed_output->{$name} = $columns[1];
|
| |
+ }
|
| |
+ }
|
| |
+ return $parsed_output;
|
| |
+ }
|
| |
+
|
| |
+ # This subroutine returns the id number of the latest history item,
|
| |
+ # based on the parsed history output, see parse_list_output.
|
| |
+ sub latest_item {
|
| |
+ # It takes the parsed data already.
|
| |
+ my $dnfdata = shift;
|
| |
+ # Take the keys list of ID numbers.
|
| |
+ my @ids = keys %$dnfdata;
|
| |
+ # Sort the keys as keys might not be sorted.
|
| |
+ @ids = sort { $b <=> $a } @ids;
|
| |
+ # The latest item, with highest number will now be at index 0.
|
| |
+ return $ids[0];
|
| |
+ }
|
| |
+
|
| |
+ # This subroutine parses info from the 'info' output
|
| |
+ # and returns a hash where the output comprises
|
| |
+ # of keys and values.
|
| |
+ sub parse_info {
|
| |
+ my $output = shift;
|
| |
+ my $info = {};
|
| |
+
|
| |
+ my @output = split("\n", $output);
|
| |
+ foreach (@output) {
|
| |
+ my ($key, $value) = split(" : ", $_);
|
| |
+ $key =~ s/^\s+|\s+$//g;
|
| |
+ $value =~ s/^\s+|\s+$//g;
|
| |
+ $info->{$key} = $value;
|
| |
+ }
|
| |
+ return $info;
|
| |
+ }
|
| |
+
|
| |
+ # This routine takes the package name
|
| |
+ # and splits it into various parts that are
|
| |
+ # returned as a hash where the values are
|
| |
+ # easily accessible.
|
| |
+ sub parse_package_name {
|
| |
+ my $rpm = shift;
|
| |
+ my @rpm = split("-", $rpm);
|
| |
+ my ($name, $version, $suffix) = split("-", $rpm);
|
| |
+ my ($release, $os, $arch);
|
| |
+ ($release, $os, $arch, $suffix) = split(/\./, $suffix);
|
| |
+ my $package = {};
|
| |
+ $package->{name} = $name;
|
| |
+ $package->{version} = $version;
|
| |
+ $package->{release} = "$release.$os";
|
| |
+ $package->{arch} = $arch;
|
| |
+ return $package;
|
| |
+ }
|
| |
+
|
| |
+
|
| |
+ # This subroutine checks all the lines of $output
|
| |
+ # for the $keyword and records all lines where
|
| |
+ # the $keyword is not found and then it dies.
|
| |
+ sub confirm_in_lines {
|
| |
+ my ($output, $keyword) = @_;
|
| |
+ my @unmatched;
|
| |
+ my @lines = split("\n", $output);
|
| |
+ # Iterate over the lines and record any line with a discrepancy
|
| |
+ # into the unmatched array.
|
| |
+ foreach (@lines) {
|
| |
+ unless ($_ =~ /Updating and loading/ or $_ =~ /Repositories loaded/) {
|
| |
+ unless ($_ =~ /$keyword/) {
|
| |
+ push(@unmatched, $_);
|
| |
+ }
|
| |
+ }
|
| |
+ }
|
| |
+ # If there are errors in unmatched, log them and die.
|
| |
+ if (scalar(@unmatched) > 0) {
|
| |
+ diag("DNF5 Repoquerry errors:");
|
| |
+ foreach (@unmatched) {
|
| |
+ diag($_);
|
| |
+ }
|
| |
+ die("dnf5 repoquery returned lines that did not match the chosen pattern.");
|
| |
+ }
|
| |
+ }
|
| |
+
|
| |
+ # This subroutine checks that the $keyword
|
| |
+ # exists in the $output and dies if it does not.
|
| |
+ sub confirm_in_output {
|
| |
+ my ($output, $keywords) = @_;
|
| |
+ my @lines = split("\n", $output);
|
| |
+ foreach (@$keywords) {
|
| |
+ unless (grep(/$_/, @lines)) {
|
| |
+ die("The $_ was not found in the output.");
|
| |
+ }
|
| |
+ }
|
| |
+ }
|
| |
+
|
| |
lruzicka
commented 2 years ago
DNF5 is a major upcoming change for the future Fedora. It is wise if we started to run tests as soon as possible, so that we could see if anything breaks during the transition process and in the early adoption days.
This test suite is based on the test cases of the DNF5 test day and performs the same (or similar steps) to test the DNF5. Some of the tests from that test day, however, have not been implemented yet, because the commands are not supported by the current version of DNF5.
Only the commands, that are at least partly supported, became part of this suite.