Various small fixes and reverse dependency tests using packit

.coderabbit.yaml

@@ -0,0 +1,14 @@
+language: "en-US"
+early_access: false
+reviews:
+  profile: "chill"
+  poem: false
+  in_progress_fortune: false
+  sequence_diagrams: false
+  related_issues: false
+  related_prs: false
+  suggested_labels: false
+  suggested_reviewers: false
+  path_filters: ["!po/**", "!plans/**", "!.github/**", "!.fmf/**"]
+chat:
+  art: false

.fmf/version

@@ -0,0 +1 @@
+1

.packit.yaml

@@ -77,14 +77,41 @@ jobs:
   dist_git_branches:
     - fedora-branched
 
+- job: tests
+  trigger: pull_request
+  targets:
+    - fedora-latest-x86_64
+    - fedora-latest-aarch64
+    - fedora-latest-stable-x86_64
+    - fedora-latest-stable-aarch64
+  branch: main
+
+# run tests for libbytesize consumers, see plans/ with `revdeps_blivet == yes`
+- job: tests
+  identifier: revdeps_blivet
+  trigger: pull_request
+  notifications:
+    failure_comment:
+      message: "Blivet tests failed for commit {commit_sha}. @vojtechtrefny please check."
+  targets:
+    - fedora-latest-stable
+  branch: main
+  tf_extra_params:
+    environments:
+      - artifacts:
+        - type: repository-file
+          id: https://copr.fedorainfracloud.org/coprs/g/storage/blivet-daily/repo/fedora-$releasever/group_storage-blivet-daily-fedora-$releasever.repo
+        tmt:
+          context:
+            revdeps_blivet: "yes"
+
 srpm_build_deps:
  - make
  - gcc
  - libtool
  - autoconf
  - automake
  - gmp-devel
- - mpfr-devel
  - pcre2-devel
  - gettext-devel
  - python3-devel

dist/libbytesize.spec.in

@@ -26,7 +26,6 @@ Source0:     https://github.com/storaged-project/libbytesize/releases/download/%
 BuildRequires: make
 BuildRequires: gcc
 BuildRequires: gmp-devel
-BuildRequires: mpfr-devel
 BuildRequires: pcre2-devel
 BuildRequires: gettext-devel
 %if %{with_python3}

plans/blivet.fmf

@@ -0,0 +1,35 @@
+# reverse dependency test for blivet
+enabled: false
+
+adjust+:
+  when: revdeps_blivet == yes
+  enabled: true
+
+prepare:
+  - name: copr
+    how: shell
+    script:
+      - sudo dnf install -y python3-libdnf5 'dnf-command(copr)'
+      - sudo dnf copr enable -y @storage/udisks-daily
+      # TF prioritizes Fedora tag repo over all others, in particular our daily COPR
+      - for f in $(grep -l -r 'testing-farm-tag-repository' /etc/yum.repos.d); do sed -i '/priority/d' "$f" ;done
+      - sudo dnf -y update
+
+  - name: ansible
+    how: shell
+    script:
+      - sudo dnf install -y curl ansible
+      - curl -Ok https://raw.githubusercontent.com/storaged-project/blivet/main/misc/install-test-dependencies.yml
+      - curl -Ok https://raw.githubusercontent.com/storaged-project/blivet/main/misc/blivet-tasks.yml
+      - ansible-playbook -K -i "localhost," -c local install-test-dependencies.yml
+
+discover:
+    how: shell
+    url: https://github.com/storaged-project/blivet
+    ref: main
+    tests:
+      - name: all
+        test: make test
+
+execute:
+    how: tmt

plans/tests.fmf

@@ -0,0 +1,14 @@
+summary: Run tests
+
+adjust+:
+  - when: revdeps_blivet == yes
+    enabled: false
+
+prepare:
+  - name: ansible
+    how: ansible
+    playbook: misc/install-test-dependencies.yml
+
+execute:
+    how: tmt
+    script: ./autogen.sh && ./configure && make -j && make ci

src/bs_size.c

@@ -7,6 +7,7 @@
 #include <string.h>
 #include <ctype.h>
 #include <limits.h>
+#include <assert.h>
 
 /* set code unit width to 8 so we can use generic macros like 'pcre2_compile'
  * instead of 'pcre2_compile_8'
@@ -398,6 +399,7 @@ void bs_clear_error (BSError **error) {
  */
 BSSize bs_size_new (void) {
     BSSize ret = (BSSize) malloc (sizeof(struct _BSSize));
+    assert (ret);
     bs_size_init (ret);
     return ret;
 }
@@ -701,6 +703,7 @@ uint64_t bs_size_get_bytes (const BSSize size, int *sgn, BSError **error) {
     mpz_set_str (max, num_str, 10);
     free (num_str);
     if (mpz_cmp (size->bytes, max) > 0) {
+        mpz_clear (max);
         set_error (error, BS_ERROR_OVER, strdup("The size is too big, cannot be returned as a 64bit number of bytes"));
         return 0;
     }
@@ -1212,7 +1215,7 @@ char* bs_size_true_div (const BSSize size1, const BSSize size2, BSError **error)
 
     if (mpz_cmp_ui (size2->bytes, 0) == 0) {
         set_error (error, BS_ERROR_ZERO_DIV, strdup_printf("Division by zero"));
-        return 0;
+        return NULL;
     }
 
     mpf_init2 (op1, BS_FLOAT_PREC_BITS);
@@ -1248,7 +1251,7 @@ char* bs_size_true_div_int (const BSSize size, uint64_t divisor, BSError **error
 
     if (divisor == 0) {
         set_error (error, BS_ERROR_ZERO_DIV, strdup_printf ("Division by zero"));
-        return 0;
+        return NULL;
     } else if (divisor > ULONG_MAX) {
         set_error (error, BS_ERROR_OVER, strdup_printf ("Divisor too big, must be less or equal to %lu", ULONG_MAX));
         return NULL;

src/bs_size.h

@@ -16,6 +16,7 @@ typedef struct _BSSize * BSSize;
  * @BS_ERROR_INVALID_SPEC: invalid size or unit spec provided
  * @BS_ERROR_OVER: a value is over the limits imposed by a type
  * @BS_ERROR_ZERO_DIV: an attempt to do division by zero
+ * @BS_ERROR_FAIL: generic failure error code
  *
  * Error codes that identify various errors that can occur while working with
  * #BSSize instances.
@@ -74,14 +75,14 @@ typedef enum {
  * @bunit: a binary unit
  * @dunit: a decimal unit
  *
- * Generic unit fo size in bytes.
+ * Generic unit for size in bytes.
  */
 typedef union {
     BSBunit bunit;
     BSDunit dunit;
 } BSUnit;
 
-/* use 256 bits of precision for floating point numbets, that should be more
+/* use 256 bits of precision for floating point numbers, that should be more
    than enough */
 /**
  * BS_FLOAT_PREC_BITS:

tests/libbytesize_unittest.py

@@ -436,12 +436,12 @@ def testSgn(self):
     def testTrueDiv(self):
         x = SizeStruct.new_from_str("1024 B")
         y = SizeStruct.new_from_str("-102.4 B") # rounds to whole bytes
-        divResult = float(x.true_div(y)[:15].replace(locale.nl_langinfo(locale.RADIXCHAR), ".")) # just some number to cover accurancy and not cross max float range
+        divResult = float(x.true_div(y)[:15].replace(locale.nl_langinfo(locale.RADIXCHAR), ".")) # just some number to cover accuracy and not cross max float range
         self.assertAlmostEqual(divResult, 1024.0/-102.0)
 
         x = SizeStruct.new_from_str("1 MiB")
         y = SizeStruct.new_from_str("1 KiB")
-        divResult = float(x.true_div(y)[:15].replace(locale.nl_langinfo(locale.RADIXCHAR), ".")) # just some number to cover accurancy and not cross max float range
+        divResult = float(x.true_div(y)[:15].replace(locale.nl_langinfo(locale.RADIXCHAR), ".")) # just some number to cover accuracy and not cross max float range
         self.assertAlmostEqual(divResult, 1024.0)
     #enddef
 
@@ -774,8 +774,8 @@ def testTrueDivInt(self):
     @requires_locales({'en_US.UTF-8'})
     def testPowerComputationRoundingIssues(self):
         """Test cases that expose rounding differences when using floating-point arithmetic.
-        
-        These test cases were discovered by fuzzing and demonstrate that both GMP and MPFR 
+
+        These test cases were discovered by fuzzing and demonstrate that both GMP and MPFR
         can produce incorrect results due to floating-point rounding errors: half of these
         fail using GMP floating-point arithmetic, and half fail using MPFR floating-point
         arithmetic. They all pass using rational arithmetic, so this can be considered a
@@ -794,7 +794,7 @@ def testPowerComputationRoundingIssues(self):
             ('0.0042875429 EB', 4287542900000000),
             ('0.0324645967 YB', 32464596700000000000000),
             ('0.1417885628 ZB', 141788562800000000000),
-            
+
             # Medium values
             ('1.2558302853 TB', 1255830285300),
             ('1.2808632839 TB', 1280863283900),
@@ -857,15 +857,15 @@ def testPowerComputationRoundingIssues(self):
             ('128.1037376252 PB', 128103737625200000),
             ('130.4743561323 ZB', 130474356132300000000000),
             ('138.2867513494 YB', 138286751349400000000000000),
-            
+
             # Large values
             ('18258.0630890156 PB', 18258063089015600000),
             ('18800.1176214700 EB', 18800117621470000000000),
             ('269636.0318459886 YB', 269636031845988600000000000000),
             ('276686.6833125990 YB', 276686683312599000000000000000),
             ('535817.4105711933 EB', 535817410571193300000000),
         ]
-        
+
         for test_str, expected_bytes in test_cases:
             with self.subTest(test_str=test_str):
                 size = SizeStruct.new_from_str(test_str)
@@ -891,7 +891,7 @@ def testPowerComputationRoundingIssues(self):
 if __name__=='__main__':
     if len(sys.argv) > 1:
         DEFAULT_LOCALE = sys.argv[1]
-        # the unittest module would try to intepret the argument too, let's
+        # the unittest module would try to interpret the argument too, let's
         # remove it
         sys.argv = [sys.argv[0]]
     unittest.main()