diff --git a/SOURCES/0001-Fix-cr_compress_file_with_stat-Memory-leak.patch b/SOURCES/0001-Fix-cr_compress_file_with_stat-Memory-leak.patch
new file mode 100644
index 0000000..126a489
--- /dev/null
+++ b/SOURCES/0001-Fix-cr_compress_file_with_stat-Memory-leak.patch
@@ -0,0 +1,25 @@
+From 1ec1e767e4a48bc2c1d09b52da9e749352f8d925 Mon Sep 17 00:00:00 2001
+From: Jaroslav Rohel <jrohel@redhat.com>
+Date: Thu, 6 May 2021 19:09:19 +0200
+Subject: [PATCH] Fix: cr_compress_file_with_stat: Memory leak
+
+---
+ src/misc.c | 2 ++
+ 1 file changed, 2 insertions(+)
+
+diff --git a/src/misc.c b/src/misc.c
+index e5350ac..59c04e6 100644
+--- a/src/misc.c
++++ b/src/misc.c
+@@ -503,6 +503,8 @@ cr_compress_file_with_stat(const char *src,
+ if (!orig) {
+ ret = tmp_err->code;
+ g_propagate_prefixed_error(err, tmp_err, "Cannot open %s: ", src);
++ if (dst != in_dst)
++ g_free(dst);
+ return ret;
+ }
+
+--
+libgit2 1.0.1
+
diff --git a/SOURCES/0002-Preserve-changed-API-for-cr_compress_file_with_stat-RhBug1973588.patch b/SOURCES/0002-Preserve-changed-API-for-cr_compress_file_with_stat-RhBug1973588.patch
new file mode 100644
index 0000000..d14a0ae
--- /dev/null
+++ b/SOURCES/0002-Preserve-changed-API-for-cr_compress_file_with_stat-RhBug1973588.patch
@@ -0,0 +1,380 @@
+From a611bb0c3729713af28031cdf1d056823d663db1 Mon Sep 17 00:00:00 2001
+From: Aleš Matěj <amatej@redhat.com>
+Date: Thu, 9 Sep 2021 08:31:03 +0200
+Subject: [PATCH] Preserve changed API for cr_compress_file_with_stat (RhBug:1973588)
+
+In order to be compatible in rhel8 we want to preserve the old API and
+behavior.
+
+Keep the fixed version as cr_compress_file_with_stat_v2 only for rhel8
+
+https://bugzilla.redhat.com/show_bug.cgi?id=1973588
+---
+ src/misc.c | 138 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-
+ src/misc.h | 42 ++++++++++++++++++++++++++++++++++++++++--
+ src/modifyrepo_shared.c | 4 ++--
+ src/python/misc-py.c | 2 +-
+ src/threads.c | 14 +++++++-------
+ tests/test_misc.c | 34 +++++++++++++++++-----------------
+ 6 files changed, 204 insertions(+), 30 deletions(-)
+
+diff --git a/src/misc.c b/src/misc.c
+index 59c04e6..b3a218c 100644
+--- a/src/misc.c
++++ b/src/misc.c
+@@ -446,7 +446,7 @@ cr_copy_file(const char *src, const char *in_dst, GError **err)
+
+ int
+ cr_compress_file_with_stat(const char *src,
+- const char *in_dst,
++ char **in_dst,
+ cr_CompressionType compression,
+ cr_ContentStat *stat,
+ const char *zck_dict_dir,
+@@ -458,6 +458,142 @@ cr_compress_file_with_stat(const char *src,
+ char buf[BUFFER_SIZE];
+ CR_FILE *orig = NULL;
+ CR_FILE *new = NULL;
++ gchar *dst = (gchar *) *in_dst;
++ GError *tmp_err = NULL;
++
++ assert(src);
++ assert(!err || *err == NULL);
++
++ const char *c_suffix = cr_compression_suffix(compression);
++
++ // Src must be a file NOT a directory
++ if (!g_file_test(src, G_FILE_TEST_IS_REGULAR)) {
++ g_debug("%s: Source (%s) must be a regular file!", __func__, src);
++ g_set_error(err, ERR_DOMAIN, CRE_NOFILE,
++ "Not a regular file: %s", src);
++ return CRE_NOFILE;
++ }
++
++ if (!dst) {
++ // If destination is NULL, use src + compression suffix
++ *in_dst = g_strconcat(src,
++ c_suffix,
++ NULL);
++ } else if (g_str_has_suffix(dst, "/")) {
++ // If destination is dir use filename from src + compression suffix
++ *in_dst = g_strconcat(dst,
++ cr_get_filename(src),
++ c_suffix,
++ NULL);
++ } else if (c_suffix && !g_str_has_suffix(dst, c_suffix)) {
++ cr_CompressionType old_type = cr_detect_compression(src, &tmp_err);
++ if (tmp_err) {
++ g_debug("%s: Unable to detect compression type of %s", __func__, src);
++ g_clear_error(&tmp_err);
++ } else if (old_type != CR_CW_NO_COMPRESSION) {
++ _cleanup_free_ gchar *tmp_file = g_strndup(dst, strlen(dst) - strlen(cr_compression_suffix(old_type)));
++ *in_dst = g_strconcat(tmp_file,
++ c_suffix,
++ NULL);
++ }
++ }
++ if (dst != *in_dst && dst)
++ g_free(dst);
++ dst = (gchar *) *in_dst;
++
++ int mode = CR_CW_AUTO_DETECT_COMPRESSION;
++
++ orig = cr_open(src,
++ CR_CW_MODE_READ,
++ mode,
++ &tmp_err);
++ if (!orig) {
++ ret = tmp_err->code;
++ g_propagate_prefixed_error(err, tmp_err, "Cannot open %s: ", src);
++ return ret;
++ }
++
++ _cleanup_free_ gchar *dict = NULL;
++ size_t dict_size = 0;
++ if (compression == CR_CW_ZCK_COMPRESSION && zck_dict_dir) {
++ /* Find zdict */
++ _cleanup_free_ gchar *file_basename = NULL;
++ if (dst) {
++ _cleanup_free_ gchar *dict_base = NULL;
++ if (g_str_has_suffix(dst, ".zck"))
++ dict_base = g_strndup(dst, strlen(dst)-4);
++ else
++ dict_base = g_strdup(dst);
++ file_basename = g_path_get_basename(dict_base);
++ } else {
++ file_basename = g_path_get_basename(src);
++ }
++ _cleanup_free_ gchar *dict_file = cr_get_dict_file(zck_dict_dir, file_basename);
++
++ /* Read dictionary from file */
++ if (dict_file && !g_file_get_contents(dict_file, &dict,
++ &dict_size, &tmp_err)) {
++ g_set_error(err, ERR_DOMAIN, CRE_IO,
++ "Error reading zchunk dict %s: %s",
++ dict_file, tmp_err->message);
++ ret = CRE_IO;
++ goto compress_file_cleanup;
++ }
++ }
++
++ new = cr_sopen(dst, CR_CW_MODE_WRITE, compression, stat, &tmp_err);
++ if (tmp_err) {
++ g_debug("%s: Cannot open destination file %s", __func__, dst);
++ g_propagate_prefixed_error(err, tmp_err, "Cannot open %s: ", dst);
++ ret = CRE_IO;
++ goto compress_file_cleanup;
++ }
++ if (compression == CR_CW_ZCK_COMPRESSION) {
++ if (dict && cr_set_dict(new, dict, dict_size, &tmp_err) != CRE_OK) {
++ ret = tmp_err->code;
++ g_propagate_prefixed_error(err, tmp_err, "Unable to set zdict for %s: ", dst);
++ goto compress_file_cleanup;
++ }
++ if (zck_auto_chunk && cr_set_autochunk(new, TRUE, &tmp_err) != CRE_OK) {
++ ret = tmp_err->code;
++ g_propagate_prefixed_error(err, tmp_err, "Unable to set auto-chunking for %s: ", dst);
++ goto compress_file_cleanup;
++ }
++ }
++ while ((readed = cr_read(orig, buf, BUFFER_SIZE, &tmp_err)) > 0) {
++ cr_write(new, buf, readed, &tmp_err);
++ if (tmp_err) {
++ ret = tmp_err->code;
++ g_propagate_prefixed_error(err, tmp_err, "Unable to write to %s: ", dst);
++ goto compress_file_cleanup;
++ }
++ }
++
++compress_file_cleanup:
++
++ if (orig)
++ cr_close(orig, NULL);
++
++ if (new)
++ cr_close(new, NULL);
++
++ return ret;
++}
++
++int
++cr_compress_file_with_stat_v2(const char *src,
++ const char *in_dst,
++ cr_CompressionType compression,
++ cr_ContentStat *stat,
++ const char *zck_dict_dir,
++ gboolean zck_auto_chunk,
++ GError **err)
++{
++ int ret = CRE_OK;
++ int readed;
++ char buf[BUFFER_SIZE];
++ CR_FILE *orig = NULL;
++ CR_FILE *new = NULL;
+ gchar *dst = (gchar *) in_dst;
+ GError *tmp_err = NULL;
+
+diff --git a/src/misc.h b/src/misc.h
+index 60f1a0f..528ccc3 100644
+--- a/src/misc.h
++++ b/src/misc.h
+@@ -184,9 +184,24 @@ gboolean cr_copy_file(const char *src,
+ cr_compress_file_with_stat(SRC, DST, COMTYPE, NULL, ZCK_DICT_DIR, \
+ ZCK_AUTO_CHUNK, ERR)
+
++/** Compress file. This function is temporary and present
++ * only in rhel 8, it will be removed in future versions.
++ * @param SRC source filename
++ * @param DST destination (If dst is dir, filename of src +
++ * compression suffix is used.
++ * If dst is NULL, src + compression suffix is used)
++ * @param COMTYPE type of compression
++ * @param ZCK_DICT_DIR Location of zchunk zdicts (if zchunk is enabled)
++ * @param ZCK_AUTO_CHUNK Whether zchunk file should be auto-chunked
++ * @param ERR GError **
++ * @return cr_Error return code
++ */
++#define cr_compress_file_v2(SRC, DST, COMTYPE, ZCK_DICT_DIR, ZCK_AUTO_CHUNK, ERR) \
++ cr_compress_file_with_stat_v2(SRC, DST, COMTYPE, NULL, ZCK_DICT_DIR, \
++ ZCK_AUTO_CHUNK, ERR)
+ /** Compress file.
+ * @param src source filename
+- * @param dst destination (If dst is dir, filename of src +
++ * @param dst pointer to destination (If dst is dir, filename of src +
+ * compression suffix is used.
+ * If dst is NULL, src + compression suffix is used)
+ * @param comtype type of compression
+@@ -197,13 +212,36 @@ gboolean cr_copy_file(const char *src,
+ * @return cr_Error return code
+ */
+ int cr_compress_file_with_stat(const char *src,
+- const char *dst,
++ char **dst,
+ cr_CompressionType comtype,
+ cr_ContentStat *stat,
+ const char *zck_dict_dir,
+ gboolean zck_auto_chunk,
+ GError **err);
+
++/** Compress file with stat versions 2. This function is temporary and present
++ * only in rhel 8, it will be removed in future versions.
++ * It is a compatibility function that preserves the API and behavior of
++ * cr_compress_file_with_stat from createrepo_c-0.12.0.
++ * @param src source filename
++ * @param dst destination (If dst is dir, filename of src +
++ * compression suffix is used.
++ * If dst is NULL, src + compression suffix is used)
++ * @param comtype type of compression
++ * @param stat pointer to cr_ContentStat or NULL
++ * @param zck_dict_dir Location of zchunk zdicts (if zchunk is enabled)
++ * @param zck_auto_chunk Whether zchunk file should be auto-chunked
++ * @param err GError **
++ * @return cr_Error return code
++ */
++int cr_compress_file_with_stat_v2(const char *src,
++ const char *dst,
++ cr_CompressionType comtype,
++ cr_ContentStat *stat,
++ const char *zck_dict_dir,
++ gboolean zck_auto_chunk,
++ GError **err);
++
+ /** Decompress file.
+ * @param SRC source filename
+ * @param DST destination (If dst is dir, filename of src without
+diff --git a/src/modifyrepo_shared.c b/src/modifyrepo_shared.c
+index 4e59660..8cf246d 100644
+--- a/src/modifyrepo_shared.c
++++ b/src/modifyrepo_shared.c
+@@ -120,8 +120,8 @@ cr_write_file(gchar *repopath, cr_ModifyRepoTask *task,
+ g_debug("%s: Copy & compress operation %s -> %s",
+ __func__, src_fn, dst_fn);
+
+- if (cr_compress_file(src_fn, dst_fn, compress_type,
+- task->zck_dict_dir, TRUE, err) != CRE_OK) {
++ if (cr_compress_file_v2(src_fn, dst_fn, compress_type,
++ task->zck_dict_dir, TRUE, err) != CRE_OK) {
+ g_debug("%s: Copy & compress operation failed", __func__);
+ return NULL;
+ }
+diff --git a/src/python/misc-py.c b/src/python/misc-py.c
+index 6a7871e..cc28448 100644
+--- a/src/python/misc-py.c
++++ b/src/python/misc-py.c
+@@ -49,7 +49,7 @@ py_compress_file_with_stat(G_GNUC_UNUSED PyObject *self, PyObject *args)
+ return NULL;
+ }
+
+- cr_compress_file_with_stat(src, dst, type, contentstat, NULL, FALSE, &tmp_err);
++ cr_compress_file_with_stat_v2(src, dst, type, contentstat, NULL, FALSE, &tmp_err);
+ if (tmp_err) {
+ nice_exception(&tmp_err, NULL);
+ return NULL;
+diff --git a/src/threads.c b/src/threads.c
+index 9ef839d..91b5430 100644
+--- a/src/threads.c
++++ b/src/threads.c
+@@ -100,13 +100,13 @@ cr_compressing_thread(gpointer data, G_GNUC_UNUSED gpointer user_data)
+ cr_compression_suffix(task->type),
+ NULL);
+
+- cr_compress_file_with_stat(task->src,
+- task->dst,
+- task->type,
+- task->stat,
+- task->zck_dict_dir,
+- task->zck_auto_chunk,
+- &tmp_err);
++ cr_compress_file_with_stat_v2(task->src,
++ task->dst,
++ task->type,
++ task->stat,
++ task->zck_dict_dir,
++ task->zck_auto_chunk,
++ &tmp_err);
+
+ if (tmp_err) {
+ // Error encountered
+diff --git a/tests/test_misc.c b/tests/test_misc.c
+index 6614809..1acccb7 100644
+--- a/tests/test_misc.c
++++ b/tests/test_misc.c
+@@ -548,8 +548,8 @@ compressfile_test_text_file(Copyfiletest *copyfiletest,
+ GError *tmp_err = NULL;
+
+ g_assert(!g_file_test(copyfiletest->dst_file, G_FILE_TEST_EXISTS));
+- ret = cr_compress_file(TEST_TEXT_FILE, copyfiletest->dst_file,
+- CR_CW_GZ_COMPRESSION, NULL, FALSE, &tmp_err);
++ ret = cr_compress_file_v2(TEST_TEXT_FILE, copyfiletest->dst_file,
++ CR_CW_GZ_COMPRESSION, NULL, FALSE, &tmp_err);
+ g_assert(!tmp_err);
+ g_assert_cmpint(ret, ==, CRE_OK);
+ g_assert(g_file_test(copyfiletest->dst_file, G_FILE_TEST_IS_REGULAR));
+@@ -574,9 +574,9 @@ compressfile_with_stat_test_text_file(Copyfiletest *copyfiletest,
+ g_assert(!tmp_err);
+
+ g_assert(!g_file_test(copyfiletest->dst_file, G_FILE_TEST_EXISTS));
+- ret = cr_compress_file_with_stat(TEST_TEXT_FILE, copyfiletest->dst_file,
+- CR_CW_GZ_COMPRESSION, stat, NULL, FALSE,
+- &tmp_err);
++ ret = cr_compress_file_with_stat_v2(TEST_TEXT_FILE, copyfiletest->dst_file,
++ CR_CW_GZ_COMPRESSION, stat, NULL, FALSE,
++ &tmp_err);
+ g_assert(!tmp_err);
+ g_assert_cmpint(ret, ==, CRE_OK);
+ g_assert(g_file_test(copyfiletest->dst_file, G_FILE_TEST_IS_REGULAR));
+@@ -603,9 +603,9 @@ compressfile_with_stat_test_gz_file_gz_output(Copyfiletest *copyfiletest,
+ char * dst_full_name = g_strconcat(copyfiletest->dst_file, ".gz", NULL);
+
+ g_assert(!g_file_test(dst_full_name, G_FILE_TEST_EXISTS));
+- ret = cr_compress_file_with_stat(TEST_TEXT_FILE_GZ, dst_full_name,
+- CR_CW_GZ_COMPRESSION, stat, NULL, FALSE,
+- &tmp_err);
++ ret = cr_compress_file_with_stat_v2(TEST_TEXT_FILE_GZ, dst_full_name,
++ CR_CW_GZ_COMPRESSION, stat, NULL, FALSE,
++ &tmp_err);
+ g_assert(!tmp_err);
+ g_assert_cmpint(ret, ==, CRE_OK);
+ g_assert(g_file_test(dst_full_name, G_FILE_TEST_IS_REGULAR));
+@@ -633,9 +633,9 @@ compressfile_test_gz_file_xz_output(Copyfiletest *copyfiletest,
+ char * dst_full_name = g_strconcat(copyfiletest->dst_file, ".xz", NULL);
+
+ g_assert(!g_file_test(dst_full_name, G_FILE_TEST_EXISTS));
+- ret = cr_compress_file(TEST_TEXT_FILE_GZ, dst_full_name,
+- CR_CW_XZ_COMPRESSION, NULL, FALSE,
+- &tmp_err);
++ ret = cr_compress_file_v2(TEST_TEXT_FILE_GZ, dst_full_name,
++ CR_CW_XZ_COMPRESSION, NULL, FALSE,
++ &tmp_err);
+ g_assert(!tmp_err);
+ g_assert_cmpint(ret, ==, CRE_OK);
+ g_assert(g_file_test(dst_full_name, G_FILE_TEST_IS_REGULAR));
+@@ -660,9 +660,9 @@ compressfile_test_xz_file_gz_output(Copyfiletest *copyfiletest,
+ char * dst_full_name = g_strconcat(copyfiletest->dst_file, ".gz", NULL);
+
+ g_assert(!g_file_test(dst_full_name, G_FILE_TEST_EXISTS));
+- ret = cr_compress_file(TEST_TEXT_FILE_XZ, dst_full_name,
+- CR_CW_GZ_COMPRESSION, NULL, FALSE,
+- &tmp_err);
++ ret = cr_compress_file_v2(TEST_TEXT_FILE_XZ, dst_full_name,
++ CR_CW_GZ_COMPRESSION, NULL, FALSE,
++ &tmp_err);
+ g_assert(!tmp_err);
+ g_assert_cmpint(ret, ==, CRE_OK);
+ g_assert(g_file_test(dst_full_name, G_FILE_TEST_IS_REGULAR));
+@@ -687,9 +687,9 @@ compressfile_test_sqlite_file_gz_output(Copyfiletest *copyfiletest,
+ char * dst_full_name = g_strconcat(copyfiletest->dst_file, ".gz", NULL);
+
+ g_assert(!g_file_test(dst_full_name, G_FILE_TEST_EXISTS));
+- ret = cr_compress_file(TEST_SQLITE_FILE, dst_full_name,
+- CR_CW_GZ_COMPRESSION, NULL, FALSE,
+- &tmp_err);
++ ret = cr_compress_file_v2(TEST_SQLITE_FILE, dst_full_name,
++ CR_CW_GZ_COMPRESSION, NULL, FALSE,
++ &tmp_err);
+ g_assert(!tmp_err);
+ g_assert_cmpint(ret, ==, CRE_OK);
+ g_assert(g_file_test(dst_full_name, G_FILE_TEST_EXISTS));
+--
+libgit2 1.0.1
+
diff --git a/SPECS/createrepo_c.spec b/SPECS/createrepo_c.spec
index fd89134..3a1f3b1 100644
--- a/SPECS/createrepo_c.spec
+++ b/SPECS/createrepo_c.spec
@@ -25,11 +25,13 @@
Summary: Creates a common metadata repository
Name: createrepo_c
Version: 0.17.2
-Release: 1.1%{?dist}
+Release: 3.1%{?dist}
License: GPLv2+
URL: https://github.com/rpm-software-management/createrepo_c
Source0: %{url}/archive/%{version}/%{name}-%{version}.tar.gz
-Patch0: https://github.com/rpm-software-management/createrepo_c/pull/269.patch
+Patch1: 0001-Fix-cr_compress_file_with_stat-Memory-leak.patch
+Patch2: 0002-Preserve-changed-API-for-cr_compress_file_with_stat-RhBug1973588.patch
+Patch3: https://github.com/rpm-software-management/createrepo_c/pull/269.patch
Provides: createrepo_c(pr269)
BuildRequires: cmake
@@ -177,9 +179,18 @@ ln -sr %{buildroot}%{_bindir}/modifyrepo_c %{buildroot}%{_bindir}/modifyrepo
%{python3_sitearch}/%{name}-%{version}-py%{python3_version}.egg-info
%changelog
+* Wed Sep 27 2021 Matthew Almond <malmond@fb.com> - 0.17.2-3.1
+- Update to 0.17.2-3 with PR269
+
+* Wed Sep 15 2021 Pavla Kratochvilova <pkratoch@redhat.com> - 0.17.2-3
+- Preserve changed API for cr_compress_file_with_stat (RhBug:1973588)
+
* Wed Aug 04 2021 Matthew Almond <malmond@fb.com> - 0.17.2-1.1
- Update to 0.17.2-1 with PR269
+* Tue Jul 27 2021 Pavla Kratochvilova <pkratoch@redhat.com> - 0.17.2-2
+- Fix: cr_compress_file_with_stat: Memory leak
+
* Mon May 24 2021 Matthew Almond <malmond@fb.com> - 0.16.2-2.1
- PR269 for --skip-filelists