squash. code review

.github/workflows/build-irods-centos.yml

@@ -18,10 +18,11 @@ jobs:
               yum -y install gcc g++ libstdc++-static make rpm-build bzip2-devel curl-devel fakeroot openssl-devel pam-devel python-devel unixODBC unixODBC-devel zlib-devel python36-distro
       - name: Install Flex and Bison
         run:  |
-              yum -y install centos-release-scl-rh
+              #yum -y install centos-release-scl-rh
               rpm --import https://www.cert.org/forensics/repository/forensics-expires-2022-04-03.asc
-              wget https://forensics.cert.org/cert-forensics-tools-release-el7.rpm
-              rpm -i ./cert-forensics-tools-release-el7.rpm
+              #wget https://forensics.cert.org/cert-forensics-tools-release-el7.rpm
+              #rpm -i ./cert-forensics-tools-release-el7.rpm
+              yum -y install https://forensics.cert.org/cert-forensics-tools-release-el7.rpm
               yum -y install flex bison
       - name: Install iRODS Externals
         run:  |

lib/api/include/irods/genquery2.h

@@ -7,20 +7,30 @@ struct RcComm;
 
 /// The input data type used to invoke #rc_genquery2.
 ///
+/// \note This data structure is part of an experimental API endpoint and may change in the future.
+///
 /// \since 4.3.2
 typedef struct GenQuery2Input // NOLINT(modernize-use-using)
 {
-    /// TODO
+    /// The GenQuery2 query string to execute.
+    ///
+    /// This member variable MUST be non-empty.
     ///
     /// \since 4.3.2
     char* query_string;
 
-    /// TODO
+    /// The zone to execute the query string against.
+    ///
+    /// This member variable is allowed to be set to NULL.
     ///
     /// \since 4.3.2
     char* zone;
 
-    /// TODO
+    /// Controls whether the SQL derived from the query string is
+    /// executed or returned to the caller.
+    ///
+    /// When set to 0, the generated SQL will be executed.
+    /// When set to 1, the generated SQL will be returned to the caller. The SQL will not be executed.
     ///
     /// \since 4.3.2
     int sql_only;
@@ -33,7 +43,70 @@ typedef struct GenQuery2Input // NOLINT(modernize-use-using)
 extern "C" {
 #endif
 
-/// TODO
+/// Query the catalog using the GenQuery2 parser.
+///
+/// \note This is an experimental API endpoint and may change in the future.
+///
+/// The GenQuery2 parser supports the following:
+/// - Enforces the iRODS permission model
+/// - Logical AND, OR, and NOT
+/// - Grouping via parentheses
+/// - SQL CAST
+/// - SQL GROUP BY
+/// - SQL aggregate functions (e.g. count, sum, avg, etc)
+/// - Per-column sorting via ORDER BY [ASC|DESC]
+/// - SQL FETCH FIRST <i>N</i> ROWS ONLY (LIMIT offered as an alias)
+/// - Metadata queries involving different iRODS entities (i.e. data objects, collections, users, and resources)
+/// - Operators: =, !=, <, <=, >, >=, LIKE, BETWEEN, IS [NOT] NULL
+/// - SQL keywords are case-insensitive
+/// - Federation
+/// - Escaping of single quotes
+/// - Bytes encoded as hexadecimal (e.g. \x21)
+///
+/// Limitations:
+/// - Groups are not yet fully supported
+/// - Cannot resolve tickets to data objects and collections using a single query
+/// - Integer values must be treated as strings, except when used for OFFSET, LIMIT, FETCH FIRST <i>N</i> ROWS ONLY
+///
+/// \param[in]     _comm   A pointer to a RcComm.
+/// \param[in]     _input  A pointer to a GenQuery2Input.
+/// \param[in,out] _output \parblock A pointer that will hold the results of the operation.
+/// On success, the pointer will either hold a JSON string or a string representing the SQL derived from
+/// the GenQuery2 query string. See GenQuery2Input::sql_only for details.
+///
+/// The string is always heap-allocated and must be free'd by the caller using free().
+///
+/// On failure, the pointer will be NULL.
+/// \endparblock
+///
+/// \return An integer.
+/// \retval  0 On success.
+/// \retval <0 On failure.
+///
+/// \b Example
+/// \code{.cpp}
+/// RcComm* comm = // Our iRODS connection.
+///
+/// // Configure the input object for the API call.
+/// struct GenQuery2Input input;
+/// memset(&input, 0, sizeof(struct GenQuery2Input));
+///
+/// // This is the query that will be executed (i.e. input.sql_only is set to 0).
+/// input.query_string = strdup("select COLL_NAME, DATA_NAME where RESC_ID = '10016'");
+///
+/// char* output = NULL;
+/// const int ec = rc_genquery2(comm, &input, &output);
+///
+/// // Handle error.
+/// if (ec < 0) {
+///     free(input.query_string);
+///     return;
+/// }
+///
+/// // At this point, "output" should represent a JSON string.
+/// // Parse the string as JSON and inspect the results. 
+/// // If "input.sql_only" was set to 1, "output" would hold the SQL derived from the GenQuery2 syntax.
+/// \endcode
 ///
 /// \since 4.3.2
 int rc_genquery2(struct RcComm* _comm, struct GenQuery2Input* _input, char** _output);

lib/core/include/irods/irods_configuration_keywords.hpp

@@ -49,6 +49,7 @@ namespace irods
     extern const char* const KW_CFG_LOG_LEVEL_CATEGORY_AGENT_FACTORY;
     extern const char* const KW_CFG_LOG_LEVEL_CATEGORY_AGENT;
     extern const char* const KW_CFG_LOG_LEVEL_CATEGORY_DELAY_SERVER;
+    extern const char* const KW_CFG_LOG_LEVEL_CATEGORY_GENQUERY2;
     extern const char* const KW_CFG_LOG_LEVEL_CATEGORY_RESOURCE;
     extern const char* const KW_CFG_LOG_LEVEL_CATEGORY_DATABASE;
     extern const char* const KW_CFG_LOG_LEVEL_CATEGORY_AUTHENTICATION;

lib/core/include/irods/irods_logger.hpp

@@ -87,6 +87,7 @@ namespace irods::experimental::log
         struct agent_factory {};
         struct agent {};
         struct delay_server {};
+        struct genquery2 {};
         struct resource {};
         struct database {};
         struct authentication {};
@@ -141,6 +142,7 @@ namespace irods::experimental::log
     using agent_factory  = logger<category::agent_factory>;
     using agent          = logger<category::agent>;
     using delay_server   = logger<category::delay_server>;
+    using genquery2      = logger<category::genquery2>;
     using resource       = logger<category::resource>;
     using database       = logger<category::database>;
     using authentication = logger<category::authentication>;
@@ -834,6 +836,15 @@ namespace irods::experimental::log
         friend class logger<category::delay_server>;
     }; // class logger_config<category::delay_server>
 
+    template <>
+    class logger_config<category::genquery2>
+    {
+        static constexpr const char* const name = "genquery2";
+        inline static level level = level::info; // NOLINT(cppcoreguidelines-avoid-non-const-global-variables)
+
+        friend class logger<category::genquery2>;
+    }; // class logger_config<category::genquery2>
+
     template <>
     class logger_config<category::resource>
     {

lib/core/src/irods_configuration_keywords.cpp

@@ -47,6 +47,7 @@ namespace irods
     const char* const KW_CFG_LOG_LEVEL_CATEGORY_AGENT_FACTORY{"agent_factory"};
     const char* const KW_CFG_LOG_LEVEL_CATEGORY_AGENT{"agent"};
     const char* const KW_CFG_LOG_LEVEL_CATEGORY_DELAY_SERVER{"delay_server"};
+    const char* const KW_CFG_LOG_LEVEL_CATEGORY_GENQUERY2{"genquery2"};
     const char* const KW_CFG_LOG_LEVEL_CATEGORY_RESOURCE{"resource"};
     const char* const KW_CFG_LOG_LEVEL_CATEGORY_DATABASE{"database"};
     const char* const KW_CFG_LOG_LEVEL_CATEGORY_AUTHENTICATION{"authentication"};

scripts/core_tests_list.json

@@ -68,6 +68,7 @@
     "test_ipwd",
     "test_iqmod",
     "test_iqstat",
+    "test_iquery",
     "test_iquest.Test_Iquest",
     "test_iquest.test_iquest_logical_or_operator_with_data_resc_hier",
     "test_iquest.test_iquest_with_data_resc_hier",

scripts/irods/test/test_genquery2_microservices.py

@@ -1,10 +1,7 @@
-import os
-import sys
 import textwrap
 import unittest
 
 from . import session
-from .. import lib
 from ..configuration import IrodsConfig
 
 rodsadmins = [('otherrods', 'rods')]

scripts/irods/test/test_iquery.py

@@ -0,0 +1,24 @@
+import json
+import unittest
+
+from . import session
+from .. import lib
+
+rodsadmins = [('otherrods', 'rods')]
+rodsusers  = [('alice', 'apass')]
+
+class Test_IQuery(session.make_sessions_mixin(rodsadmins, rodsusers), unittest.TestCase):
+
+    def setUp(self):
+        super(Test_IQuery, self).setUp()
+
+        self.admin = self.admin_sessions[0]
+        self.user = self.user_sessions[0]
+
+    def tearDown(self):
+        super(Test_IQuery, self).tearDown()
+
+    def test_iquery_can_run_a_query__issue_7570(self):
+        json_string = json.dumps([[self.user.session_collection]])
+        self.user.assert_icommand(
+            ['iquery', f"select COLL_NAME where COLL_NAME = '{self.user.session_collection}'"], 'STDOUT', [json_string])

server/api/include/irods/rs_genquery2.hpp

@@ -7,7 +7,70 @@
 
 struct RsComm;
 
-/// TODO
+/// Query the catalog using the GenQuery2 parser.
+///
+/// \note This is an experimental API endpoint and may change in the future.
+///
+/// The GenQuery2 parser supports the following:
+/// - Enforces the iRODS permission model
+/// - Logical AND, OR, and NOT
+/// - Grouping via parentheses
+/// - SQL CAST
+/// - SQL GROUP BY
+/// - SQL aggregate functions (e.g. count, sum, avg, etc)
+/// - Per-column sorting via ORDER BY [ASC|DESC]
+/// - SQL FETCH FIRST <i>N</i> ROWS ONLY (LIMIT offered as an alias)
+/// - Metadata queries involving different iRODS entities (i.e. data objects, collections, users, and resources)
+/// - Operators: =, !=, <, <=, >, >=, LIKE, BETWEEN, IS [NOT] NULL
+/// - SQL keywords are case-insensitive
+/// - Federation
+/// - Escaping of single quotes
+/// - Bytes encoded as hexadecimal (e.g. \x21)
+///
+/// Limitations:
+/// - Groups are not yet fully supported
+/// - Cannot resolve tickets to data objects and collections using a single query
+/// - Integer values must be treated as strings, except when used for OFFSET, LIMIT, FETCH FIRST <i>N</i> ROWS ONLY
+///
+/// \param[in]     _comm   A pointer to a RsComm.
+/// \param[in]     _input  A pointer to a GenQuery2Input.
+/// \param[in,out] _output \parblock A pointer that will hold the results of the operation.
+/// On success, the pointer will either hold a JSON string or a string representing the SQL derived from
+/// the GenQuery2 query string. See GenQuery2Input::sql_only for details.
+///
+/// The string is always heap-allocated and must be free'd by the caller using free().
+///
+/// On failure, the pointer will be NULL.
+/// \endparblock
+///
+/// \return An integer.
+/// \retval  0 On success.
+/// \retval <0 On failure.
+///
+/// \b Example
+/// \code{.cpp}
+/// RsComm* comm = // Our iRODS connection.
+///
+/// // Configure the input object for the API call.
+/// struct GenQuery2Input input;
+/// memset(&input, 0, sizeof(struct GenQuery2Input));
+///
+/// // This is the query that will be executed (i.e. input.sql_only is set to 0).
+/// input.query_string = strdup("select COLL_NAME, DATA_NAME where RESC_ID = '10016'");
+///
+/// char* output = NULL;
+/// const int ec = rc_genquery2(comm, &input, &output);
+///
+/// // Handle error.
+/// if (ec < 0) {
+///     free(input.query_string);
+///     return;
+/// }
+///
+/// // At this point, "output" should represent a JSON string.
+/// // Parse the string as JSON and inspect the results. 
+/// // If "input.sql_only" was set to 1, "output" would hold the SQL derived from the GenQuery2 syntax.
+/// \endcode
 ///
 /// \since 4.3.2
 int rs_genquery2(RsComm* _comm, GenQuery2Input* _input, char** _output);

server/api/src/rs_genquery2.cpp

@@ -8,22 +8,17 @@
 #include "irods/genquery2_sql.hpp"
 
 #include "irods/apiHandler.hpp"
-//#include "irods/catalog.hpp" // Requires linking against libnanodbc.so
 #include "irods/irods_logger.hpp"
 #include "irods/irods_rs_comm_query.hpp"
 #include "irods/irods_server_properties.hpp"
 #include "irods/irods_version.h"
-//#include "irods/procApiRequest.h"
 #include "irods/rodsConnect.h"
 #include "irods/rodsDef.h"
 #include "irods/rodsErrorTable.h"
 #include "irods/icatHighLevelRoutines.hpp"
 
-//#include <fmt/format.h>
-//#include <nanodbc/nanodbc.h>
 #include <nlohmann/json.hpp>
 
-//#include <cstdlib> // For std::malloc.
 #include <cstring> // For strdup.
 #include <string_view>
 #include <vector>
@@ -52,6 +47,8 @@ auto rs_genquery2(RsComm* _comm, GenQuery2Input* _input, char** _output) -> int
         log_api::trace("{}: Received: query_string=[{}], zone=[nullptr]", __func__, _input->query_string);
     }
 
+    *_output = nullptr;
+
     rodsServerHost* host_info{};
 
     if (const auto ec = getAndConnRcatHost(_comm, PRIMARY_RCAT, _input->zone, &host_info); ec < 0) {

server/core/src/irods_api_calling_functions.cpp

@@ -1,4 +1,3 @@
-#include "irods/genquery2.h"
 #include "irods/rcConnect.h"
 #include "irods/apiHeaderAll.h"
 #include "irods/apiHandler.hpp"

server/core/src/rodsAgent.cpp

@@ -314,6 +314,7 @@ void set_log_levels_for_all_log_categories()
     log_ns::agent::set_level(log_ns::get_level_from_config(irods::KW_CFG_LOG_LEVEL_CATEGORY_AGENT));
     log_ns::legacy::set_level(log_ns::get_level_from_config(irods::KW_CFG_LOG_LEVEL_CATEGORY_LEGACY));
     log_ns::resource::set_level(log_ns::get_level_from_config(irods::KW_CFG_LOG_LEVEL_CATEGORY_RESOURCE));
+    log_ns::genquery2::set_level(log_ns::get_level_from_config(irods::KW_CFG_LOG_LEVEL_CATEGORY_GENQUERY2));
     log_ns::database::set_level(log_ns::get_level_from_config(irods::KW_CFG_LOG_LEVEL_CATEGORY_DATABASE));
     log_ns::authentication::set_level(log_ns::get_level_from_config(irods::KW_CFG_LOG_LEVEL_CATEGORY_AUTHENTICATION));
     log_ns::api::set_level(log_ns::get_level_from_config(irods::KW_CFG_LOG_LEVEL_CATEGORY_API));