This is an automated email from the ASF dual-hosted git repository.
dbarnes pushed a commit to branch develop
in repository https://gitbox.apache.org/repos/asf/geode-native.git
The following commit(s) were added to refs/heads/develop by this push:
new ac0c1f0 GEODE-4728: update serialization docs - WIP
ac0c1f0 is described below
commit ac0c1f0771bd7d4de8a90bea3c5f3a2904ceebf1
Author: Dave Barnes <[email protected]>
AuthorDate: Fri Jul 20 14:34:52 2018 -0700
GEODE-4728: update serialization docs - WIP
---
.../pdxserializable-example.html.md.erb | 22 ++++------------------
1 file changed, 4 insertions(+), 18 deletions(-)
diff --git
a/docs/geode-native-docs/serialization/cpp-serialization/pdxserializable-example.html.md.erb
b/docs/geode-native-docs/serialization/cpp-serialization/pdxserializable-example.html.md.erb
index 780b138..714980c 100644
---
a/docs/geode-native-docs/serialization/cpp-serialization/pdxserializable-example.html.md.erb
+++
b/docs/geode-native-docs/serialization/cpp-serialization/pdxserializable-example.html.md.erb
@@ -19,10 +19,6 @@ See the License for the specific language governing
permissions and
limitations under the License.
-->
-DRAFT - Dave's notes interspersed.
-
-Why is this called "customserializable", rather than "PdxSerializable"?
-
## Order.cpp:
Front matter:
@@ -37,9 +33,7 @@ Front matter:
namespace customserializable {
- Must provide `fromData()` and `toData()`.
-- Specify the keys in the same order in fromData and toData.
-- The order (no pun intended) in which the keys appear IS SIGNIFICANT in PDX
Serialization, though it probably
-shouldn't be since we're supposedly indexing by name, not by position in the
byte stream.
+- Specify the keys in the same order in `fromData()` and `toData()`. Use
`PdxReader` and `PdxWriter`:
```
void Order::fromData(PdxReader& pdxReader) {
@@ -60,14 +54,13 @@ shouldn't be since we're supposedly indexing by name, not
by position in the byt
}
```
-Ivan says "toString is an override." What does he mean by that? Required by
interface spec, or what?
+Optionally override and add detailed methods as needed. In this example, we
override `toString()` and implement `size()` and `classname()`.
std::string Order::toString() const {
return "OrderID: " + std::to_string(order_id_) + " Product Name: " +
name_ +
" Quantity: " + std::to_string(quantity_);
}
-Do we need size?
size_t Order::objectSize() const {
auto objectSize = sizeof(Order);
@@ -75,14 +68,13 @@ Do we need size?
return objectSize;
}
-Do we need classname?
const std::string& Order::getClassName() const {
static const std::string CLASS_NAME = "com.example.Order";
return CLASS_NAME;
}
-Why `createDeserializable`?
+Provide a constructor in a parameter-less wrapper that can be called to create
a prototypical instance (called `createDeserializable()` here).
std::shared_ptr<PdxSerializable> Order::createDeserializable() {
return std::make_shared<Order>(0, "", 0);
@@ -96,7 +88,7 @@ Why `createDeserializable`?
## main.cpp
-Front matter. Any of these **required** for serialization (e.g. TypeRegistry)?
+Front matter. You will need `TypeRegistry` to register your PdxSerializable
class. Bring in the customary includes for I/O and cache, pool, and region
creation.
#include <iostream>
#include <sstream>
@@ -109,12 +101,9 @@ Front matter. Any of these **required** for serialization
(e.g. TypeRegistry)?
#include "Order.hpp"
-Is `namespace apache::geode::client;` required for a client app?
-
using namespace apache::geode::client;
using namespace customserializable;
-Safe to assume that region & pool creation follow a template?
int main(int argc, char** argv) {
auto cacheFactory = CacheFactory();
@@ -128,11 +117,9 @@ Safe to assume that region & pool creation follow a
template?
auto regionFactory = cache.createRegionFactory(RegionShortcut::PROXY);
auto region = regionFactory.setPoolName("pool").create("custom_orders");
-Again, why `createDeserializable`?
cache.getTypeRegistry().registerPdxType(Order::createDeserializable);
-Invoke the example code, report on progress...
std::cout << "Create orders" << std::endl;
auto order1 = std::make_shared<Order>(1, "product x", 23);
@@ -159,7 +146,6 @@ Invoke the example code, report on progress...
std::cout << "Order 2 not found." << std::endl;
}
-Is `cache.close();` conventional or required?
cache.close();
}
