Re: [PR] Guard Mapper.write against null writer [johnzon]
jungm merged PR #123: URL: https://github.com/apache/johnzon/pull/123 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
Re: [PR] Guard Mapper.write against null writer [johnzon]
pingpingy1 commented on PR #123: URL: https://github.com/apache/johnzon/pull/123#issuecomment-2007101918 @rmannibucau No worries! Just pushed those changes. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
Re: [PR] Guard Mapper.write against null writer [johnzon]
rmannibucau commented on PR #123: URL: https://github.com/apache/johnzon/pull/123#issuecomment-2006761132 Ok for me, to be picky (not a blocker) I'd like to get the param and field `final` in the delegating classes and `// no-op` in the payload instead of aside the method (an autoreformat will break it I think in most IDE) but can be handled after the merge easily - I can even do it for you if you feel you did enough (I fully agree, thanks for that ;)). -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
Re: [PR] Guard Mapper.write against null writer [johnzon]
pingpingy1 commented on PR #123: URL: https://github.com/apache/johnzon/pull/123#issuecomment-2006431214 @rmannibucau Implemented all four delegated classes! PTAL -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
Re: [PR] Guard Mapper.write against null writer [johnzon]
pingpingy1 commented on code in PR #123:
URL: https://github.com/apache/johnzon/pull/123#discussion_r1529954373
##
johnzon-mapper/src/main/java/org/apache/johnzon/mapper/internal/DelegatingWriter.java:
##
@@ -0,0 +1,81 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ *http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+package org.apache.johnzon.mapper.internal;
+
+import java.io.IOException;
+import java.io.Writer;
+import java.util.Objects;
+
+/* A delegating class for a no-close Writer */
+public class DelegatingWriter extends Writer {
+private Writer writer;
+
+public DelegatingWriter(Writer writer) {
+this.writer = Objects.requireNonNull(writer, "writer");
+}
+
+@Override
+public void write(final int c) throws IOException {
+writer.write(c);
+}
+
+@Override
+public void write(final char[] cbuf) throws IOException {
+writer.write(cbuf);
+}
+
+@Override
+public void write(final char[] cbuf, final int off, final int len) throws
IOException {
+writer.write(cbuf, off, len);
+}
+
+@Override
+public void write(final String str) throws IOException {
+writer.write(str);
+}
+
+@Override
+public void write(final String str, final int off, final int len) throws
IOException {
+writer.write(str, off, len);
+}
+
+@Override
+public Writer append(final CharSequence csq) throws IOException {
+return writer.append(csq);
+}
+
+@Override
+public Writer append(final CharSequence csq, final int start, final int
end) throws IOException {
+return writer.append(csq, start, end);
+}
+
+@Override
+public Writer append(final char c) throws IOException {
+return writer.append(c);
+}
+
+@Override
+public void flush() throws IOException {
+writer.flush();
+}
+
+// Flushes instead of closing
+@Override
+public void close() throws IOException {
Review Comment:
I got what you mean. Let me make those changes and notify you then.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
Re: [PR] Guard Mapper.write against null writer [johnzon]
rmannibucau commented on code in PR #123:
URL: https://github.com/apache/johnzon/pull/123#discussion_r1529829826
##
johnzon-mapper/src/main/java/org/apache/johnzon/mapper/internal/DelegatingWriter.java:
##
@@ -0,0 +1,81 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ *http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+package org.apache.johnzon.mapper.internal;
+
+import java.io.IOException;
+import java.io.Writer;
+import java.util.Objects;
+
+/* A delegating class for a no-close Writer */
+public class DelegatingWriter extends Writer {
+private Writer writer;
+
+public DelegatingWriter(Writer writer) {
+this.writer = Objects.requireNonNull(writer, "writer");
+}
+
+@Override
+public void write(final int c) throws IOException {
+writer.write(c);
+}
+
+@Override
+public void write(final char[] cbuf) throws IOException {
+writer.write(cbuf);
+}
+
+@Override
+public void write(final char[] cbuf, final int off, final int len) throws
IOException {
+writer.write(cbuf, off, len);
+}
+
+@Override
+public void write(final String str) throws IOException {
+writer.write(str);
+}
+
+@Override
+public void write(final String str, final int off, final int len) throws
IOException {
+writer.write(str, off, len);
+}
+
+@Override
+public Writer append(final CharSequence csq) throws IOException {
+return writer.append(csq);
+}
+
+@Override
+public Writer append(final CharSequence csq, final int start, final int
end) throws IOException {
+return writer.append(csq, start, end);
+}
+
+@Override
+public Writer append(final char c) throws IOException {
+return writer.append(c);
+}
+
+@Override
+public void flush() throws IOException {
+writer.flush();
+}
+
+// Flushes instead of closing
+@Override
+public void close() throws IOException {
Review Comment:
(just explaining for this one but idea is the same for all types) here it
should be `writer.*close()*`, then in `Streams` you do`return new
DelegatingWriter(writer) { @Override public void close() {flush();} }`;
Idea was just to make most of the impl "standard" and just the specific part
"hidden".
side note: even if I'm still not a fan of requireNonNull the minimum would
be a valid error message (`writer = requireNonNull(writer, "delegate writer
must not be null")`) but as mentionned in comment it is always wrong before in
any code path so I would be interested to know how you got it and if the null
check wouldn't sit better before in the codepath.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
Re: [PR] Guard Mapper.write against null writer [johnzon]
pingpingy1 commented on PR #123: URL: https://github.com/apache/johnzon/pull/123#issuecomment-2005319070 @jungm @rmannibucau Not quite sure if I understand the delegation pattern, so I pushed a test commit that moves the anonymous `Writer` class to the `DelegatingWriter` class. Could you please check if this is what you intended? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
Re: [PR] Guard Mapper.write against null writer [johnzon]
rmannibucau commented on PR #123: URL: https://github.com/apache/johnzon/pull/123#issuecomment-2003948812 @jungm maybe let do internal.DelegatingXStream classes (1-1 delegation pattern/decorator) and in `Stream` just override the close by a flush delegation, will enable to "expose" the delegating classes and maintain most of the code there and still prevent anyone to use the "skip the close" impl with an IDE magic trick. > code that is literally in a package named *.internal... it is not uncommon to see sun.internal in imports from time to time and then a "fix commit" and 99% of the time is it cause the IDE imported the wrong class and there it could be a big silent issue (leak with its impacts either in mem or ulimit). So I prefer to not assume it. > capturing variables from an "outside class" makes code super hard to read IMHO guess we are not more at that stage, lambdas made it more than common ;). -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
Re: [PR] Guard Mapper.write against null writer [johnzon]
jungm commented on PR #123: URL: https://github.com/apache/johnzon/pull/123#issuecomment-2003881839 Generally I'm not a friend of inner classes, yet alone anonymous classes - capturing variables from an "outside class" makes code super hard to read IMHO I wouldn't worry too much about users assuming backwards compatibility with code that is literally in a package named `*.internal` tbh -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
Re: [PR] Guard Mapper.write against null writer [johnzon]
rmannibucau commented on PR #123: URL: https://github.com/apache/johnzon/pull/123#issuecomment-2003838137 +1 to just rename the parameter if it helps, having null there should have been identified way earlier in the chain anyway so shouldn't even be possible in terms of code flow. `requireNonNull` does not look required to me. Side note: one required test is that it ends as a `JsonbException` in `johnzon-jsonb` and not a `NullPointerException` or anything else and that the runtime exception thrown in mapper stays of the same type for backward compat. @jungm no big issue for me to make Streams anonymous classes inner classes but I'd like to avoid to make them surfacing as accessible or root classes, they are 100% dedicated to a particular internal need and should be droppable without caring about backward compat - was why it was done this way. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
Re: [PR] Guard Mapper.write against null writer [johnzon]
jungm commented on PR #123: URL: https://github.com/apache/johnzon/pull/123#issuecomment-2003786756 Thanks, I understand a bit more what you want to accomplish now :slightly_smiling_face: Arguably `Cannot invoke "java.io.Writer.write(char[], int, int)" because "this.val$from" is null` can be a bit confusing (a user might think: What is this.val$from suppposed to be? What have I done wrong to make it null?") I do believe renaming this parameter would be enough to clarify what has gone wrong though, adding null guards in the Mapper blows up the code with no real benefit. The exception is thrown immediately and easily debuggable, just not as clear why it is happening right now You want to try going that path? Would also be an awesome opportunity to move the anonymous classes from `org.apache.johnzon.mapper.internal.Streams` into proper classes IMHO -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
Re: [PR] Guard Mapper.write against null writer [johnzon]
pingpingy1 commented on PR #123:
URL: https://github.com/apache/johnzon/pull/123#issuecomment-2002889653
@jungm Thanks for the response. I don't think I was quite thinking right
when I wrote those tests.
The following is the NPE that I intended to address:
```java
public static void main(String args[]) throws Exception {
Mapper mapper = MapperLocatorDelegate.locate();
con_recv0.writeObject((Writer) null, (Object) "string");
}
```
```console
Exception in thread "main" java.lang.NullPointerException: Cannot invoke
"java.io.Writer.write(char[], int, int)" because "this.val$from" is null
at
org.apache.johnzon.mapper.internal.Streams$1.write(Streams.java:45)
at
org.apache.johnzon.core.JsonGeneratorImpl.flushBuffer(JsonGeneratorImpl.java:481)
at
org.apache.johnzon.core.JsonGeneratorImpl.close(JsonGeneratorImpl.java:451)
at org.apache.johnzon.mapper.Mapper.writeObject(Mapper.java:188)
at Test.main(Test.java:17)
```
I know that the explicit `null` here makes the problem obvious, but if it
weren't visible, the NPE is thrown not upon writing but later when flushing the
buffer, which I thought could cause problems in some scenarios.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
Re: [PR] Guard Mapper.write against null writer [johnzon]
jungm commented on PR #123: URL: https://github.com/apache/johnzon/pull/123#issuecomment-1999198687 We're already implicitly throwing NPEs when writer is null by invoking methods on it, don't think there is any need to explicitly guard for this Your test is also already passing without modifying any other source code -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
