On Mon, 17 Jun 2024 05:54:37 GMT, Renjith Kannath Pariyangad
<rkannathp...@openjdk.org> wrote:
> Hi Reviewers,
>
> This fix will resolve page range not printing proper pages if the rage begin
> from 2 or above on Mac machines.
> I have verified the manual range related tests like PageRanges.java,
> ClippedImages.java and test.java and confirmed its fixing the issue.
>
> Please review and let me know your suggestions if any.
I added some traces and ran the test `PageRanges.java` with and without your
proposed changes.
The diff:
diff --git a/src/java.desktop/macosx/classes/sun/lwawt/macosx/CPrinterJob.java
b/src/java.desktop/macosx/classes/sun/lwawt/macosx/CPrinterJob.java
index 416a3ee002b..6ddf46896b2 100644
--- a/src/java.desktop/macosx/classes/sun/lwawt/macosx/CPrinterJob.java
+++ b/src/java.desktop/macosx/classes/sun/lwawt/macosx/CPrinterJob.java
@@ -33,6 +33,7 @@
import java.net.URI;
import java.security.AccessController;
import java.security.PrivilegedAction;
+import java.util.Arrays;
import java.util.concurrent.atomic.AtomicReference;
import javax.print.*;
@@ -325,6 +325,8 @@ public void print(PrintRequestAttributeSet attributes)
throws PrinterException {
lastPage = mDocument.getNumberOfPages() - 1;
}
}
+ System.out.println("print: firstPage = " + firstPage + "; "
+ + "lastPage = " + lastPage);
try {
synchronized (this) {
@@ -338,7 +340,11 @@ public void print(PrintRequestAttributeSet attributes)
throws PrinterException {
:
(PageRanges)attributes.get(PageRanges.class);
int[][] prMembers = (pr == null) ? new int[0][0] : pr.getMembers();
int loopi = 0;
+ System.out.println("pr = " + pr);
+ System.out.println("prMembers = " +
Arrays.deepToString(prMembers));
do {
+ System.out.println("printLoop: firstPage = " + firstPage + "; "
+ + "lastPage = " + lastPage);
if (EventQueue.isDispatchThread()) {
// This is an AWT EventQueue, and this print rendering
loop needs to block it.
The output:
2-3 + 3-5 without Renjith changes
print: firstPage = 1; lastPage = 2
pr = 2-3
prMembers = [[2, 3]]
printLoop: firstPage = 1; lastPage = 2
print: firstPage = 2; lastPage = 4
pr = 3-5
prMembers = [[3, 5]]
printLoop: firstPage = 2; lastPage = 4
2-3 + 3-5 with Renjith changes
print: firstPage = 0; lastPage = -1
pr = 2-3
prMembers = [[2, 3]]
printLoop: firstPage = 0; lastPage = -1
print: firstPage = 0; lastPage = -1
pr = 3-5
prMembers = [[3, 5]]
printLoop: firstPage = 0; lastPage = -1
In the first case, I got page 3 and page 5 printed; with Renjith's changes, I
got the correct range printed 2-3 and 3-5.
So it works… kind of. I believe the bug is somewhere in native code.
What would be printed if `PageRanges` contains several ranges? This case should
be supported as well.
Since the last page to be printed is -1, `PrinterView->knowsPageRange` sets
`aRange->length` to `NSIntegerMax`. This does not look right, even though it
works.
https://github.com/openjdk/jdk/blob/08ace27da1d9cd215c77471eabf41417ff6282d2/src/java.desktop/macosx/native/libawt_lwawt/awt/PrinterView.m#L130-L152
Then, the Print dialog itself: with Renjith's changes the range of pages is
preserved. Without the changes, the range of pages isn't preserved. It's a
different bug: the dialog should be initialised using the attributes.
If `PageRanges` attribute is set and is supported, the dialog should select
**Range** radio button and set the correct page range.
A new test could be written (if it doesn't exist already) to confirm whether
attributes are taken into account when initialising the Print dialog and then a
new bug should be submitted.
-------------
PR Comment: https://git.openjdk.org/jdk/pull/19740#issuecomment-2182540891
PR Comment: https://git.openjdk.org/jdk/pull/19740#issuecomment-2182543047
