gmane.comp.java.openjdk.distro-packaging.devel

Re: [rfc][icedtea-web] Ignore invalid .jar files specified for plugin

Danesh Dadachanji — 2012-05-25T21:19:28

Hi Adam,

On 25/05/12 03:17 PM, Adam Domurad wrote:

Just to summarize what we discussed on IRC, these changes do not belong in JCV. IMO, IcedTea-Web should never think of this extra dir 
as a jar. Going through the stack trace, the verification call happens in JNLPClassLoader#initializeResources. The first line of that 
method (eventually) points to where the change should be, in grabbing the resources from the HTML page. IMO these changes are more 
suited for PluginBridge#getResources[1], right before the names passed into the archive attribute are turned into JARDescs. Keep in 
mind  CodeBaseClassLoader for adding things that are actually in the specified dir. Thanks Omair for pointing this out!

Remember to watch out for the plugin being called using jnlp_href, these dirs should not be accepted then.

There are a few more comments below.


It went from one space to no spaces! :O 2




In general, I would prefer things like refactoring that aren't directly related to your changes to be in separate patches

Re: [rfc][icedtea-web] fix for RH816592

Thomas Meyer — 2012-05-25T21:06:37

Am Donnerstag, den 24.05.2012, 15:41 +0200 schrieb Jiri Vanek:

Hi,

is this the same as PR858?

I mailed this patch to fix it:
http://old.nabble.com/Re%3A--RFC--PR858-p33525848.html

with kind regards
thomas

Re: [rfc][icedtea-web] Ignore invalid .jar files specified for plugin

Adam Domurad — 2012-05-25T19:54:46

So, I realize after some thought that this probably shouldn't be in
JarCertVerifier. However this is where the .jar files are uncompressed
and thus the location where the ZipException could be caught and dealt
with. Suggestions ? To be sure the .jar files are valid they must be
unzipped first, via new JarFile(...).

On Fri, 2012-05-25 at 15:17 -0400, Adam Domurad wrote:

[rfc][icedtea-web] Ignore invalid .jar files specified for plugin

Adam Domurad — 2012-05-25T19:17:48

So, first patch that isn't something trivial, I'll definitely need
people to weigh in on this one. 

The proprietary plug-in seems to just skip over any malformed .jar files
and carry on loading. This patch emulates that behaviour. 
This alleviates some of the symptoms of
http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=1011 
However, the plugin should still parse folders differently and look for
resources in folders like the proprietary plug-in does.

ChangeLog:
2012-05-25  Adam Domurad <adomurad-H+wXaHxf7aLQT0dZR+AlfA< at >public.gmane.org>

Ignore invalid jar files, like the oracle plugin does.
* netx/net/sourceforge/jnlp/tools/JarCertVerifier.java: Added extra 
possible verification result, INVALID_JAR. This value occurs if the
.jar file could not be properly loaded. The .jar file is then ignored.

Re: [rfc] [icedtea-web] get rid of repated sout/serr in reproducerstestcases

Danesh Dadachanji — 2012-05-25T19:08:46

Thanks for letting us know. I have some comments below. Could you also add a ChangeLog entry?




Add a space for indenting please! (Surprised I caught this even, guess it just happened to stand :/  )

[snip]


Indentation of this chunk looks a bit off, could you double check it?

[snip]


s/ooutput/output/
I think also for style, we use single lined comments for class vars - /** <comment-on-one-line */


Could you refactor this so that it's s/PROCES/PROCESS/

Also, there are lots of missing spaces in this patch, e.g. around the '=' above. Could you please space correctly, thanks!


Could you name these logStdErr and logStdOut instead? Either that or add javadoc please. Seeing these outside of this class may throw 
some newcomers off.


Again, could you name the bools printToOut and printToErr?

Thanks for doing this!

Cheers,
Danesh

/hg/icedtea-web: Changed for-loops over iterators and indices to...

2012-05-25T15:44:55

changeset 6df151bb5320 in /hg/icedtea-web
details: http://icedtea.classpath.org/hg/icedtea-web?cmd=changeset;node=6df151bb5320
author: Adam Domurad <adomurad-H+wXaHxf7aLQT0dZR+AlfA< at >public.gmane.org>
date: Fri May 25 11:44:13 2012 -0400

Changed for-loops over iterators and indices to for-each loops if they
were sufficient and clearer.


diffstat:

 ChangeLog                                                         |  18 ++++
 netx/net/sourceforge/jnlp/JNLPFile.java                           |   4 +-
 netx/net/sourceforge/jnlp/PluginBridge.java                       |   4 +-
 netx/net/sourceforge/jnlp/ResourcesDesc.java                      |  17 ++--
 netx/net/sourceforge/jnlp/cache/CacheUtil.java                    |  31 ++++----
 netx/net/sourceforge/jnlp/cache/DefaultDownloadIndicator.java     |   8 +-
 netx/net/sourceforge/jnlp/cache/Resource.java                     |   3 +-
 netx/net/sourceforge/jnlp/cache/ResourceTracker.java              |  37 ++++-----
 netx/net/sourceforge/jnlp/runtime/AppletEnvir

/hg/icedtea-web: Introduced constant NUM_STR_BUFFER_SIZE to repl...

2012-05-25T15:24:26

changeset 082f04822f12 in /hg/icedtea-web
details: http://icedtea.classpath.org/hg/icedtea-web?cmd=changeset;node=082f04822f12
author: Adam Domurad <adomurad-H+wXaHxf7aLQT0dZR+AlfA< at >public.gmane.org>
date: Fri May 25 11:23:47 2012 -0400

Introduced constant NUM_STR_BUFFER_SIZE to replace magic constants.
Clarified some uses of snprintf.


diffstat:

 ChangeLog                              |   13 ++++
 plugin/icedteanp/IcedTeaNPPlugin.cc    |    6 +-
 plugin/icedteanp/IcedTeaPluginUtils.cc |  106 +++++++++++++++-----------------
 plugin/icedteanp/IcedTeaPluginUtils.h  |    4 +
 4 files changed, 71 insertions(+), 58 deletions(-)

diffs (199 lines):

diff -r 9d422db340b8 -r 082f04822f12 ChangeLog
--- a/ChangeLogFri May 25 10:34:24 2012 -0400
+++ b/ChangeLogFri May 25 11:23:47 2012 -0400
< at >< at > -1,3 +1,16 < at >< at >
+2012-05-23  Adam Domurad  <adomurad-H+wXaHxf7aLQT0dZR+AlfA< at >public.gmane.org>
+Removed instances of snprintf where buffer size was not known. Added
+buffer size constant for allocating buffers for numeric c

Re: [rfc][icedtea-web] Simple for-each loop style patch

Pavel Tisnovsky — 2012-05-25T15:25:14

Hi Adam,

your changes are correct, ok for HEAD.

Cheers,
Pavel

vm_start_mutex use in IcedTeaNPPlugin.cc looks fishy

Martin Olsson — 2012-05-23T14:50:19

The usage of "vm_start_mutex" in IcedTeaNPPlugin.cc looks buggy to me.
Won't each thread just create and lock it's own mutex, thereby voiding the
raison d'être of the mutex?


Martin

Re: /hg/icedtea-web: 3 new changesets

Martin Olsson — 2012-05-24T18:43:55

dont_free_stack_var.patch is still not merged.


Martin


On 05/24/2012 04:40 PM, Adam Domurad wrote:

[PATCH] fix mutex related crash + cleanup .hgignore

Martin Olsson — 2012-05-23T14:46:58

Two patches for icedtea-web attached.

which package do I need to avoid: "package org.junit does not exist"

Martin Olsson — 2012-05-23T14:45:06

When I try to run tests using:

        rm -rf ~/ice
        mkdir ~/ice
        ./autogen.sh
        ./configure --prefix=/home/molsson/ice
--with-jdk-home=/usr/lib/jvm/java-6-openjdk-amd64
        make
        make check
        make install
        make run-netx-dist-tests

Then I get this error:

tests/netx/jnlp_testsengine/net/sourceforge/jnlp/ResourcesTest.java:42:
package org.junit does not exist

Full error:

        http://pastebin.com/4MTg03ni

...I assume I need to install some missing package (junit?), any ideas on
which one? I'm on Ubuntu 12.04

Re: RTC Thumb2 JIT enhancements

Andrew Dinn — 2012-05-25T14:40:39

Ok, so this has been reviewed by eyeballing the patch and by building it
and running SpecJVM.

such a thing by eye) modulo one tiny error in thumb2.cpp at line 6141
(that's post patch)

< at >< at > -6558,7 +6137,7 < at >< at >
     r = JSTACK_REG(jstack);
     PUSH(jstack, r);
     if (tos_type == btos)
-      ldrsb_imm(jinfo->codebuf, r, r_obj, field_offset, 1, 0);
+      ldrb_imm(jinfo->codebuf, r, r_obj, field_offset, 1, 0);
     else if (tos_type == ctos)
       ldrh_imm(jinfo->codebuf, r, r_obj, field_offset, 1, 0);
     else if (tos_type == stos)

For some reason the byte load has been changed to unsigned but it really
ought to be signed.

Running SpecJVM unearthed an error which appears to have been there all
along and so is only being made manifest by the patch, viz that stacked
return values are not copied through from old to new (nursery) space if
we enter a garbage collect at a return point in either compiled /or/
interpreted code. They need to be written back to the frame before a
safepoint is entered. Oth

Re: [rfc][icedtea-web] improvements to reproducers engine

Danesh Dadachanji — 2012-05-25T14:39:11


On 25/05/12 03:54 AM, Jiri Vanek wrote:

No problem =) Pushed:

http://icedtea.classpath.org/hg/icedtea-web/rev/9d422db340b8

Cheers,
Danesh

/hg/icedtea-web: Fix whitelisting to use src dir instead of buil...

2012-05-25T14:35:06

changeset 9d422db340b8 in /hg/icedtea-web
details: http://icedtea.classpath.org/hg/icedtea-web?cmd=changeset;node=9d422db340b8
author: Danesh Dadachanji <ddadacha-H+wXaHxf7aLQT0dZR+AlfA< at >public.gmane.org>
date: Fri May 25 10:34:24 2012 -0400

Fix whitelisting to use src dir instead of build dir.


diffstat:

 ChangeLog   |  6 ++++++
 Makefile.am |  2 +-
 2 files changed, 7 insertions(+), 1 deletions(-)

diffs (25 lines):

diff -r 2c84a314c411 -r 9d422db340b8 ChangeLog
--- a/ChangeLogThu May 24 15:37:30 2012 -0400
+++ b/ChangeLogFri May 25 10:34:24 2012 -0400
< at >< at > -1,3 +1,9 < at >< at >
+2012-05-24  Danesh Dadachanji <ddadacha-H+wXaHxf7aLQT0dZR+AlfA< at >public.gmane.org>
+
+Fix use of src dir instead of build dir when whitelisting.
+* Makefile.am (REPRODUCERS_CLASS_WHITELIST): Use abs_top_srcdir
+instead of abs_top_builddir.
+
 2012-05-23  Martin Olsson <martin-Gwph8TgvYLDLoDKTGw+V6w< at >public.gmane.org>
 
 * plugin/icedteanp/IcedTeaPluginUtils.cc:
diff -r 2c84a314c411 -r 9d422db340b8 Makefile.am
--- a/Makefile.amThu May

Re: [rfc][icedtea-web] Removal of 'magic' numbers and more consistentbuffer sizes in C++ side of plugin

Pavel Tisnovsky — 2012-05-25T14:28:06

Hi Adam,

it looks good.

Thank you,
Pavel

Adam Domurad wrote:

[rfc][icedtea-web] Simple for-each loop style patch

Adam Domurad — 2012-05-25T13:56:22

Hey all. I find the for-each syntax that Java 1.5 to be pretty clear in
intent when it comes to data structure iteration. I went through the
code and changed some occurrences of explicit index/iterator loops to
for-each.

OK for HEAD ? (Ran with run-netx-dist-tests with 21/104 failures)

Changelog:
2012-05-25  Adam Domurad <adomurad-H+wXaHxf7aLQT0dZR+AlfA< at >public.gmane.org>

Changed for-loops over iterators and indices to for-each loops if they
were sufficient and clearer.
*netx/net/sourceforge/jnlp/JNLPFile.java: Changed for-loops that could
be expressed more clearly as for-each loops.
*netx/net/sourceforge/jnlp/PluginBridge.java: Same
*netx/net/sourceforge/jnlp/ResourcesDesc.java: Same
*netx/net/sourceforge/jnlp/cache/CacheUtil.java: Same
*netx/net/sourceforge/jnlp/cache/DefaultDownloadIndicator.java: Same
*netx/net/sourceforge/jnlp/cache/Resource.java: Same
*netx/net/sourceforge/jnlp/cache/ResourceTracker.java: Same
*netx/net/sourceforge/jnlp/runtime/AppletEnvironment.java: Same
*netx/net/sour

/hg/icedtea7: Bump to OpenJDK7u6 b10.

2012-05-25T13:05:11

changeset 2c3c929e038c in /hg/icedtea7
details: http://icedtea.classpath.org/hg/icedtea7?cmd=changeset;node=2c3c929e038c
author: Andrew John Hughes <ahughes-H+wXaHxf7aLQT0dZR+AlfA< at >public.gmane.org>
date: Fri May 25 14:04:55 2012 +0100

Bump to OpenJDK7u6 b10.

2012-05-25  Andrew John Hughes  <ahughes-H+wXaHxf7aLQT0dZR+AlfA< at >public.gmane.org>

* Makefile.am:
(OPENJDK_VERSION): Bump to b10.
(JDK_UPDATE_VERSION): Bump to u6.
(CORBA_CHANGESET): Update to IcedTea7 forest head.
(HOTSPOT_CHANGESET): Likewise.
(JAXP_CHANGESET): Likewise.
(JAXWS_CHANGESET): Likewise.
(JDK_CHANGESET): Likewise.
(LANGTOOLS_CHANGESET): Likewise.
(OPENJDK_CHANGESET): Likewise.
(CORBA_SHA256SUM): Likewise.
(HOTSPOT_SHA256SUM): Likewise.
(JAXP_SHA256SUM): Likewise.
(JAXWS_SHA256SUM): Likewise.
(JDK_SHA256SUM): Likewise.
(STRIP_POLICY): Changed from ALT_STRIP_POLICY
due to upstream changes.
* patches/boot/ecj-diamond.patch:
Recreated, new case added in java.io.FileDescriptor.
* patches/boot/ecj-trywi

[rfc] [icedtea-web] get rid of repated sout/serr in reproducerstestcases

Jiri Vanek — 2012-05-25T12:55:28

Hi!

I would like to get rid of repeating std out/err reprint in reproducers by moving this into 
ServerAccess class. This patch is adding simple loging interface for stdout/err and (mainly) is 
hiding reprinting (logging since now:-) of processes out/err and "connectng: " message into 
executeProcess.

My goal in longer term is to have html file (generated from logged xml file probably) which I will 
be able to link (anchor inlcuded) from current test results html files.And so I will be able to 
quickly find whats wrong. Now it is becoming little bit painfull. >(Both stdou/err of processes and 
searching in looong output file)

Saad, Danesh - you have number of reproducers in development. Please count with this change.


Thanx in advance
    J.
diff -r 2c84a314c411 tests/jnlp_tests/signed/AppletTestSigned/testcases/AppletTestSignedTests.java
--- a/tests/jnlp_tests/signed/AppletTestSigned/testcases/AppletTestSignedTests.javaThu May 24 15:37:30 2012 -0400
+++ b/tests/jnlp_tests/signed/AppletTestSigned/testca

/hg/gfx-test: * src/org/gfxtest/testsuites/PrintTestLines.java:

2012-05-25T10:04:21

changeset b56f62052e46 in /hg/gfx-test
details: http://icedtea.classpath.org/hg/gfx-test?cmd=changeset;node=b56f62052e46
author: Pavel Tisnovsky <ptisnovs-H+wXaHxf7aLQT0dZR+AlfA< at >public.gmane.org>
date: Fri May 25 12:07:03 2012 +0200

* src/org/gfxtest/testsuites/PrintTestLines.java:
Added new tests to this test suite.


diffstat:

 ChangeLog                                      |    5 +
 src/org/gfxtest/testsuites/PrintTestLines.java |  239 ++++++++++++++++++++++++-
 2 files changed, 241 insertions(+), 3 deletions(-)

diffs (317 lines):

diff -r ab51d1f4d8ba -r b56f62052e46 ChangeLog
--- a/ChangeLogThu May 24 11:41:36 2012 +0200
+++ b/ChangeLogFri May 25 12:07:03 2012 +0200
< at >< at > -1,3 +1,8 < at >< at >
+2012-05-25  Pavel Tisnovsky  <ptisnovs-H+wXaHxf7aLQT0dZR+AlfA< at >public.gmane.org>
+
+* src/org/gfxtest/testsuites/PrintTestLines.java:
+Added new tests to this test suite.
+
 2012-05-24  Pavel Tisnovsky  <ptisnovs-H+wXaHxf7aLQT0dZR+AlfA< at >public.gmane.org>
 
 * src/org/gfxtest/framework/PrintTest.java:
diff -r ab51d1f4

Re: [rfc] [icedtea-web] providing little bit more debug outputs forfew methods

Jiri Vanek — 2012-05-25T09:32:58

HA!
I got it >:) One condition really escaped. I fixed the block like in patch. It is much more readable 
now and much more bulletproof. Thanx for catch!
With attached patch the jmol is working fine.

2012-05-25  Jiri Vanek  <jvanek-H+wXaHxf7aLQT0dZR+AlfA< at >public.gmane.org>

* netx/net/sourceforge/jnlp/runtime/JNLPClassLoader.java: (getPermissions):
New rethrow of exceptions and following condition make more accurate.
diff -r 2c84a314c411 netx/net/sourceforge/jnlp/runtime/JNLPClassLoader.java
--- a/netx/net/sourceforge/jnlp/runtime/JNLPClassLoader.javaThu May 24 15:37:30 2012 -0400
+++ b/netx/net/sourceforge/jnlp/runtime/JNLPClassLoader.javaFri May 25 11:13:42 2012 +0200
< at >< at > -914,22 +914,26 < at >< at >
                 // 1. Code must be signed
                 // 2. ALL or J2EE permissions must be requested (note: plugin requests ALL automatically)
                 if (cs == null) {
-                    throw new RuntimeException("Code source was null");
+                    throw new NullPointerException("Code so