#26709

Hello ,

we got occasionally a server hotspot crash during dynamic compiling the obfuscated class method

# Java VM: Java HotSpot(TM) Server VM (1.5.0_12-b04 mixed mode)

---

T H R E A D

---

Current thread (0x53e1c180): JavaThread “CompilerThread1” daemon [_thread_in_vm, id=3180]

Stack: [0x54150000,0x54190000)
[error occurred during error reporting, step 110, id 0xc0000005]

Current CompileTask:
opto:3052 org.zefer.html.doc.o.o00000(Ljava/lang/String;Ljava/lang/String;)Ljava/lang/String; (8 bytes)
…

and get none further information (i.e. java stacktrace) in server or system log’s.
(pd4ml.enableDebugInfo() has been set also)

As a result we need a complete web app server restart (tomcat 5)

Without any clue what’s going on it seams not very promising to get a bugfix for it.

So i suggest:
1. We would activate the heap log with param
-XX:+HeapDumpOnOutOfMemoryError
and wait for next error (hopefully with java stack trace)

2. Afterwards – as a workaround – we will set the param
-XX:CompileCommand=exclude,org/zefer/html/doc/o,o00000
to (hopefully) get rid of other server crashes

Do you agree with that ?
Will this lead to further problems or
have your any other proposals for to solve our problem ?

Thanks in advance

#28939

It looks like you’re hitting hotspot compiler bug.

If for some reason an upgrade to the most recent JVM version is not an option for you, as a workaround try to create a .hotspot_compiler file that excludes the call for org.zefer.html.doc.o.o00000

for example:

[language=java:fc72q1tt]#.hotspot_compiler file
exclude org/zefer/html/doc/o o00000[/language:fc72q1tt]
(the library is obfuscated, so the method signatures look strange)

BTW: could you please pubblish “VM Arguments” section contents of the crash report? Didi you overrire the default max heap size value with -Xmx parameter?

#26710

I am using .NET version of PD4ML to render an HTML file to PDF using generatePdfForms(true,””) call. Radio and checkbox come correctly on the form but when form is printed in Acrobat radio and check boxes are missing on printout.
When I inspect created pdf file in Acrobat I see that those check and radio boxes has attribute “Visible but doesn’t print” it should be”Visible” instead. What I am missing? Please help

#28898

I am also seeing this issue on pd4ml version 3.8.0fx3 pro
Seems to have to do with heavy amounts of CSS.

Unfortunately for me, the actual CSS is generally out of my control, so trimming it is not an option.

It is not a server issue. Please benchmark and see for yourselves.

Additionally: It seems to work very fast on a Windows Desktop, but is very noticeably slower on a multi-server Linux Machine. I’m using this through the provided jar, so this shouldn’t be an issue.

#28920

Could I please have a response to this, even if its a request for me to provide more information?

#28940

Here is HTML file I was use for test




Item1Item2Item3Item4

Text Box

Check 1
Check 2
Check 3

R1
R2
R3

#26711

pdf is blank when trying to generate a html file in the same server/domain but when try to upload the html in other domain and use that for the html parameter it works.

any idea with this one?

#26712

For some reason PD4ML suddenly doesn’t function on my system (which happens to be Macintosh).
I have 3.8.0

The only reason I can think of is my recent system update to Lion. Anybody else had a problem with PD4ML on Lion?

The problem is pretty odd. I get no errors in my logs.
After turning on debugging, in one case it tells me that it’s created a .pdf when it’s done nothing at all. Then everything just stops. It took me ages to work out that the problem was with PD4ML because there were absolutely no errors in my logs anywhere.

In another debugging – for a different page – the debug just tells me the version number of PD4ML then does nothing else. Nothing is produced any more in any of the programmes that use PD4ML, at least in the tagged versions. I haven’t tried to check the .java api.

If anybody’s got any ideas, they would be very welcome.

#26713

I am using pd4ml lib which is old, spanish chars are visible on html file, but when i am converting it using pd4ml it is not showing those on pdf file. can u provide trial version of pd4ml which support this.
do u have any other solution for this problem

#28945

Could it be a 64-Bit issue?

#26714

The attached example zip contains a header.html, footer.html, body.html and the generated pdf file.

In the PDF file an additional blank first and blank last page is generated.

Is it my html/CSS code that is causing this or is it something else?

Your helps is highly appreciated.

#28946

I’ve just updated JDK on my Mac workstation to the most actual state and tested PD4ML – everything works fine (as before).

Currently Lion comes with no JDK – you need to install it manually. Also Lion does not support 32bit JDK anymore. So I suspect you installed Lion, but JDK remains from your previous installation, which is not compatible with Lion.

The actual JDK can be downloaded from http://connect.apple.com – please give a try.

#28976

If the spanish characters are not part of ISO-8859-1 you need to use TTF embedding feature of PD4ML Pro.

See: http://pd4ml.com/cookbook/pd4ml_pdf_true_type_fonts.htm

In PD4ML download area you may always find fully functional trial versions of PD4ML Pro: http://pd4ml.com/download.htm

#28942

In order to answer your question we need more input from your side:

1. Do you use PD4ML API or taglib?
2. Does “blank pdf” mean no PDF opened in viewer? Or the document appears as a regular PDF, but page(s) has no content?
3. Code and document samples (HTML/JSP/Java) would help a lot in the issue analyze.

#28921

Sorry for the delay with the reply.

PD4ML supports soft hyphens (­), but I am afraid it is not exactly what you need. We’ll try to add ‍ and ‌ support to the next release.

[Author]

Posts