Saxon

Return-Path: <anna.benton@gmail.com>

Received: from mi008.mc1.hosteurope.de ([80.237.138.247]) by wp245.webpack.hosteurope.de running ExIM with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) id 1Yb9H4-0002MA-S9; Thu, 26 Mar 2015 16:00:46 +0100

Received: from mail-wi0-f171.google.com ([209.85.212.171]) by mx0.webpack.hosteurope.de (mi008.mc1.hosteurope.de) with esmtps (TLSv1.2:AES128-GCM-SHA256:128) id 1Yb9H3-0008Gw-L3 for dropbox+saxonica+f38e@plan.io; Thu, 26 Mar 2015 16:00:46 +0100

Received: by wibg7 with SMTP id g7so151858164wib.1 for <dropbox+saxonica+f38e@plan.io>; Thu, 26 Mar 2015 08:00:45 -0700

Received: by 10.194.43.105 with HTTP; Thu, 26 Mar 2015 08:00:44 -0700

Date: Thu, 26 Mar 2015 08:00:44 -0700

From: Anna Benton <anna.benton@gmail.com>

To: Saxonica Developer Community <dropbox+saxonica+f38e@plan.io>

Message-ID: <CALoy=bjCLxU0tQT3U4HmGv_CTHSLeToa187kp4KBef50w4AbZg@mail.gmail.com>

In-Reply-To: <redmine.journal-4116.20150324180940.3c37b7a5f3fe5965@plan.io>

References: <redmine.issue-2338.20150324172731@plan.io>

<redmine.journal-4116.20150324180940.3c37b7a5f3fe5965@plan.io>

Subject: Re: [Saxon - Bug #2338] Configuration's ErrorListener not thread safe

Mime-Version: 1.0

Content-Type: multipart/alternative;

boundary=001a11c26baab29a8d05123247a3;

charset=UTF-8

Content-Transfer-Encoding: 7bit

Delivery-date: Thu, 26 Mar 2015 16:00:46 +0100

DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113;

h=mime-version:in-reply-to:references:date:message-id:subject:from:to

:content-type; bh=0pf14x+M0PhqfSqoWi18mdQgtTNsa5/V6e23Cp6PJ3w=;

b=gYj5d6CUf4ma9bm0HfCfLBd91ZixQ0KPi0Ly5CFXiBOwiHR5VsmL5nZB2nYlhp0Pgx

Vg4rl8COEmqOpcbkN1vnf6gYOsCyMqgLmvcFsAXGyUrytixT0IkOw14LrxGvU9rlMU3x

wQWiFRRQlzGUPTyPDxk6DAAAre8m37hD2yEitx4nZEcBj1hMjVUm/4R9A1EPI6nM11F3

qGLjiXmFqkQwH9dvyIZ60fD/W3WL3f7T+2AGaifxCfx2lTtHDiJPEA787TTiK2UkBrn3

wBOSuHUr8gitcYT2fnft6/mn89Cg2tQWjEA2KupPeH+VpPOawp6VXCdcGENtJOFNeA+K k17A==

X-Received: by 10.180.187.200 with SMTP id fu8mr48054785wic.2.1427382044103;

Thu, 26 Mar 2015 08:00:44 -0700 (PDT)

X-HE-Spam-Level: +

X-HE-Spam-Score: 1.8

X-HE-Spam-Report: Content analysis details: (1.8 points) pts rule name

description ---- ----------------------

-------------------------------------------------- -0.7 RCVD_IN_DNSWL_LOW RBL:

Sender listed at http://www.dnswl.org/, low trust [209.85.212.171 listed in

list.dnswl.org] 2.5 RCVD_IN_SORBS_HTTP RBL: SORBS: sender is open HTTP proxy

server [209.85.212.171 listed in dnsbl.sorbs.net] 0.0 FREEMAIL_FROM Sender

email is commonly abused enduser mail provider (anna.benton[at]gmail.com) 0.1

HTML_MESSAGE BODY: HTML included in message -0.1 DKIM_VALID_AU Message has a

valid DKIM or DK signature from author's domain -0.1 DKIM_VALID Message has at

least one valid DKIM or DK signature 0.1 DKIM_SIGNED Message has a DKIM or DK

signature, not necessarily valid

X-HE-SPF: PASSED

Envelope-to: dropbox+saxonica+f38e@plan.io

--001a11c26baab29a8d05123247a3

Content-Type: text/plain;

charset=UTF-8

Content-Transfer-Encoding: 7bit

Dear Dr. Kay,

I've sent you a link to a zip file containing our test suite (took us a

while to simplify things) via Google Docs.

Please let me know if you have any questions or if it doesn't work for you!

I forgot to mention in the note on the google doc, we've been running it

with java1.6.

Thanks,

Anna Benton

On Tue, Mar 24, 2015 at 11:09 AM, Saxonica Developer Community <

dropbox+saxonica+f38e@plan.io> wrote:

> --- In your reply, please do not write below this line ---

> Issue #2338 has been updated by Michael Kay.

>

>

> Yes, this area is pretty messy. Given a free choice, we wouldn't have an

> ErrorListener at the Processor/Configuration level, but we need it because

> our Configuration corresponds to JAXP's TransformerFactory, and JAXP puts

> the ErrorListener on the TransformerFactory.

>

> The intention is that setting an ErrorListener on the Configuration should

> implicitly set the ErrorListener on queries and transforms run under that

> Configuration, but not vice versa. You seem to suggest that it's happening

> the other way, and I don't immediately see how that can happen.

>

> Complicating this further is that if the ErrorListener implements Saxon's

> StandardErrorListener interface, then we can clone it using its

> getAnother() method, so that different instances can be used in different

> threads.

>

> I'd be grateful if you could put together a repro that illustrates the

> problem more specifically.

> ------------------------------

> Bug #2338: Configuration's ErrorListener not thread safe

> <https://saxonica.plan.io/issues/2338#change-4116>

>

> - Author: Anna Benton

> - Status: New

> - Priority: Normal

> - Assignee:

> - Category:

> - Sprint/Milestone:

> - Legacy ID:

> - Found in version: 9.5.1.8

> - Fixed in version:

>

> When we get an error in DocumentBuilder.build() it's using the

> ErrorListener of another thread to report it.

>

> Details:

>

> Document builder part:

> We create a SAXSource, passing in an XMLReader on which we have set an

> ErrorHandler.

> There's a problem with the source file (it is empty) and the ErrorHandler

> is used to output an error. At the same time an error goes out to an

> ErrorListener for a completely different thread. I poked around the

> DocumentBuilder source a little and found that DocumentBuilder.build()

> calls buildDocument from the Configuration, and that has a note that says:

> "if any errors occur during document parsing or validation. Detailed

> errors occurring during schema validation will be written to the

> ErrorListener associated with the AugmentedSource, if supplied, or with the

> Configuration otherwise."

>

> The transform part:

> While we are failing to build a document with one thread we are

> successfully transforming some xml in another. When we do this we create a

> new XsltTransformer off of a cached XSLTExecutable and we attach an

> ErrorListener that is specific to that transformation to it. Note that we

> call load() on this single XsltExecutable across many threads at once to

> get an XsltTransformer for each xml file in a batch, but we do not re-use

> the XsltTransformers (although to isolate this problem I limited our inputs

> to a single xml file being transformed and a single xml file going through

> DocumentBuilder).

>

> When we actually do our transform() the XsltTransformer's ErrorListener

> becomes the ErrorListener of the Configuration. I've tracked this by

> outputting the ERROR_LISTENER_CLASS configuration property from the

> Processor right before and right after the transform() call.

>

> A clearer example:

>

> Thread 1:

> Configuration ErrorListener Originally:

> net.sf.saxon.lib.StandardErrorListener

> Source #1 runs a transform() (with one of our ErrorListeners set on the

> XsltTransformer)

> Configuration ErrorListener is now set to one of our ErrorListeners (it is

> using our LogWriter class)

>

> Thread 2:

> Source #2 is an empty file. When we go to build a document off of it, an

> error is emitted to two places:

>

> 1. The ErrorHandler we set up on the XMLReader which we pass in when we

> build the SAXSource (this is great, it's what we want).

> 2. The Configuration's ErrorListener, which, thanks to thread #1, is now

> Source #1's LogWriter.

>

> We depend on our ErrorListeners to write out our logs for our

> transformations, which we then parse and use.

>

> Are we not using ErrorListeners as intended here? Is there a way to keep

> the ErrorListener we set on the XsltTransformer from being set on the

> Configuration? That would be our first choice since we're not sure what

> other circumstances might trigger its use at the Configuration level. I

> haven't run through the Saxon code itself very far, so I'm not sure of the

> exact point at which transforming sets the Configuration's ErrorListener.

> Note that we're using the EnterpriseConfiguration with saxonEE 9.5.1.8.

> ------------------------------

>

> You have received this notification because you have either subscribed to

> or are involved in a project on Saxonica Developer Community site.

> To change your notification preferences, please click here:

> https://saxonica.plan.io/my/account?tour=mail_preferences

>

> This notification was cheerfully delivered by <https://plan.io/>

> [image: Planio] <https://plan.io/>

>

--001a11c26baab29a8d05123247a3

Content-Type: text/html;

charset=UTF-8

Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Dear Dr. Kay,<div><br></div><div>I&#39;ve sent you a lin=

k to a zip file containing our test suite (took us a while to simplify th=

ings) via Google Docs.</div><div><br></div><div>Please let me know if you=

have any questions or if it doesn&#39;t work for you!</div><div><br></di=

v><div>I forgot to mention in the note on the google doc, we&#39;ve been =

running it with java1.6.</div><div><br></div><div>Thanks,</div><div>Anna =

Benton</div></div><div class=3D"gmail_extra"><br><div class=3D"gmail_quot=

e">On Tue, Mar 24, 2015 at 11:09 AM, Saxonica Developer Community <span d=

ir=3D"ltr">&lt;<a href=3D"mailto:dropbox+saxonica+f38e@plan.io" target=3D=

"_blank">dropbox+saxonica+f38e@plan.io</a>&gt;</span> wrote:<br><blockquo=

te class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc =

solid;padding-left:1ex">

<div>

<table width=3D"100%">

<tbody><tr><td style=3D"font-family:MarketWeb,Verdana,sans-serif;font-siz=

e:0.8em;text-align:center;width:100%;color:#d7d7d7"><p>--- In your reply,=

please do not write below this line ---</p></td></tr>

<tr><td>Issue #2338 has been updated by Michael Kay.

<ul>

</ul>

<p>Yes, this area is pretty messy. Given a free choice, we wouldn&#39;t h=

ave an ErrorListener at the Processor/Configuration level, but we need it=

because our Configuration corresponds to JAXP&#39;s TransformerFactory, =

and JAXP puts the ErrorListener on the TransformerFactory.</p>

<p>The intention is that setting an ErrorListener on the Configuration s=

hould implicitly set the ErrorListener on queries and transforms run unde=

r that Configuration, but not vice versa. You seem to suggest that it&#39=

;s happening the other way, and I don&#39;t immediately see how that can =

happen.</p>

<p>Complicating this further is that if the ErrorListener implements Sax=

on&#39;s StandardErrorListener interface, then we can clone it using its =

getAnother() method, so that different instances can be used in different=

threads.</p>

<p>I&#39;d be grateful if you could put together a repro that illustrate=

s the problem more specifically.</p>

<hr>

<h1><a href=3D"https://saxonica.plan.io/issues/2338#change-4116" target=3D=

"_blank">Bug #2338: Configuration&#39;s ErrorListener not thread safe</a>=

</h1>

<ul><li>Author: Anna Benton</li>

<li>Status: New</li>

<li>Priority: Normal</li>

<li>Assignee: </li>

<li>Category: </li>

<li>Sprint/Milestone: </li>

<li>Legacy ID: </li>

<li>Found in version: 9.5.1.8</li>

<li>Fixed in version: </li></ul>

<p>When we get an error in DocumentBuilder.build() it&#39;s using the Err=

orListener of another thread to report it.</p>

<p>Details:</p>

<p>Document builder part:<br>We create a SAXSource, passing in an XMLRea=

der on which we have set an ErrorHandler.<br>There&#39;s a problem with t=

he source file (it is empty) and the ErrorHandler is used to output an er=

ror. At the same time an error goes out to an ErrorListener for a complet=

ely different thread. I poked around the DocumentBuilder source a little =

and found that DocumentBuilder.build() calls buildDocument from the Confi=

guration, and that has a note that says:<br>&quot;if any errors occur dur=

ing document parsing or validation. Detailed errors occurring during sche=

ma validation will be written to the ErrorListener associated with the Au=

gmentedSource, if supplied, or with the Configuration otherwise.&quot;</p=

>

<p>The transform part:<br>While we are failing to build a document with =

one thread we are successfully transforming some xml in another. When we =

do this we create a new XsltTransformer off of a cached XSLTExecutable an=

d we attach an ErrorListener that is specific to that transformation to i=

t. Note that we call load() on this single XsltExecutable across many thr=

eads at once to get an XsltTransformer for each xml file in a batch, but =

we do not re-use the XsltTransformers (although to isolate this problem I=

limited our inputs to a single xml file being transformed and a single x=

ml file going through DocumentBuilder).</p>

<p>When we actually do our transform() the XsltTransformer&#39;s ErrorLi=

stener becomes the ErrorListener of the Configuration. I&#39;ve tracked t=

his by outputting the ERROR_LISTENER_CLASS configuration property from th=

e Processor right before and right after the transform() call.</p>

<p>A clearer example:</p>

<p>Thread 1:<br>Configuration ErrorListener Originally: net.sf.saxon.lib=

.StandardErrorListener<br>Source #1 runs a transform() (with one of our E=

rrorListeners set on the XsltTransformer)<br>Configuration ErrorListener =

is now set to one of our ErrorListeners (it is using our LogWriter class)=

</p>

<p>Thread 2:<br>Source #2 is an empty file. When we go to build a docume=

nt off of it, an error is emitted to two places:</p>

<p>1. The ErrorHandler we set up on the XMLReader which we pass in when =

we build the SAXSource (this is great, it&#39;s what we want).<br>2. The =

Configuration&#39;s ErrorListener, which, thanks to thread #1, is now Sou=

rce #1&#39;s LogWriter.</p>

<p>We depend on our ErrorListeners to write out our logs for our transfo=

rmations, which we then parse and use.</p>

<p>Are we not using ErrorListeners as intended here? Is there a way to k=

eep the ErrorListener we set on the XsltTransformer from being set on the=

Configuration? That would be our first choice since we&#39;re not sure w=

hat other circumstances might trigger its use at the Configuration level.=

I haven&#39;t run through the Saxon code itself very far, so I&#39;m not=

sure of the exact point at which transforming sets the Configuration&#39=

;s ErrorListener. Note that we&#39;re using the EnterpriseConfiguration w=

ith saxonEE 9.5.1.8.</p>

</td></tr>

<tr><td style=3D"font-size:0.8em;width:100%"><hr><p>You have received thi=

s notification because you have either subscribed to or are involved in a=

project on Saxonica Developer Community site.<br>To change your notifica=

tion preferences, please click here: <a href=3D"https://saxonica.plan.io/=

my/account?tour=3Dmail_preferences" target=3D"_blank">https://saxonica.pl=

an.io/my/account?tour=3Dmail_preferences</a></p></td><td></td></tr>

<tr><td style=3D"font-family:MarketWeb,Verdana,sans-serif;font-size:1.2em=

;text-align:center;width:100%;color:#d7d7d7"><br><div><a href=3D"https://=

plan.io/" style=3D"color:#d7d7d7;text-decoration:none" target=3D"_blank">=

This notification was cheerfully delivered by</a></div></td><td></td></tr=

>

<tr><td style=3D"text-align:center;width:100%"><a href=3D"https://plan.io=

/" title=3D"Planio" target=3D"_blank"><img src=3D"https://assets.plan.io/=

images/planio_logo_gray_204x50.png" height=3D"25" width=3D"102" alt=3D"Pl=

anio" style=3D"vertical-align:middle"></a></td></tr>

</tbody></table>

</div>

</blockquote></div><br></div>

--001a11c26baab29a8d05123247a3--