I working with Orbeon XForms 4.7 on eXist-db 2.2. We've always had session problems where a session would expire in eXist-db after a given point in time, but where Orbeon retains a session token and 'thinks' it is still logged on. This may or may not be at the heart of what I see in the logging:
2016-11-07 15:06:27,643 INFO ProcessorService - /xforms-server - Received request
2016-11-07 15:06:27,650 INFO ProcessorService - /xforms-server - Timing: 7
2016-11-07 15:06:29,368 INFO ProcessorService - /xforms-server - Received request
2016-11-07 15:06:29,372 ERROR XFormsServer - Unable to retrieve XForms engine state. Please reload the current page. Note that you will lose any unsaved changes. {}
2016-11-07 15:06:29,372 INFO ProcessorService - /xforms-server - Timing: 4
2016-11-07 15:06:29,373 ERROR ProcessorService -
+----------------------------------------------------------------------------------------------------------------------+
|An Error has Occurred |
|----------------------------------------------------------------------------------------------------------------------|
|Unable to retrieve XForms engine state. Please reload the current page. Note that you will lose any unsaved changes. |
|----------------------------------------------------------------------------------------------------------------------|
|Application Call Stack |
|----------------------------------------------------------------------------------------------------------------------|
|oxf:/config/prologue-servlet.xpl |executing processor | 36|
|······················································································································|
|element=<p:processor name="oxf:pipeline">[...]</p:processor> |
|name ={http://www.orbeon.com/oxf/processors}pipeline |
|----------------------------------------------------------------------------------------------------------------------|
|oxf:/ops/xforms/xforms-server.xpl |reading processor output | 43|
|······················································································································|
|element=<p:output name="response" id="xforms-response"/> |
|name =response |
|id =xforms-response |
|----------------------------------------------------------------------------------------------------------------------|
|----------------------------------------------------------------------------------------------------------------------|
|Exception: org.orbeon.oxf.common.OXFException |
|----------------------------------------------------------------------------------------------------------------------|
|org.orbeon.oxf.xforms.state.XFormsStateManager |createDocumentFromStore |XFormsStateManager.java | 489|
|org.orbeon.oxf.xforms.state.XFormsStateManager |findOrRestoreDocument |XFormsStateManager.java | 448|
|org.orbeon.oxf.xforms.state.XFormsStateManager |beforeUpdate |XFormsStateManager.java | 328|
|org.orbeon.oxf.xforms.processor.XFormsServer |doIt |XFormsServer.java | 169|
|org.orbeon.oxf.xforms.processor.XFormsServer |access$000 |XFormsServer.java | 59|
|org.orbeon.oxf.xforms.processor.XFormsServer$1 |readImpl |XFormsServer.java | 85|
|essor.impl.ProcessorOutputImpl$TopLevelOutputFilter|read |ProcessorOutputImpl.java | 257|
|org.orbeon.oxf.processor.impl.ProcessorOutputImpl |read |ProcessorOutputImpl.java | 394|
|----------------------------------------------------------------------------------------------------------------------|
|Exception: org.orbeon.oxf.common.ValidationException |
|----------------------------------------------------------------------------------------------------------------------|
|org.orbeon.oxf.common.OrbeonLocationException$ |wrapException |OrbeonLocationException.scala | 60|
|org.orbeon.oxf.common.OrbeonLocationException |wrapException |OrbeonLocationException.scala | |
This error comes up every such and so seconds. There is no mention what XForms triggers it, so a targeted search is out of the question.
Does anyone know how to interpret this error and possibly how to prevent it? Note that there's no error at the user side at all, and everything appears to work fine. This is purely server side.
I'm running into some odd behavior when trying to setup an ActionFixture test using Fitnesse (with FitSharp as the test runner)
When creating an actionFixture I'll get an error that the class (Namespace.TestClassName in example below) cannot be found. If I create a wiki page for it the test will work.
| actionFixture |
| start | Namespace.TestClassName |
Is it required to have a page for each class? If so can I reference the same page for all tests (different location in hierarchy)?
Sorry for the naive question, sure I'm missing something simple here.
Simple error on my part;
Needed to enclose the class name as shown below
| actionFixture |
| start | !-Namespace.TestClassName-! |
...or simply don't write the name as a wikiword;
| actionFixture |
| start | namespace.testClassName |
I'm using CircleCI to check for security issues and this is cropping up as an error, though I'm not sure that it is.
This is the line of code that is causing one of the scripting errors:
= link_to t(:delete), main_app.board_comment_path(#board, comment), method: :delete
Is this a valid security issue? Is there any way for me to make Brakeman accept these parameters as safe? I read up on --url-safe-methods but I couldn't figure out a way to make it work.
Used this link as a guide https://github.com/presidentbeef/brakeman/pull/45
Running bundle exec brakeman -A -q --exit-on-warn, this is the error report:
+BRAKEMAN REPORT+
Application path: ****
Rails version: 4.2.2
Brakeman version: 3.0.4
Started at 2015-06-26 14:10:14 -0700
Duration: 1.8311 seconds
Checks run: BasicAuth, ContentTag, CreateWith, CrossSiteScripting, DefaultRoutes, Deserialize, DetailedExceptions, DigestDoS, EscapeFunction, Evaluation, Execute, FileAccess, FileDisclosure, FilterSkipping, ForgerySetting, HeaderDoS, I18nXSS, JRubyXML, JSONEncoding, JSONParsing, LinkTo, LinkToHref, MailTo, MassAssignment, ModelAttrAccessible, ModelAttributes, ModelSerialize, NestedAttributes, NumberToCurrency, QuoteTableName, Redirect, RegexDoS, Render, RenderDoS, RenderInline, ResponseSplitting, SQL, SQLCVEs, SSLVerify, SafeBufferManipulation, SanitizeMethods, SelectTag, SelectVulnerability, Send, SendFile, SessionSettings, SimpleFormat, SingleQuotes, SkipBeforeFilter, StripTags, SymbolDoS, SymbolDoSCVE, TranslateBug, UnsafeReflection, UnscopedFind, ValidationRegex, WithoutProtection, XMLDoS, YAMLParsing
+SUMMARY+
+-------------------+-------+
| Scanned/Reported | Total |
+-------------------+-------+
| Controllers | 23 |
| Models | 9 |
| Templates | 53 |
| Errors | 0 |
| Security Warnings | 2 (0) |
+-------------------+-------+
+----------------------+-------+
| Warning Type | Total |
+----------------------+-------+
| Cross Site Scripting | 2 |
+----------------------+-------+
View Warnings:
+------------+------------------------------------------------------------------+----------------------+-------------------->>
| Confidence | Template | Warning Type | Message >>
+------------+------------------------------------------------------------------+----------------------+-------------------->>
| Medium | boards/show (BoardsController#show) | Cross Site Scripting | Unsafe parameter va>>
| Medium | boards/show (BoardsController#show) | Cross Site Scripting | Unsafe parameter va>>
+------------+------------------------------------------------------------------+----------------------+-------------------->>
This is (almost certainly) a false positive, assuming board_comment_path returns a path.
The reason Brakeman warns about URLs in link_to is because it is possible to set URLs like javascript:dangerous_stuff_here(). A common example would be user profiles linking to a user's website.
--url-safe-methods only applies to methods wrapping input to link_to. For example, link_to 'stuff', safe_url(some_input).
However, after https://github.com/presidentbeef/brakeman/pull/674 Brakeman will stop warning about path helpers in URLs and also expand --safe-methods/--url-safe-methods to match all types of methods.
I have an application that requests data from a database triggered by a timer on a form. If there is an error (the connection to the database is lost), I sometimes I get the expected exception (EIBO_ISCError) and sometimes I get an access violation in RtlLeaveCriticalSection of ntdll.dll. Here is the corresponing Eurekalog stack:
------------------------------------------------------------------------------------------------------
|Adresse |Modul |Unit |Klasse |Prozedur/Methode |Zeile |
------------------------------------------------------------------------------------------------------
|Laufender Thread: ID=1320; Priorität=0; Klasse=; [Haupt Thread] |
|----------------------------------------------------------------------------------------------------|
|76FD2280|ntdll.dll | | |RtlLeaveCriticalSection | |
|76FDE0ED|ntdll.dll | | |RtlAllocateHeap | |
|76FE6CC5|ntdll.dll | | |LdrUnlockLoaderLock | |
|7552EF19|KERNELBASE.dll| | |VirtualQueryEx | |
|7552EF02|KERNELBASE.dll| | |VirtualQueryEx | |
|7552EFE6|KERNELBASE.dll| | |VirtualQuery | |
|76FC012E|ntdll.dll | | |KiUserExceptionDispatcher | |
|0069D997|Program.exe |IBODataset.pas |TIBOInternalDataset|DoHandleError |8407[23] |
|0063B3F7|Program.exe |IB_Components.pas |TIB_Session |DoHandleError |13181[2] |
|0068B36C|Program.exe |IB_Session.pas |TIB_SessionBase |HandleException |1442[58] |
|0068B03C|Program.exe |IB_Session.pas |TIB_SessionBase |HandleException |1384[0] |
|0064EE74|Program.exe |IB_Components.pas |TIB_Statement |API_Execute |22927[14]|
|0064EE10|Program.exe |IB_Components.pas |TIB_Statement |API_Execute |22913[0] |
|00655D1D|Program.exe |IB_Components.pas |TIB_Dataset |SysExecSelect |26432[1] |
|0064DA60|Program.exe |IB_Components.pas |TIB_Statement |SysExecStatement |22259[9] |
|0064D7A1|Program.exe |IB_Components.pas |TIB_Statement |SysExecute |22173[12]|
|0064D708|Program.exe |IB_Components.pas |TIB_Statement |SysExecute |22161[0] |
|00655A9F|Program.exe |IB_Components.pas |TIB_Dataset |SysExecute |26373[7] |
|00655210|Program.exe |IB_Components.pas |TIB_Dataset |SysOpen |26160[23]|
|006550F8|Program.exe |IB_Components.pas |TIB_Dataset |SysOpen |26137[0] |
|006994E5|Program.exe |IBODataset.pas |TIBODataset |DoBeforeOpen |6312[17] |
|0061FBEA|Program.exe |mvdb.pas |TImvDatabase |QueryRun |1393[10] |
...
|00B1D440|Program.exe |StartDialogForm.pas|TFormStartDialog |UpdateStartBar |494[0] |
|00B1D4C3|Program.exe |StartDialogForm.pas|TFormStartDialog |TimerExBarTimer |521[6] |
|76667BC5|USER32.dll | | |DispatchMessageA | |
|76667BBB|USER32.dll | | |DispatchMessageA | |
|00BF1178|Program.exe |Program.dpr | | |884[399] |
------------------------------------------------------------------------------------------------------
The code, which is executed, is nothing special. It boils down to:
qry := TIBOQuery.Create(nil); //IBObjects
qry.SQL := 'SELECT COUNT(IDX) FROM TABLE';
qry.Prepare;
when creating the form and
qry.Open; //<-- Exception
TotalCount := qry.Fields[0].AsVariant;
qry.Close;
in the OnTimer event of the MDI form.
The code line in IBObjects that is called in DoHandleError is
raise EIBO_ISCError.CreateISC( ... );
The underlaying exception is likely to be caused by a lost database connection in qry.Open. What I want to know is, which circumstances (read defects in my code) can lead to the behaviour, that sometimes this exception is handled as expected (EIBO_ISCError in Eurekalog) and sometimes the same exception leads to an access violation in RtlLeaveCriticalSection.
It looks like you have heap corruption. Somewhere, your program has written to memory that you're not supposed to write to.
That might mean you've written to a critical-section data structure belonging to the heap, but it might mean you've written somewhere else that caused the memory manager to think there's a critical-section object where there really isn't one.
The stack trace suggests you're still getting the usual exception you expected to get, but in attempting to handle that exception, something goes wrong.
You could try using the debugger to inspect other memory near where the invalid read occurs. See whether there are any strings or numbers you recognize from your program. They could indicate which section of code is writing where it shouldn't.
From the comments, for reference:
The cause of this access violation was, that I freed an exception that occured while handling an other exception (like originally suggested here).
Citation from the comments:
you should never free an exception, ever
The exception was handled normally if during the handling of the original exception no other exception occured. If an other exception occured I got the access violation.
I have a grails application which is failing at runtime in a cryptic way
(cyptic to me anyway)
with a ArrayIndexOutOfBoundsException when I visit the scaffolded /imca2/imcaReferral/index.
* now edited to put solution at end *
There are about a dozen domain classes.
I haven't got round to worrying about the UI yet so the controllers are all dynamically scaffolded.
All the other controllers work OK.
This Controller:
package com.ubergen
class ImcaReferralController {
def scaffold = ImcaReferral
}
For this Domain:
package com.ubergen
class ImcaReferral {
private def todayDate = new Date()
String advocacyReferenceNum = ""
[snip a lot of code]
String toString() {
"${this.advocacyReferenceNum}: ${this.client?this.client:'-'}${this.referralIssue?', '+this.referralIssue:''}"
}
}
(I don't want to post the domain class here as its huge).
Produces this stacktrace:
|Server running. Browse to http://localhost:8080/imca2
| Error 2014-03-12 18:48:24,935 [http-bio-8080-exec-3] ERROR errors.GrailsExceptionResolver - ArrayIndexOutOfBoundsException occurred when processing request: [GET] /imca2/imcaReferral/index
0. Stacktrace follows:
Message: 0
Line | Method
->> 55 | <init> in grails.orm.PagedResultList
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
| 15 | $tt__index in com.ubergen.ImcaReferralController
| 191 | doFilter . in grails.plugin.cache.web.filter.PageFragmentCachingFilter
| 63 | doFilter in grails.plugin.cache.web.filter.AbstractFilter
| 1146 | runWorker in java.util.concurrent.ThreadPoolExecutor
| 615 | run in java.util.concurrent.ThreadPoolExecutor$Worker
^ 701 | run . . . in java.lang.Thread
Cleaning and (re)compiling make no difference.
The domain class is being used during bootstrapping to push data successfully into the database, so it works to that extent.
I can run the application from the command line instead of from inside eclipse/STS. The same error is thrown.
run-app --noreloading makes no difference either (clutching at straws now). And run-war also produces the same error.
run-app --verbose shows:
| Error 2014-03-12 19:58:37,745 [http-bio-8080-exec-1] ERROR errors.GrailsExceptionResolver - ArrayIndexOutOfBoundsException occurred when processing request: [GET] /imca2/imcaReferral/index
0. Stacktrace follows:
java.lang.ArrayIndexOutOfBoundsException: 0
at org.hibernate.criterion.Order.toSqlString(Order.java:73)
at org.hibernate.loader.criteria.CriteriaQueryTranslator.getOrderBy(CriteriaQueryTranslator.java:394)
[snip]
at grails.orm.PagedResultList.<init>(PagedResultList.java:55)
[snip]
at com.ubergen.ImcaReferral.list(ImcaReferral.groovy)
[snip]
at com.ubergen.ImcaReferralController.$tt__index(script1394654146228610896735.groovy:15)
[snip]
So the index page calls the domain's list() and this is a problem in some way but not enough of a way that its getting mentioned in the stacktrace.
Where should I look first for the problem?
Versions:
ubuntu 10.04
eclipse / SpringToolSuite 3.4.0
grails 2.3.6
groovy 2.1.9 (for both project and workspace)
Update 13/03/2014
I followed Joe's suggestions (below) and found that the problem is indeed in the ImcaReferral.list() method.
In the grails console simply running:
package com.ubergen
ImcaReferral.withTransaction { status ->
ImcaReferral.list()
}
Returns
java.lang.ArrayIndexOutOfBoundsException: 0
at org.hibernate.criterion.Order.toSqlString(Order.java:73)
at org.hibernate.loader.criteria.CriteriaQueryTranslator.getOrderBy(CriteriaQueryTranslator.ja a:394)
[snip]
at com.ubergen.ImcaReferral.list(ImcaReferral.groovy)
Looking at the domain's sort order information BINGO! its incorrectly defined, there are two competing definitions of how to sort the domain.
I Comment out the erroneous sort order information:
package com.ubergen
class ImcaReferral {
...
static hasMany = [challenges:Challenge]
static mapping = {
...
sort dateReceived:'asc'
// sort challenges:'challengeRoute' // *** ERROR ***
}
}
and (after restarting the console) the call to list works fine and returns an empty array.
Correcting the sort order of the child records:
package com.ubergen
class ImcaReferral {
...
static hasMany = [challenges:Challenge]
static mapping = {
...
sort dateReceived:'asc'
challenges sort: 'challengeRoute', order: 'asc' // *** CORRECT ***
}
}
Fixes the problem. The scaffolding now works.
Conclusions
Trust the full stacktrace even if its rather verbose. It shows the classes and methods to look at.
Learn to use the console.
grails -reloading console
Read your more code carefully!
You could try to generate the Static Scaffolding and see if you get a different result. You could also try running the list in a integration test to see what happens.