You are viewing a plain text version of this content. The canonical link for it is here.
Posted to kato-spec@incubator.apache.org by Steve Poole <sp...@googlemail.com> on 2010/01/11 15:21:58 UTC
System Properties and Env vars
Is it just me or does anyone else think its an unneccessary restriction that
the JavaRuntime API does not offer the ability to get to System Properties
or Environmental variables? You can get them from a running JVM so why
should we not have the same capability on our API?
--
Steve
Re: System Properties and Env vars
Posted by Daniel Julin <dp...@us.ibm.com>.
David Griffiths <da...@gmail.com> wrote on 2010-01-12 06:01:00
AM:
> BTW Dan - is it really possible to have multiple copies of
java.lang.System?
Hi David:
I have not personally witnessed multiple copes of java.lang.System, but I
don't see any technical reason why this could not happen in a complex J2EE
system with a big hierarchy of classloaders.
Certainly, in WebSphere Application Server, we have a multitude of "custom
properties" classes in many components at many levels of the classloader
tree. They're not quite as central as java.lang.System, but they are
important to keep track of nonetheless, and they motivate the need for
special analyzer modules that understand about that particular structure in
a particular runtime like WebSphere Application Server...
==============================================================
Daniel P. Julin
Senior Technical Staff Member
Technical Area Lead - IBM AIM WebSphere Serviceability Team
EMail: dpj@us.ibm.com
Phone: 347-273-2093 Tie-line 930-3468
Notes: Daniel Julin/Pittsburgh/IBM@IBMUS
Re: System Properties and Env vars
Posted by Steve Poole <sp...@googlemail.com>.
Dan - thanks for sharing your list. Before we go further I'd like to poll
the mailing list for a reaction. Who thinks that some or all of the
functions list should be part of the API and if so which ones?
On Thu, Jan 14, 2010 at 3:17 AM, Daniel Julin <dp...@us.ibm.com> wrote:
>
>
> Stuart Monteith <st...@stoo.me.uk> wrote on 2010-01-12 06:51:50 AM:
> > Hi David,
> > This brings us back to the whole "Collections" problem where classes
> > can be implemented differently by different JREs. Which is, of course, a
> > more general problem. We possibly need to build on the basic API a means
> > of abstracting access to SDK classes and provide means of extending this.
>
> Hah yes... Not to sound like a broken record :-) but we have that too in
> the "utility" library from DTFJ-DumpAnalyzer. We have a concept of a
> "wrapper" analyzer module, that encapsulates all the information that is
> necessary to interpret the fields that implement a particular collection
> class (or other common classes) in the underlying dump, and provides
> programmatic access to the logical information contained in the underlying
> object (e.g. list all the entries in a collection).
>
> Currently, we have wrappers for:
>
> ArrayList
> ConcurrentHashMap
> ConcurrentLinkedQueue
> Date
> HashMap
> HashSet
> Hashtable
> LinkedList
> Properties
> String
> SynchronizedMap
> SynchronizedSet
> ThreadLocal
> TreeMap
> TreeSet
> Vector
>
> This list has grown organically over time, simply based on the classes that
> we came across while writing our more complex analyzers. It's obviously
> very spotty.
>
> Several of these wrappers have multiple alternative implementations, that
> correspond to different implementation of the underlying class that they
> "wrap" (e.g. one implementation for the 1.4.2 class library, another for
> 5.0, another for 6.0...)
>
>
> ==============================================================
> Daniel P. Julin
> Senior Technical Staff Member
> Technical Area Lead - IBM AIM WebSphere Serviceability Team
>
> EMail: dpj@us.ibm.com
> Phone: 347-273-2093 Tie-line 930-3468
> Notes: Daniel Julin/Pittsburgh/IBM@IBMUS
>
--
Steve
Re: System Properties and Env vars
Posted by Daniel Julin <dp...@us.ibm.com>.
Stuart Monteith <st...@stoo.me.uk> wrote on 2010-01-12 06:51:50 AM:
> Hi David,
> This brings us back to the whole "Collections" problem where classes
> can be implemented differently by different JREs. Which is, of course, a
> more general problem. We possibly need to build on the basic API a means
> of abstracting access to SDK classes and provide means of extending this.
Hah yes... Not to sound like a broken record :-) but we have that too in
the "utility" library from DTFJ-DumpAnalyzer. We have a concept of a
"wrapper" analyzer module, that encapsulates all the information that is
necessary to interpret the fields that implement a particular collection
class (or other common classes) in the underlying dump, and provides
programmatic access to the logical information contained in the underlying
object (e.g. list all the entries in a collection).
Currently, we have wrappers for:
ArrayList
ConcurrentHashMap
ConcurrentLinkedQueue
Date
HashMap
HashSet
Hashtable
LinkedList
Properties
String
SynchronizedMap
SynchronizedSet
ThreadLocal
TreeMap
TreeSet
Vector
This list has grown organically over time, simply based on the classes that
we came across while writing our more complex analyzers. It's obviously
very spotty.
Several of these wrappers have multiple alternative implementations, that
correspond to different implementation of the underlying class that they
"wrap" (e.g. one implementation for the 1.4.2 class library, another for
5.0, another for 6.0...)
==============================================================
Daniel P. Julin
Senior Technical Staff Member
Technical Area Lead - IBM AIM WebSphere Serviceability Team
EMail: dpj@us.ibm.com
Phone: 347-273-2093 Tie-line 930-3468
Notes: Daniel Julin/Pittsburgh/IBM@IBMUS
Re: System Properties and Env vars
Posted by Stuart Monteith <st...@stoo.me.uk>.
Hi David,
This brings us back to the whole "Collections" problem where classes
can be implemented differently by different JREs. Which is, of course, a
more general problem. We possibly need to build on the basic API a means
of abstracting access to SDK classes and provide means of extending this.
Regards,
Stuart
David Griffiths wrote:
> One thing to bear in mind is that the implementation of Hashtable (the
> superclass of Properties) can vary between JVMs. For instance our Java
> 1.4 and Java 5 versions use private fields with different names. Of
> course you can still have code above Kato that says "if table field
> exists do this, else if elementData field exists do this etc". But
> given how useful access to the system properties is it may be easier
> to just add it to Kato?
>
--
Stuart Monteith
http://blog.stoo.me.uk/
Re: System Properties and Env vars
Posted by David Griffiths <da...@gmail.com>.
One thing to bear in mind is that the implementation of Hashtable (the
superclass of Properties) can vary between JVMs. For instance our Java
1.4 and Java 5 versions use private fields with different names. Of
course you can still have code above Kato that says "if table field
exists do this, else if elementData field exists do this etc". But
given how useful access to the system properties is it may be easier
to just add it to Kato?
BTW Dan - is it really possible to have multiple copies of java.lang.System?
Cheers,
Dave
Re: System Properties and Env vars
Posted by Daniel Julin <dp...@us.ibm.com>.
Steve Poole <sp...@googlemail.com> wrote on 2010-01-12 03:33:21 AM:
> Yes that's right. I'm sure we'd all be interested in knowing more
about
> these utilitiy modules you mention - could you provide a basic 1 liner
> description for the most important ones?
Here is a list of the most common utility modules/functions from the
DTFJ-DumpAnalyzer library, that I had prepared quite a while ago... The
name of the module is in square brackets.
(1) Given a JavaObject, extract the value of a field specified by its name
(without having to enumerate through all the fields to find the right one).
Also deal with field inherited from superclasses, etc. - [OneJavaObject]
(2) Given a JavaClass, extract the value of a static field specified by its
name, taking into account superclasses, etc. (same as (1) above, but for
static fields) - [OneJavaClass]
(3) Given a JavaObject or JavaClass as a starting point, extract the value
at the end of a sequence of fields, represented as path expression (e.g.
X/Y/Z: field X is a pointer to object that contains field Y, which itself
is a pointer to an object that contains field Z) - [ObjectWrapper]
(4) Given a JavaObject and a field that happens to be a reference to an
array or an instance of the Java Collections framework, extract the
contents of that collection, rather than just a reference to the collection
or the fields that happen to be part of the internal implementation of that
collection class - [ObjectWrapper + specialized wrappers]
(5) Extend function (4) to deal not just with arrays and the standard Java
Collection classes, but with arbitrary application classes, by providing a
specialized "wrapper" analyzer that defines how to interpret the contents
of that particular application class - [ObjectWrapper + specialized
wrappers]
(6) Print the content of a field or collection as extracted by one of the
functions (1), (2), (3), (4), (5) above, in a way such that all possible
errors are handled internally and simply printed instead of the desired
value, rather than forcing the programmer to explicitly catch and handle
all exceptions in his/her code - [ObjectWrapper]
(7) Extend function (6) above to allow the caller to specify a special
"wrapper" analyzer that prints the desired contents in a particular format
instead of the default format. For example, instead of printing all the
entries in a collection, the "wrapper" could filter the entries and only
print the relevant ones. Several variations to deal with Sets, Maps,
individual objects, etc. [ObjectWrapper + specialized wrappers]
(8) List all definitions of a given class, specified by name, accounting
for the fact that there could be multiple definitions of the same class
present in different classloaders in the dump. [ClassFinder]
(9) List all definitions of a given class, specified by name, and all its
subclasses, present in the dump. [ClassFinder]
(10) Extend function (9) above to accept the name of an interface instead
of a class, and find all the sub-interfaces and all classes that implement
that interface or any of its sub-interfaces. [ClassFinder]
(11) Find the JavaClass that represents a given interface, specified by its
name. [ClassFinder]
(12) Get a count of all the class definitions present in the dump
[ClassFinder]
(13) List all instances of a given class (accounting for the fact that the
class may have multiple definitions in different classloaders, as noted in
function (8) above) [ObjectFinder]
(14) List all instances of a given class or any of it subclasses, or all
instances that implement a given interface or any of its sub-interfaces.
[ObjectFinder]
(15) Get a count of all the object instances of a given class, or all the
object instances of any class, or all the instances of primitive arrays, or
all the instances of object arrays [ObjectFinder]
(16) Given an address (ImagePointer), find the JavaClass or JavaObject or
JavaThread or other entity whose ID matches that address
[LookupObjectByAddress]
(17) List all the object references outbound from a given JavaObject
[OneJavaObject]
(18) Get a canonical type signature for a given JavaObject or JavaClass
(e.g. "[Lmypackage.myclass;") [OneJavaObject, OneJavaClass]
(19) Enumerate all the fields from a given JavaClass, including all field
inherited from superclasses and not masked [OneJavaClass]
(20) Variations on function (19) above, that filters to enumerate only
static fields, or only fields that match particular criteria [OneJavaClass]
(21) Print the content of a field in a given JavaObject, but mask that
content if the field happens to be a password (as specified by a regular
expression for the name of the field) [ObjectWrapper]
(22) Get a count of all the JavaThreads, all the ImageThreads, all the
methods with bytecodes, all the compiled methods, etc. without having to
explicitly enumerate through them [AllJavaThreads, AllImageThreads,
AllBytecodeMethods, AllCompiledMethods]
(23) Given a thread ID, find the corresponding ImageThread
[AllImageThreads]
(24) Given a ImageThread, find the JavaThread that is associated with that
ImageThread [AllImageThreads, AllJavaThreads]
(25) List all the ImageThreads that are not associated with a JavaThread
[not currently implemented]
(26) Given a JavaThread, find the JavaMonitor that this thread is currently
waiting on (if any) [OneJavaThread]
(27) Given a JavaThread, find all the JavaMonitors that are currently owned
by this thread [AllJavaMonitors]
(28) Given a JavaObject that happens to represent a java/lang/Thread, find
the corresponding JavaThread [AllJavaThreads]
(29) Given a JavaObject that happens to represent a java/lang/Class, find
the corresponding JavaClass [not implemented?]
(30) Given a ImageModule or JavaHeap, find the total size allocated to this
area [OneImageModule, OneJavaHeap]
(31) Given a ImagePointer and a ImageSection, check if that address falls
within that section [OneImageSection]
(32) Annotate an ImageSection with a user-specified description, that
replaces the default section name which if often not useful
[OneImageSection]
(33) Given a ImageStackFrame or JavaStackFrame, check if it is identical to
another frame (to detect duplicate frames in stack traces)
[OneImageStackFrame, OneJavaStackFrame]
(34) Given a ImageThread and a register name, get the value of that
register without having to enumerate through all registers [OneImageThread]
(35) Given a ImageThread, get the address at the top of the stack,
accounting for the fact that the stack pointer is maintained in different
registers (and different directions) on different platforms
[OneImageThread, partial implementation]
(36) Given a JavaHeap, get the current used space, free space, etc. and as
many other GC-related statistics as possible [not implemented]
(37) List all the ImageSections associated with all JavaHeaps, or all
JavaThreads, or all ImageModules, or all compiled methods, etc.
[SectionsTableBase + All* analyzers for each type]
==============================================================
Daniel P. Julin
Senior Technical Staff Member
Technical Area Lead - IBM AIM WebSphere Serviceability Team
EMail: dpj@us.ibm.com
Phone: 347-273-2093 Tie-line 930-3468
Notes: Daniel Julin/Pittsburgh/IBM@IBMUS
Re: System Properties and Env vars
Posted by Steve Poole <sp...@googlemail.com>.
On Tue, Jan 12, 2010 at 2:38 AM, Daniel Julin <dp...@us.ibm.com> wrote:
>
> For what it's worth, the DTFJ-based DumpAnalyzer tool also has a module for
> extracting system properties: com.ibm.dtfj.analyzer.jvm.SystemProperties.
> It works by using DTFJ API calls to find the java.lang.System class
> instance in the list of classloaders, then traverse it, then interpret the
> fields in the java.util.Properties instance that it finds there.
>
> This is one of about 100 "utility" modules that have been built on top of
> DTFJ to facilitate the implementation of various common functions in
> DumpAnalyzer. I think that this may demonstrate that there are too many
> "common functions" and variations to hard-wire into a single low-level API.
> We really do need a low level API _and_ a powerful library on top of it.
> Similar to how in the runtime, we have the "low-level" JVM and a powerful
> class library on top of it... I guess this is similar to the "commons"
> library that Steve mentions below
>
Yes that's right. I'm sure we'd all be interested in knowing more about
these utilitiy modules you mention - could you provide a basic 1 liner
description for the most important ones?
> Incidentally, this simple example of a module for system properties is not
> as simple as it sounds. In a complex J2EE or OSGI world with a complex
> hierarchy of classloaders, we can't be sure that there will always be a
> single definition of java.lang.System, or of other similar classes that
> hold similar properties associated with various J2EE or OSGI modules. If
> that's the case, we would need to have an equivalent collection of analysis
> modules that deal with these variations.
>
>
> As for environment variables, I have not looked in Kato recently, but in
> DTFJ, they used to be available, but from the ImageProcess, not from the
> JavaRuntime.
>
> Your right - they are available from ImageProcess but I wanted to make them
available from JavaRuntime as well .
>
>
> -- Daniel --
>
> ==============================================================
> Daniel P. Julin
> Senior Technical Staff Member
> Technical Area Lead - IBM AIM WebSphere Serviceability Team
>
> EMail: dpj@us.ibm.com
> Phone: 347-273-2093 Tie-line 930-3468
> Notes: Daniel Julin/Pittsburgh/IBM@IBMUS
>
>
>
> Steve Poole <sp...@googlemail.com> wrote on 2010-01-11 11:23:53 AM:
>
> > [image removed]
> >
> > Re: System Properties and Env vars
> >
> > Steve Poole
> >
> > to:
> >
> > kato-spec
> >
> > 10-01-11 11:24 AM
> >
> > Please respond to kato-spec
> >
> > On Mon, Jan 11, 2010 at 4:19 PM, David Griffiths
> > <da...@gmail.com>wrote:
> >
> > > Agreed about the system properties (as a Map). Env vars used to be
> > > available via ImageProcess but I guess you want to move away from
> > > that?
> > >
> > > I would like have the same data available from the JavaRuntime for
> > completeness since java.lang.System provides both properties and envvars
> > and , as you've guessed, because the ImageProcess might not be
> available.
> >
> >
> >
> > Cheers,
> > >
> > > Dave
> > >
> > > On Mon, Jan 11, 2010 at 3:21 PM, Steve Poole <spoole167@googlemail.com
> >
> > > wrote:
> > > > Is it just me or does anyone else think its an unneccessary
> restriction
> > > that
> > > > the JavaRuntime API does not offer the ability to get to System
> > > Properties
> > > > or Environmental variables? You can get them from a running JVM so
> why
> > > > should we not have the same capability on our API?
> > > >
> > > >
> > > > --
> > > > Steve
> > > >
> > >
> >
> >
> >
> > --
> > Steve
>
--
Steve
Re: System Properties and Env vars
Posted by Daniel Julin <dp...@us.ibm.com>.
For what it's worth, the DTFJ-based DumpAnalyzer tool also has a module for
extracting system properties: com.ibm.dtfj.analyzer.jvm.SystemProperties.
It works by using DTFJ API calls to find the java.lang.System class
instance in the list of classloaders, then traverse it, then interpret the
fields in the java.util.Properties instance that it finds there.
This is one of about 100 "utility" modules that have been built on top of
DTFJ to facilitate the implementation of various common functions in
DumpAnalyzer. I think that this may demonstrate that there are too many
"common functions" and variations to hard-wire into a single low-level API.
We really do need a low level API _and_ a powerful library on top of it.
Similar to how in the runtime, we have the "low-level" JVM and a powerful
class library on top of it... I guess this is similar to the "commons"
library that Steve mentions below?
Incidentally, this simple example of a module for system properties is not
as simple as it sounds. In a complex J2EE or OSGI world with a complex
hierarchy of classloaders, we can't be sure that there will always be a
single definition of java.lang.System, or of other similar classes that
hold similar properties associated with various J2EE or OSGI modules. If
that's the case, we would need to have an equivalent collection of analysis
modules that deal with these variations.
As for environment variables, I have not looked in Kato recently, but in
DTFJ, they used to be available, but from the ImageProcess, not from the
JavaRuntime.
-- Daniel --
==============================================================
Daniel P. Julin
Senior Technical Staff Member
Technical Area Lead - IBM AIM WebSphere Serviceability Team
EMail: dpj@us.ibm.com
Phone: 347-273-2093 Tie-line 930-3468
Notes: Daniel Julin/Pittsburgh/IBM@IBMUS
Steve Poole <sp...@googlemail.com> wrote on 2010-01-11 11:23:53 AM:
> [image removed]
>
> Re: System Properties and Env vars
>
> Steve Poole
>
> to:
>
> kato-spec
>
> 10-01-11 11:24 AM
>
> Please respond to kato-spec
>
> On Mon, Jan 11, 2010 at 4:19 PM, David Griffiths
> <da...@gmail.com>wrote:
>
> > Agreed about the system properties (as a Map). Env vars used to be
> > available via ImageProcess but I guess you want to move away from
> > that?
> >
> > I would like have the same data available from the JavaRuntime for
> completeness since java.lang.System provides both properties and envvars
> and , as you've guessed, because the ImageProcess might not be
available.
>
>
>
> Cheers,
> >
> > Dave
> >
> > On Mon, Jan 11, 2010 at 3:21 PM, Steve Poole <sp...@googlemail.com>
> > wrote:
> > > Is it just me or does anyone else think its an unneccessary
restriction
> > that
> > > the JavaRuntime API does not offer the ability to get to System
> > Properties
> > > or Environmental variables? You can get them from a running JVM so
why
> > > should we not have the same capability on our API?
> > >
> > >
> > > --
> > > Steve
> > >
> >
>
>
>
> --
> Steve
Re: System Properties and Env vars
Posted by Steve Poole <sp...@googlemail.com>.
On Mon, Jan 11, 2010 at 4:19 PM, David Griffiths
<da...@gmail.com>wrote:
> Agreed about the system properties (as a Map). Env vars used to be
> available via ImageProcess but I guess you want to move away from
> that?
>
> I would like have the same data available from the JavaRuntime for
completeness since java.lang.System provides both properties and envvars
and , as you've guessed, because the ImageProcess might not be available.
Cheers,
>
> Dave
>
> On Mon, Jan 11, 2010 at 3:21 PM, Steve Poole <sp...@googlemail.com>
> wrote:
> > Is it just me or does anyone else think its an unneccessary restriction
> that
> > the JavaRuntime API does not offer the ability to get to System
> Properties
> > or Environmental variables? You can get them from a running JVM so why
> > should we not have the same capability on our API?
> >
> >
> > --
> > Steve
> >
>
--
Steve
Re: System Properties and Env vars
Posted by David Griffiths <da...@gmail.com>.
Agreed about the system properties (as a Map). Env vars used to be
available via ImageProcess but I guess you want to move away from
that?
Cheers,
Dave
On Mon, Jan 11, 2010 at 3:21 PM, Steve Poole <sp...@googlemail.com> wrote:
> Is it just me or does anyone else think its an unneccessary restriction that
> the JavaRuntime API does not offer the ability to get to System Properties
> or Environmental variables? You can get them from a running JVM so why
> should we not have the same capability on our API?
>
>
> --
> Steve
>
Re: System Properties and Env vars
Posted by Steve Poole <sp...@googlemail.com>.
Also what form should the return value be for these- a Properties object or
a Map<String,String> ? (I favour the second cos its explicit )
On Mon, Jan 11, 2010 at 3:39 PM, Steve Poole <sp...@googlemail.com>wrote:
> I rest my case then - we can't ask users to do that sort of thing to get
> something as useful as System Properties etc. In my JDump case I have a
> dump with the data but its not stored as JavaObjects. The data is sitting
> to one side. If we don't add an API to get at the data I'd have to fake
> out JavaObjects just so the user can go searching :-(
>
> I wonder if we should factor out your code Stuart so it could be used by
> implementations that do not have the data to one side? Do you think we
> need the equivalent of the commons module that provides common functions for
> using the API? (the commons module is not the right place as its designed
> to offer common services to implementers - not consumers)
>
> On Mon, Jan 11, 2010 at 3:27 PM, Stuart Monteith <st...@stoo.me.uk>wrote:
>
>> I've implemented a command to extract system properties in KatoView.
>> The source is here:
>>
>>
>> https://svn.apache.org/repos/asf/incubator/kato/trunk/org.apache.kato/kato.tools.katoview/src/main/java/org/apache/kato/katoview/commands/infocommands/InfoPropertyCommand.java
>>
>> Regards,
>> Stuart
>>
>>
>> Steve Poole wrote:
>>
>>> Is it just me or does anyone else think its an unneccessary restriction
>>> that
>>> the JavaRuntime API does not offer the ability to get to System
>>> Properties
>>> or Environmental variables? You can get them from a running JVM so why
>>> should we not have the same capability on our API?
>>>
>>>
>>>
>>>
>>
>> --
>> Stuart Monteith
>> http://blog.stoo.me.uk/
>>
>>
>
>
> --
> Steve
>
--
Steve
Re: System Properties and Env vars
Posted by Steve Poole <sp...@googlemail.com>.
I rest my case then - we can't ask users to do that sort of thing to get
something as useful as System Properties etc. In my JDump case I have a
dump with the data but its not stored as JavaObjects. The data is sitting
to one side. If we don't add an API to get at the data I'd have to fake
out JavaObjects just so the user can go searching :-(
I wonder if we should factor out your code Stuart so it could be used by
implementations that do not have the data to one side? Do you think we
need the equivalent of the commons module that provides common functions for
using the API? (the commons module is not the right place as its designed
to offer common services to implementers - not consumers)
On Mon, Jan 11, 2010 at 3:27 PM, Stuart Monteith <st...@stoo.me.uk> wrote:
> I've implemented a command to extract system properties in KatoView.
> The source is here:
>
>
> https://svn.apache.org/repos/asf/incubator/kato/trunk/org.apache.kato/kato.tools.katoview/src/main/java/org/apache/kato/katoview/commands/infocommands/InfoPropertyCommand.java
>
> Regards,
> Stuart
>
>
> Steve Poole wrote:
>
>> Is it just me or does anyone else think its an unneccessary restriction
>> that
>> the JavaRuntime API does not offer the ability to get to System Properties
>> or Environmental variables? You can get them from a running JVM so why
>> should we not have the same capability on our API?
>>
>>
>>
>>
>
> --
> Stuart Monteith
> http://blog.stoo.me.uk/
>
>
--
Steve
Re: System Properties and Env vars
Posted by Stuart Monteith <st...@stoo.me.uk>.
I've implemented a command to extract system properties in KatoView.
The source is here:
https://svn.apache.org/repos/asf/incubator/kato/trunk/org.apache.kato/kato.tools.katoview/src/main/java/org/apache/kato/katoview/commands/infocommands/InfoPropertyCommand.java
Regards,
Stuart
Steve Poole wrote:
> Is it just me or does anyone else think its an unneccessary restriction that
> the JavaRuntime API does not offer the ability to get to System Properties
> or Environmental variables? You can get them from a running JVM so why
> should we not have the same capability on our API?
>
>
>
--
Stuart Monteith
http://blog.stoo.me.uk/