You are viewing a plain text version of this content. The canonical link for it is here.
Posted to c-user@axis.apache.org by John Hawkins <HA...@uk.ibm.com> on 2006/09/19 10:37:07 UTC
Re: xsd deserialization for integer types broken
I won't comment on the broken code but I will comment on the reason why
max and min values on reply msgs aren't validated.
The view was taken that the server was already going to do validation and
therefore that the client could have a reasonable trust in the values
coming across the wire. You can argue either way - it would be ideal if
someone would write it so that validation could be toggled - but there
were more important things to do at the time.
regards,
John.
Kamlesh kumar <ka...@yahoo.com>
18/09/2006 21:30
Please respond to
"Apache AXIS C User List" <ax...@ws.apache.org>
To
axis-c-user@ws.apache.org
cc
Subject
xsd deserialization for integer types broken
The xsd deserialization for integer types is awfully
broken.
For instance, calling the getElementAsInt() API for
the XML fragment
"<foo>30a</foo>"
I would have expected it to either return 30 or
ideally throw an exception. Instead, I get a value of
300 :(
Similarly, calling
"<foo>30 </foo>" returns a value of 300 !
I think the following change might have introduced
this bug.
http://svn.apache.org/viewvc/webservices/axis/trunk/c/src/soap/xsd/Integer.cpp?r1=261408&r2=261497&diff_format=h
Further, why aren't we validating against the maximum
and minimum value for each type ?
For instance, getElementAsShort() for the XML fragment
"<foo1>32768</foo1>" returns -1. I would have expected
AxisC++ to throw an exception saying it exceeds the
max value of short instead of silently returning wrong
values.
I am surprised why this bug hasn't been seen till now.
Regardless, I think it should be top priority to fix
this issue.
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
---------------------------------------------------------------------
To unsubscribe, e-mail: axis-c-user-unsubscribe@ws.apache.org
For additional commands, e-mail: axis-c-user-help@ws.apache.org
Re: xsd deserialization for integer types broken
Posted by Kamlesh kumar <ka...@yahoo.com>.
Thanks, that explains why things are done the way it
is. However, even toggling won't help in my case since
I wany any non-Axis client to be able to send request
to my Axis C++ server. I have no guarantee on what the
client will be sending and therefore we need to
validate during both serializing and deserializing.
-Kamlesh
--- John Hawkins <HA...@uk.ibm.com> wrote:
> I won't comment on the broken code but I will
> comment on the reason why
> max and min values on reply msgs aren't validated.
>
> The view was taken that the server was already going
> to do validation and
> therefore that the client could have a reasonable
> trust in the values
> coming across the wire. You can argue either way -
> it would be ideal if
> someone would write it so that validation could be
> toggled - but there
> were more important things to do at the time.
>
> regards,
> John.
>
>
>
>
>
> Kamlesh kumar <ka...@yahoo.com>
> 18/09/2006 21:30
> Please respond to
> "Apache AXIS C User List"
> <ax...@ws.apache.org>
>
>
> To
> axis-c-user@ws.apache.org
> cc
>
> Subject
> xsd deserialization for integer types broken
>
>
>
>
>
>
> The xsd deserialization for integer types is awfully
> broken.
>
> For instance, calling the getElementAsInt() API for
> the XML fragment
>
> "<foo>30a</foo>"
>
> I would have expected it to either return 30 or
> ideally throw an exception. Instead, I get a value
> of
> 300 :(
>
> Similarly, calling
>
> "<foo>30 </foo>" returns a value of 300 !
>
> I think the following change might have introduced
> this bug.
>
http://svn.apache.org/viewvc/webservices/axis/trunk/c/src/soap/xsd/Integer.cpp?r1=261408&r2=261497&diff_format=h
>
>
> Further, why aren't we validating against the
> maximum
> and minimum value for each type ?
>
> For instance, getElementAsShort() for the XML
> fragment
>
> "<foo1>32768</foo1>" returns -1. I would have
> expected
> AxisC++ to throw an exception saying it exceeds the
> max value of short instead of silently returning
> wrong
> values.
>
> I am surprised why this bug hasn't been seen till
> now.
>
> Regardless, I think it should be top priority to fix
> this issue.
>
>
>
>
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam? Yahoo! Mail has the best spam
> protection around
> http://mail.yahoo.com
>
>
---------------------------------------------------------------------
> To unsubscribe, e-mail:
> axis-c-user-unsubscribe@ws.apache.org
> For additional commands, e-mail:
> axis-c-user-help@ws.apache.org
>
>
>
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
---------------------------------------------------------------------
To unsubscribe, e-mail: axis-c-user-unsubscribe@ws.apache.org
For additional commands, e-mail: axis-c-user-help@ws.apache.org