< Previous by Date Date Index Next by Date >
< Previous in Thread Thread Index Next in Thread >

Re: [reSIProcate] HeaderNames issue


Hi Yannik,

What version/revision are you using?  Do you have sample code that illustrates this trap?

Thanks,
Scott

Sent from my iPhone

On 2013-01-17, at 3:40 PM, Yannick Guay <yannick.guay@xxxxxxxxx> wrote:

Hi,

I've recently played around with the resiprocate's sip headers code and came across an issue with Headers::getHeaderName(int type).

My code basically parses sip headers looking for user parameters and whenever they come in as empty, ParseBuffer::fail()  is getting called but is unable to throw the proper exception and is segfaulting instead.

I was able to trace it down to using an uninitialized item inside the HeaderNames array when trying to access an item with an index of "header type = NONE". The uninitialized Data object then gets used in the ParseBuffer::fail to construct the exception text description.

I've temporarily patched the code in 2 places (in our repository):
  • Ensured the given index (Headers::getHeaderName(int type)) is lesser then MAX_HEADERS. Otherwise returns a properly initialized but empty Data object.
  • Appended mErrorContext to the exception description only if it's size is greater than zero.
While this solution works as expected, it is meant as a stop gap measure as seems rather inefficient given that this method is consistently called over and over again. So I'm looking for a way to avoid the need to go through a condition every time Headers::getHeaderName() is called.

Ideally all indexes in the array should be initialized such that I don't end up getting a reference to uninitialized memory but I'm not clear how this should be done. I've explored both resip/stack/Headers.cxx and resip/stack/Headers.hxx and would need guidance as to how Headers::HeaderNames[NONE] should be initialized to something valid.

Can you guys please give me a hand with this issue?

best regards,
-Yannick Guay
_______________________________________________
resiprocate-devel mailing list
resiprocate-devel@xxxxxxxxxxxxxxx
https://list.resiprocate.org/mailman/listinfo/resiprocate-devel