[redland-dev] [Redland RDF API 0000372]: serialize_model_to_iostream closes iostream
Mantis Bug Tracker
mantis-bug-sender at librdf.org
Sun May 23 15:22:30 CEST 2010
The following issue has been SUBMITTED.
======================================================================
http://bugs.librdf.org/mantis/view.php?id=372
======================================================================
Reported By: normang
Assigned To:
======================================================================
Project: Redland RDF API
Issue ID: 372
Category: api
Reproducibility: always
Severity: major
Priority: normal
Status: new
======================================================================
Date Submitted: 2010-05-23 13:22
Last Modified: 2010-05-23 13:22
======================================================================
Summary: serialize_model_to_iostream closes iostream
Description:
Consider:
status = librdf_serializer_serialize_model_to_iostream(serializer,
base_uri,
model,
iostr);
raptor_free_iostream(iostr);
Tracing this in the debugger, it turns out that
librdf_serializer_serialize_model_to_iostream calls raptor_free_iostream
internally:
(gdb) where
#0 0x014cf219 in raptor_free_iostream ()
#1 0x014c199e in raptor_serialize_end ()
#2 0x01446a2a in librdf_serializer_raptor_serialize_stream_to_iostream
()
#3 0x01446a92 in librdf_serializer_raptor_serialize_model_to_iostream ()
This causes a segfault when the application calls raptor_free_iostream
separately, which one would expect was the right thing to do, and which
isn't warned against in the docs (or at least not anywhere obvious).
Suggestions:
* avoid freeing the iostream within the serializer function
* document that this happens, so that the iostream should not be freed
* (optimal?) make raptor_free_iostream robust against re-freeing objects
======================================================================
Issue History
Date Modified Username Field Change
======================================================================
2010-05-23 13:22 normang New Issue
======================================================================
More information about the redland-dev
mailing list