summaryrefslogtreecommitdiffstats
path: root/doc/src/snippets/code/src_sql_kernel_qsqlindex.cpp
diff options
context:
space:
mode:
authorAaron Kennedy <aaron.kennedy@nokia.com>2009-11-09 01:33:14 (GMT)
committerAaron Kennedy <aaron.kennedy@nokia.com>2009-11-10 03:44:16 (GMT)
commit8d7c74f560c263b3933b21d780c6a1980425a4ba (patch)
tree5cfd47381f502722a0666af613fffbf177705396 /doc/src/snippets/code/src_sql_kernel_qsqlindex.cpp
parent74d3a4bf9654c8bbe9139951cc5c9352015438a6 (diff)
downloadQt-8d7c74f560c263b3933b21d780c6a1980425a4ba.zip
Qt-8d7c74f560c263b3933b21d780c6a1980425a4ba.tar.gz
Qt-8d7c74f560c263b3933b21d780c6a1980425a4ba.tar.bz2
During a GC mark the scope chain of QScriptContexts
It appears that simply being in the scope chain of an existant frame isn't sufficient to be marked. This can lead to a QScriptContext scope chain that contains a JSObject that has been collected. For example, this code: QScriptContext *ctxt = engine->pushContext(); ctxt.pushScope(engine->newObject()); previouslyCreatedFunctionObject.call(); // causes a GC can lead to the object added to the scope chain to have become invalid. This leads to hilarity later on. Reviewed-by: Kent Hansen
Diffstat (limited to 'doc/src/snippets/code/src_sql_kernel_qsqlindex.cpp')
0 files changed, 0 insertions, 0 deletions