hi,
this patch properly addresses escaping < and > which might appear
(e.g. for C++ operators) in the xml dump.
I also re-enabled the unit test for ast-print-xml (or so I think)
at least, make test didn't fail...
cheers,
sebastien.
hi,
this patch properly addresses escaping < and > which might appear
(e.g. for C++ operators) in the xml dump.
I also re-enabled the unit test for ast-print-xml (or so I think)
at least, make test didn't fail...
cheers,
sebastien.
ENOPATCH
Excerpts from Chris Lattner's message of 2010-05-11 19:04:16 +0200:
ENOPATCH
arf !
Is this hunk intended?
+++ lib/Frontend/DeclXML.cpp (working copy)
@@ -49,7 +49,7 @@
addSubNodes(cast<RecordDecl>(RD));
if (RD->isDefinition()) {
- Doc.addAttribute("num_bases", RD->getNumBases());
+ //Doc.addAttribute("num_bases", RD->getNumBases());
-Chris
Excerpts from Chris Lattner's message of 2010-05-11 19:13:42 +0200:
> Excerpts from Chris Lattner's message of 2010-05-11 19:04:16 +0200:
>> ENOPATCH
>
> arf !Is this hunk intended?
+++ lib/Frontend/DeclXML.cpp (working copy)
@@ -49,7 +49,7 @@
addSubNodes(cast<RecordDecl>(RD));if (RD->isDefinition()) {
- Doc.addAttribute("num_bases", RD->getNumBases());
+ //Doc.addAttribute("num_bases", RD->getNumBases());
yes (I should indeed have documented it)
the problem is that by re-using the dumping code for RecordDecl, the
CXXRecordDecl node is already closed by the time one tries to add this
new num_bases attribute.
something better needs to be done here, but fixing the XML generation
was the priority
cheers,
sebastien.
Thanks, applied in r103671