Re: [libxml++] SIGABRT received during xmlpp::Document::write_to_string_formatted

[Jonathan W. <co...@co...>]

On Mon, Jan 10, 2005 at 10:27:09PM +0100, Cyril PICARD wrote:

> Hi
> 
> I'm writing an XML file with the xmlpp interface.
> Everything is fine with xmlpp::Document::write_to_file_formatted();
> But when I'm calling xmlpp::Document::write_to_string_formatted(), I 
> receive a SIGABRT signal.
> 
> Any idea around ?
> 
> Thanks

I'll have a guess ...

> * gdb info st :
> 
> #0  0x403aa621 in kill () from /lib/libc.so.6
> #1  0x4017826b in raise (sig=6) at signals.c:65
> #2  0x403aba53 in abort () from /lib/libc.so.6
> #3  0x402f7895 in __cxxabiv1::__terminate(void (*)()) () from 
> /usr/lib/libstdc++.so.3
> #4  0x402f7880 in __cxxabiv1::__terminate(void (*)()) () from 
> /usr/lib/libstdc++.so.3
> #5  0x402f79e0 in __cxa_rethrow () from /usr/lib/libstdc++.so.3
> #6  0x402ede29 in std::__throw_length_error(char const*) () from 
> /usr/lib/libstdc++.so.3
> #7  0x0804aef1 in std::string::_Rep::_S_create(unsigned, 
> std::allocator<char> const&) (__capacity=4294967295,
>     __alloc=@0xbffff500) at /usr/include/g++-v3/bits/basic_string.tcc:371
> #8  0x0804ab28 in char* std::string::_S_construct<char const*>(char 
> const*, char const*, std::allocator<char> const&, 
> std::forward_iterator_tag) (
>     __beg=0x80516e0 "<?xml version=\"1.0\" 
> encoding=\"ISO-8859-1\"?>\n<!DOCTYPE GEDCOM SYSTEM 
> \"gedcom60.dtd\">\n<!--Generated by libgedcomparser - part of GHOSTS 
> project-->\n<GEDCOM>blah ?</GEDCOM>\n", __end=0x80516df "",
>     __a=@0xbffff500) at /usr/include/g++-v3/bits/basic_string.tcc:143

The frame above has __end < __beg, which is why an exception is thrown.

> #9  0x40034674 in std::string::string(char const*, unsigned, 
> std::allocator<char> const&) (this=0xbffff560,
>     __s=0x80516e0 "<?xml version=\"1.0\" 
> encoding=\"ISO-8859-1\"?>\n<!DOCTYPE GEDCOM SYSTEM 
> \"gedcom60.dtd\">\n<!--Generated by libgedcomparser - part of GHOSTS 
> project-->\n<GEDCOM>blah ?</GEDCOM>\n", __n=4294967295,
>     __a=@0xbffff500) at /usr/include/g++-v3/bits/basic_string.h:666

That seems to be because __n = std::string::npos = size_t(-1) in frame #9,
which must be passed by:

> #10 0x401c40c4 in Glib::ustring::ustring(char const*, unsigned) 
> (this=0xbffff560,
>     src=0x80516e0 "<?xml version=\"1.0\" 
> encoding=\"ISO-8859-1\"?>\n<!DOCTYPE GEDCOM SYSTEM 
> \"gedcom60.dtd\">\n<!--Generated by libgedcomparser - part of GHOSTS 
> project-->\n<GEDCOM>blah ?</GEDCOM>\n", n=169)
>     at /usr/include/g++-v3/bits/stl_alloc.h:571

n = 169 in that frame, but Glib::ustring seems to pass -1 to the
std::string ctor. That looks as though utf8_byte_offset(src,n) returns
-1, which I think means there was an invalid UTF8 character in the
string. Is the data you pass to the string correctly UTF8-encoded?
The gdb output shows the 'é' wasn't converted from latin1 to utf-8.

> #11 0x40025d27 in xmlpp::Document::do_write_to_string(Glib::ustring 
> const&, bool) (this=0xbffff6a0,
>     encoding=@0xbffff5d0, format=true) at document.cc:309

jon

-- 
"A woman drove me to drink, I never had the courtesy to thank her."
	- W.C. Fields