diff options
| author | Stefan Behnel <stefan_ml@behnel.de> | 2012-12-21 22:48:02 +0100 |
|---|---|---|
| committer | Stefan Behnel <stefan_ml@behnel.de> | 2012-12-21 22:48:02 +0100 |
| commit | c37f36f2843f0f9c1ef938920b47978d6e28ed4e (patch) | |
| tree | afea03f1df470ebb8cda16fc7728d4358e06101e /src | |
| parent | 4c6a7f4eba4208cbedc653e3277ebd7dc79b9614 (diff) | |
| download | python-lxml-c37f36f2843f0f9c1ef938920b47978d6e28ed4e.tar.gz | |
add compile time option for Py2 that makes the API always return Unicode strings for names and text instead of byte strings for pure ASCII
Diffstat (limited to 'src')
| -rw-r--r-- | src/lxml/apihelpers.pxi | 6 | ||||
| -rw-r--r-- | src/lxml/includes/etree_defs.h | 9 | ||||
| -rw-r--r-- | src/lxml/python.pxd | 1 |
3 files changed, 13 insertions, 3 deletions
diff --git a/src/lxml/apihelpers.pxi b/src/lxml/apihelpers.pxi index 67f5ff48..18ebd134 100644 --- a/src/lxml/apihelpers.pxi +++ b/src/lxml/apihelpers.pxi @@ -1301,7 +1301,7 @@ cdef object funicode(const_xmlChar* s): cdef Py_ssize_t slen cdef const_xmlChar* spos cdef bint is_non_ascii - if python.IS_PYTHON3: + if python.LXML_UNICODE_STRINGS: return s.decode('UTF-8') spos = s is_non_ascii = 0 @@ -1551,11 +1551,11 @@ cdef inline object _namespacedName(xmlNode* c_node): cdef object _namespacedNameFromNsName(const_xmlChar* href, const_xmlChar* name): if href is NULL: return funicode(name) - elif python.IS_PYTHON3: + elif python.LXML_UNICODE_STRINGS and python.PY_VERSION_HEX >= 0x02060000: return python.PyUnicode_FromFormat("{%s}%s", href, name) else: s = python.PyBytes_FromFormat("{%s}%s", href, name) - if isutf8(_xcstr(s)): + if python.LXML_UNICODE_STRINGS or isutf8(_xcstr(s)): return python.PyUnicode_FromEncodedObject(s, 'UTF-8', NULL) else: return s diff --git a/src/lxml/includes/etree_defs.h b/src/lxml/includes/etree_defs.h index 7440e427..8d4d963e 100644 --- a/src/lxml/includes/etree_defs.h +++ b/src/lxml/includes/etree_defs.h @@ -32,6 +32,15 @@ # define IS_PYTHON3 0 #endif +#if IS_PYTHON3 +#undef LXML_UNICODE_STRINGS +#define LXML_UNICODE_STRINGS 1 +#else +#ifndef LXML_UNICODE_STRINGS +#define LXML_UNICODE_STRINGS 0 +#endif +#endif + #if !IS_PYPY # define PyWeakref_LockObject(obj) (NULL) #endif diff --git a/src/lxml/python.pxd b/src/lxml/python.pxd index 99e19ad9..980f6d0d 100644 --- a/src/lxml/python.pxd +++ b/src/lxml/python.pxd @@ -129,5 +129,6 @@ cdef extern from "etree_defs.h": # redefines some functions as macros cdef bint _isString(object obj) cdef const_char* _fqtypename(object t) cdef object PY_NEW(object t) + cdef bint LXML_UNICODE_STRINGS cdef bint IS_PYTHON3 cdef bint IS_PYPY |