How to print wchar_t string in lldb? - lldb

Question

Given wchar_t* str; which points to a null-terminated utf32 (or utf16) string, what command should I use to print it in lldb?

Answer 1

I assume you want to print it as utf8. It's a little involved - you need to create a summary provider for the type in python that returns a utf8 string for printing. It's not especially complicated though. Create a little python file like ~/lldb/wcharsummary.py with contents like
import lldb
def wchar_SummaryProvider(valobj, dict):
e = lldb.SBError()
s = u'"'
if valobj.GetValue() != 0:
i = 0
newchar = -1
while newchar != 0:
# read next wchar character out of memory
data_val = valobj.GetPointeeData(i, 1)
size = data_val.GetByteSize()
if size == 1:
newchar = data_val.GetUnsignedInt8(e, 0) # utf-8
if size == 2:
newchar = data_val.GetUnsignedInt16(e, 0) # utf-16
if size == 4:
newchar = data_val.GetUnsignedInt32(e, 0) # utf-32
else:
return '<error>'
if e.fail:
return '<error>'
i = i + 1
# add the character to our string 's'
if newchar != 0:
s = s + unichr(newchar)
s = s + u'"'
return s.encode('utf-8')
Load this in to lldb and set this python function as the summary provider for wchar_t*; easiest to put this in your ~/.lldbinit file for re-use:
command script import ~/lldb/wcharsummary.py
type summary add -F wcharsummary.wchar_SummaryProvider "wchar_t *"
then given some source that has some utf32 encoded characters in 32-bit wchar_t's,
NSString *str = #"こんにちは"; // 5 characters long
wchar_t *str_utf32_wchar = (wchar_t*) [[str dataUsingEncoding:NSUTF32StringEncoding] bytes];
lldb will print them in utf8 for us:
Process 22278 stopped
* thread #1: tid = 0x1c03, 0x0000000100000e92 a.out`main + 146 at a.m:11, stop reason = step over
#0: 0x0000000100000e92 a.out`main + 146 at a.m:11
8
9 NSString *str = #"こんにちは"; // 5 characters long
10 wchar_t *str_utf32_wchar = (wchar_t*) [[str dataUsingEncoding:NSUTF32StringEncoding] bytes];
-> 11 printf ("0x%llx 0x%llx 0x%llx 0x%llx\n", (uint64_t) str_utf32_wchar[0], (uint64_t) str_utf32_wchar[1],
12 (uint64_t) str_utf32_wchar[2], (uint64_t) str_utf32_wchar[3]);
13
14 [pool release];
(lldb) fr va
(NSAutoreleasePool *) pool = 0x0000000100108190
(NSString *) str = 0x0000000100001068 #"こんにちは"
(wchar_t *) str_utf32_wchar = 0x0000000100107f80 "こんにちは"
(lldb) p str_utf32_wchar
(wchar_t *) $0 = 0x0000000100107f80 "こんにちは"
(lldb) x/16b `str_utf32_wchar`
0x100107f80: 0xff 0xfe 0x00 0x00 0x53 0x30 0x00 0x00
0x100107f88: 0x93 0x30 0x00 0x00 0x6b 0x30 0x00 0x00
(lldb)

Answer 2

I've modified Jason's code a little bit to handle wxString directly instead of having to change the summary for int* pointers.
Test it by typing the script command in Xcode debugger console, then pasting the code below and pressing ctrl-D. Then at the lldb prompt again, type type summary add --python-function wxString_SummaryProvider "wxString". Works well with my wxWidgets build.
def wxString_SummaryProvider(valobj, dict):
e = lldb.SBError()
charPointer = valobj.GetChildMemberWithName('m_impl').GetChildMemberWithName('_M_dataplus').GetChildMemberWithName('_M_p')
valobj = charPointer
s = u'"'
if valobj.GetValue() != 0:
i = 0
newchar = -1
while newchar != 0:
# read next wchar character out of memory
data_val = valobj.GetPointeeData(i, 1)
size = data_val.GetByteSize()
if size == 1:
newchar = data_val.GetUnsignedInt8(e, 0) # utf-8
if size == 2:
newchar = data_val.GetUnsignedInt16(e, 0) # utf-16
if size == 4:
newchar = data_val.GetUnsignedInt32(e, 0) # utf-32
else:
return '<error>'
if e.fail:
return '<error>'
i = i + 1
# add the character to our string 's'
# print "char2 = %s" % newchar
if newchar != 0:
s = s + unichr(newchar)
s = s + u'"'
return s.encode('utf-8')