[lldb-dev] UnicodeDecodeError for serialize SBValue description

Jeffrey Tan via lldb-dev lldb-dev at lists.llvm.org
Wed Apr 6 19:31:07 PDT 2016


Thanks Enrico. This is very detailed! I will take a look.
Btw: originally, I was hoping that data formatter can be added without
changing the source code. Like giving a xml/json format file telling lldb
the memory layout/structure of the data structure, lldb can parse the
xml/json and deduce the formatting. This is approach used by data
visualizer in VS debugger:
https://msdn.microsoft.com/en-us/library/jj620914.aspx
This will make adding data formatter more extensible/flexible. Any reason
we did not take this approach?

Jeffrey

On Wed, Apr 6, 2016 at 11:49 AM, Enrico Granata <egranata at apple.com> wrote:

>
> On Apr 5, 2016, at 2:42 PM, Jeffrey Tan <jeffrey.fudan at gmail.com> wrote:
>
> Hi Enrico,
>
> Any suggestion/example how to add a data formatter for our own STL string?
> From the output below I can see we are using our own "*fbstring_core*"
> which I assume I need to write a type summary for this type:
>
> frame variable corpus -T
> (const string &const) corpus = error: summary string parsing error: {
>   (std::*fbstring_core*<char>) store_ = {
>     (std::*fbstring_core*<char>::(anonymous union))  = {
>       (char [24]) small_ = "www"
>       (std::fbstring_core<char>::MediumLarge) ml_ = {
>         (char *) data_ = 0x0000000000777777
> "H\x89U\xa8H\x89M\xa0L\x89E\x98H\x8bE\xa8H\x89��_U��D\x88e�H\x8bE\xa0H\x89��]U��H\x89�H\x8dE�H\x89�H\x89���
> ��L\x8dm�H\x8bE\x98H\x89��IU��\x88]�L\x8be\xb0L\x89��
>         (std::size_t) size_ = 0
>         (std::size_t) capacity_ = 1441151880758558720
>       }
>     }
>   }
> }
>
>
> Admittedly, this is going to be a little vague since I haven’t really seen
> your code and I am only working off of one sample
>
> There’s going to be two parts to getting this to work:
>
> *Part 1 - Formatting fbstring_core*
>
> At a glance, an fbstring_core<char> can be backed by two representations.
> A “small” representation (a char array), and a “medium/large"
> representation (a char* + a size)
> I assume that the way you tell one from the other is
>
> if (size == 0) small
> else medium-large
>
> If my assumption is not correct, you’ll need to discover what the correct
> discriminator logic is - the class has to know, and so do you :-)
>
> Armed with that knowledge, look in lldb
> source/Plugins/Language/CPlusPlus/Formatters/LibCxx.cpp
> There’s a bunch of code that deals with formatting llvm’s libc++
> std::string - which follows a very similar logic to your class
>
> ExtractLibcxxStringInfo() is the function that handles discovering which
> layout the string uses - where the data lives - and how much data there is
>
> Once you have told yourself how much data there is (the size) and where it
> lives (array or pointer), LibcxxStringSummaryProvider() has the easy task
> - it sets up a StringPrinter, tells it how much data to print, where to get
> it from, and then delegates the StringPrinter to do the grunt work
> StringPrinter is a nifty little tool - it can handle generating summaries
> for different kinds of strings (UTF8? UTF16? we got it - is a \0 a
> terminator? what quote character would you like? …) - you point it at some
> data, set up a few options, and it will generate a printable representation
> for you - if your string type is doing anything out of the ordinary, let’s
> talk - I am definitely open to extending StringPrinter to handle even more
> magic
>
> *Part 2 - Teaching std::string that it can be backed by an fbstring_core*
>
> At the end of part 1, you’ll probably end up with a
> FBStringCoreSummaryProvider() - now you need to teach LLDB about it
> The obvious thing you could do would be to go in CPlusPlusLanguage
> ::GetFormatters() add a LoadFBStringFormatter(g_category) to it - and
> then imitate - say - LoadLibCxxFormatters()
>
>     AddCXXSummary(cpp_category_sp, lldb_private::formatters::
> FBStringCoreSummaryProvider, “fbstringcore summary provider", ConstString(
> “std::fbstring_core<.+>"), stl_summary_flags, true);
>
> That will work - but what you would see is:
>
> (const string &const) corpus = error: summary string parsing error: {
>   (std::*fbstring_core*<char>) store_ = “www"
>
>
> You wanna do
>
> (lldb) log enable lldb formatters
> (lldb) frame variable -T corpus
>
> It will list one or more typenames - the most specific one is the one you
> like (e.g. for libc++ we get std::__1::string - this is how we tell
> ourselves this is the std::string from libc++)
> Once you find that typename, you’ll make a new formatter -
> FBStringSummaryProvider() - and register that formatter with that very
> specific typename
>
> All that FBStringSummaryProvider() has to do is get the “store_” member
> (ValueObject::GetChildMemberWithName() is your friend) - and pass it down
> to FBStringCoreSummaryProvider()
>
>
> I understand this may seem a little convoluted and arcane at first - but
> feel free to ask more questions, and I’ll try to help out!
>
> Thanks.
> Jeffrey
>
> On Mon, Mar 28, 2016 at 11:38 AM, Enrico Granata <egranata at apple.com>
> wrote:
>
>> This is kind of orthogonal to your problem, but the reason why you are
>> not seeing the kind of simplified printing Greg is suggesting, is because
>> your std::string doesn’t look like any of the kinds we recognize
>>
>> Specifically, LLDB data formatters work by matching against type names,
>> and once they recognize a typename, then they try to inspect the variable
>> in order to grab a summary
>> In your example, your std::string exposes a layout that we are not
>> handling - hence we bail out of the formatter and we fall back to the raw
>> view
>>
>> If you want pretty printing to work, you’ll need to write a data formatter
>>
>> There are a few avenues. The obvious easy one is to extend the existing
>> std::string formatter to recognize your type’s internal layout.
>> If one were signing up for more infrastructure work, they could decide to
>> try and detect shared library loads and load formatters that match with
>> whatever libraries are being loaded.
>>
>> On Mar 28, 2016, at 9:47 AM, Greg Clayton via lldb-dev <
>> lldb-dev at lists.llvm.org> wrote:
>>
>> So you need to be prepared to escape any text that can have special
>> characters. A "std::string" or any container can contain special
>> characters. If you are encoding stuff into JSON, you will either need to
>> escape any special characters, or hex encode the string into ASCII hex
>> bytes.
>>
>> In debuggers we often get bogus data because variables are not
>> initialized, but the compiler tells us that a variable is valid in address
>> range [0x1000-0x2000), but it actually is [0x1200-0x2000). If we read a
>> variable in this case, a std::string might contain bogus data and the bytes
>> might not make sense. So you always have to be prepared for bad data.
>>
>> If we look at:
>>
>>  store_ = {
>>     = {
>>      small_ = "www"
>>      ml_ = (data_ =
>>
>> "��UH\x89�H�}�H\x8bE�]ÐUH\x89�H��H\x89}�H\x8bE�H\x89��~\xb4��\x90��UH\x89�SH\x83�H\x89}�H�u�H�E�H���\x9e���H\x8b\x18H\x8bE�H���O\xb4��H\x89ƿ\b",
>> size_ = 0, capacity_ = 1441151880758558720)
>>    }
>>  }
>> }
>>
>> We can see the "size_" is zero, and capacity_ is 1441151880758558720
>> (which is 0x1400000000000000). "data_" seems to be some random pointer.
>>
>> On MacOSX, we have a special formatting code that displays std::string in
>> CPlusPlusLanguage.cpp that gets installed in the LoadLibCxxFormatters() or
>> LoadLibStdcppFormatters() functions with code like:
>>
>>    lldb::TypeSummaryImplSP std_string_summary_sp(new
>> CXXFunctionSummaryFormat(stl_summary_flags,
>> lldb_private::formatters::LibcxxStringSummaryProvider, "std::string summary
>> provider"));
>>    cpp_category_sp->GetTypeSummariesContainer()->Add(ConstString("std::__1::string"),
>> std_string_summary_sp);
>>
>> Special flags are set on std::string to say "don't show children of this
>> and just show a summary" So if a std::string contained "hello". So for the
>> following code:
>>
>> std::string h ("hello");
>>
>> You should just see:
>>
>> (lldb) fr var h
>> (std::__1::string) h = "hello"
>>
>> If you take a look at the normal value in the raw we see:
>>
>> (lldb) fr var --raw h
>> (std::__1::string) h = {
>>  __r_ = {
>>    std::__1::__libcpp_compressed_pair_imp<std::__1::basic_string<char,
>> std::__1::char_traits<char>, std::__1::allocator<char> >::__rep,
>> std::__1::allocator<char>, 2> = {
>>      __first_ = {
>>         = {
>>          __l = {
>>            __cap_ = 122511465736202
>>            __size_ = 0
>>            __data_ = 0x0000000000000000
>>          }
>>          __s = {
>>             = {
>>              __size_ = '\n'
>>              __lx = '\n'
>>            }
>>            __data_ = {
>>              [0] = 'h'
>>              [1] = 'e'
>>              [2] = 'l'
>>              [3] = 'l'
>>              [4] = 'o'
>>              [5] = '\0'
>>              [6] = '\0'
>>              [7] = '\0'
>>              [8] = '\0'
>>              [9] = '\0'
>>              [10] = '\0'
>>              [11] = '\0'
>>              [12] = '\0'
>>              [13] = '\0'
>>              [14] = '\0'
>>              [15] = '\0'
>>              [16] = '\0'
>>              [17] = '\0'
>>              [18] = '\0'
>>              [19] = '\0'
>>              [20] = '\0'
>>              [21] = '\0'
>>              [22] = '\0'
>>            }
>>          }
>>          __r = {
>>            __words = {
>>              [0] = 122511465736202
>>              [1] = 0
>>              [2] = 0
>>            }
>>          }
>>        }
>>      }
>>    }
>>  }
>> }
>>
>> So the main question is why are our "std::string" formatters not kicking
>> in for you. That comes down to a typename match, or the format of the
>> string isn't what the formatter is expecting.
>>
>> But again, since you std::string can contain anything, you will need to
>> escape any and all text that is encoded into JSON to ensure it doesn't
>> contain anything JSON can't deal with.
>>
>> On Mar 27, 2016, at 9:20 PM, Jeffrey Tan via lldb-dev <
>> lldb-dev at lists.llvm.org> wrote:
>>
>> Thanks Siva. All the DW_TAG_member related errors seems to go away after
>> patching with your fix. The current problem is handling the decoding.
>>
>> Here is the correct decoding from gdb whic might be useful:
>> (gdb) p corpus
>> $3 = (const std::string &) @0x7fd133cfb888: {
>>  static npos = 18446744073709551615, store_ = {
>>    static kIsLittleEndian = <optimized out>,
>>    static kIsBigEndian = <optimized out>, {
>>      small_ = "www", '\000' <repeats 20 times>, "\024", ml_ = {
>>        data_ = 0x777777 <std::_Any_data::_M_access<void
>> folly::fibers::Baton::waitFiber<folly::fibers::FirstArgOf<facebook::servicerouter::RequestDispatcherBase<facebook::servicerouter::ThriftDispatcher>::prepareForSelection(facebook::servicerouter::DispatchContext&)::{lambda(folly::fibers::Promise<facebook::servicerouter::RequestDispatcherBase<facebook::servicerouter::ThriftDispatcher>::prepareForSelection(facebook::servicerouter::DispatchContext&)::SelectionResult>)#1},
>> void>::type::value_type
>> folly::fibers::await<facebook::servicerouter::RequestDispatcherBase<facebook::servicerouter::ThriftDispatcher>::prepareForSelection(facebook::servicerouter::DispatchContext&)::{lambda(folly::fibers::Promise<facebook::servicerouter::RequestDispatcherBase<facebook::servicerouter::ThriftDispatcher>::prepareForSelection(facebook::servicerouter::DispatchContext&)::SelectionResult>)#1}>(folly::fibers::FirstArgOf&&)::{lambda()#1}>(folly::fibers::FiberManager&,
>> folly::fibers::FirstArgOf<folly::fibers::FirstArgOf<facebook::servicerouter::RequestDispatcherBase<facebook::servicerouter::ThriftDispatcher>::prepareForSelection(facebook::servicerouter::DispatchContext&)::{lambda(folly::fibers::Promise<facebook::servicerouter::RequestDispatcherBase<facebook::servicerouter::ThriftDispatcher>::prepareForSelection(facebook::servicerouter::DispatchContext&)::SelectionResult>)#1},
>> void>::type::value_type
>> folly::fibers::await<facebook::servicerouter::RequestDispatcherBase<facebook::servicerouter::ThriftDispatcher>::prepareForSelection(facebook::servicerouter::DispatchContext&)::{lambda(folly::fibers::Promise<facebook::servicerouter::RequestDispatcherBase<facebook::servicerouter::ThriftDispatcher>::prepareForSelection(facebook::servicerouter::DispatchContext&)::SelectionResult>)#1}>(folly::fibers::FirstArgOf&&)::{lambda()#1},
>> void>::type::value_type)::{lambda(folly::fibers::Fiber&)#1}*>() const+25>
>> "\311\303UH\211\345H\211}\370H\213E\370]ÐUH\211\345H\203\354\020H\211}\370H\213E\370H\211\307\350~\264\312\377\220\311\303UH\211\345SH\203\354\030H\211}\350H\211u\340H\213E\340H\211\307\350\236\377\377\377H\213\030H\213E\350H\211\307\350O\264\312\377H\211ƿ\b",
>> size_ = 0,
>>        capacity_ = 1441151880758558720}}}}
>>
>> Utf-16 does not seem to decode it, while 'latin-1' does:
>>
>> '\xc9'.decode('utf-16')
>>
>> Traceback (most recent call last):
>>  File "<stdin>", line 1, in <module>
>>  File
>> "/mnt/gvfs/third-party2/python/55c1fd79d91c77c95932db31a4769919611c12bb/2.7.8/centos6-native/da39a3e/lib/python2.7/encodings/utf_16.py",
>> line 16, in decode
>>    return codecs.utf_16_decode(input, errors, True)
>> UnicodeDecodeError: 'utf16' codec can't decode byte 0xc9 in position 0:
>> truncated data
>>
>> '\xc9'.decode('latin-1')
>>
>> u'\xc9'
>>
>> Instead of guessing what kind of decoding I should use, I would use
>> 'ensure_ascii=False' to prevent the crash for now.
>>
>> I tried to reproduce this crash, but it seems that the crash might be
>> related with some internal stl implementation we are using. I will see if I
>> can narrow down to a small repro later.
>>
>> Thanks
>> Jeffrey
>>
>> On Sun, Mar 27, 2016 at 2:49 PM, Siva Chandra <sivachandra at gmail.com>
>> wrote:
>> On Sat, Mar 26, 2016 at 11:58 PM, Jeffrey Tan <jeffrey.fudan at gmail.com>
>> wrote:
>>
>> Btw: after patching with Siva's fix http://reviews.llvm.org/D18008, the
>> first field 'small_' is fixed, however the second field 'ml_' still emits
>> garbage:
>>
>> (lldb) fr v corpus
>> (const string &const) corpus = error: summary string parsing error: {
>>  store_ = {
>>     = {
>>      small_ = "www"
>>      ml_ = (data_ =
>>
>> "��UH\x89�H�}�H\x8bE�]ÐUH\x89�H��H\x89}�H\x8bE�H\x89��~\xb4��\x90��UH\x89�SH\x83�H\x89}�H�u�H�E�H���\x9e���H\x8b\x18H\x8bE�H���O\xb4��H\x89ƿ\b",
>> size_ = 0, capacity_ = 1441151880758558720)
>>    }
>>  }
>> }
>>
>>
>> Do you still see the DW_TAG_member related error?
>>
>> A wild (and really wild at that) guess: Is it utf16 data that is being
>> decoded as utf8?
>>
>> As David Blaikie mentioned on the other thread, it would really help
>> if you provide us with a minimal example to repro this. Atleast, repro
>> instructions.
>>
>> _______________________________________________
>> lldb-dev mailing list
>> lldb-dev at lists.llvm.org
>> http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-dev
>>
>>
>> _______________________________________________
>> lldb-dev mailing list
>> lldb-dev at lists.llvm.org
>> http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-dev
>>
>>
>>
>> Thanks,
>> *- Enrico*
>> 📩 egranata@.com ☎️ 27683
>>
>>
>
>
> Thanks,
> *- Enrico*
> 📩 egranata@.com ☎️ 27683
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/lldb-dev/attachments/20160406/24c736f3/attachment-0001.html>


More information about the lldb-dev mailing list