> On Apr 5, 2016, at 2:42 PM, Jeffrey Tan <[email protected]> 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 <[email protected]
> <mailto:[email protected]>> 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
>> <[email protected] <mailto:[email protected]>> 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
>>> <[email protected] <mailto:[email protected]>> 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 <[email protected]
>>> <mailto:[email protected]>> wrote:
>>> On Sat, Mar 26, 2016 at 11:58 PM, Jeffrey Tan <[email protected]
>>> <mailto:[email protected]>> wrote:
>>>> Btw: after patching with Siva's fix http://reviews.llvm.org/D18008
>>>> <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
>>> [email protected] <mailto:[email protected]>
>>> http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-dev
>>> <http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-dev>
>>
>> _______________________________________________
>> lldb-dev mailing list
>> [email protected] <mailto:[email protected]>
>> http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-dev
>> <http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-dev>
>
>
> Thanks,
> - Enrico
> 📩 egranata@.com ☎️ 27683
>
>
Thanks,
- Enrico
📩 egranata@.com ☎️ 27683
_______________________________________________
lldb-dev mailing list
[email protected]
http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-dev
