‘VerifyNestedFlexBuffer’ is not a member of ‘flexbuffers’ [C++, gcc 11.2.0, Arch Linux, Flatbuffers v2.0.6]

[technobaboo]

As it says in the title, after using flatc to rebuild the flatbuffer schemas that include flexbuffers (https://github.com/StardustXR/libstardustxr/blob/main/src/common/flatbuffers/message.fbs#L9) the error below occurs:

In file included from ../src/fusion/types/data/../../../common/messenger.hpp:9,
                 from ../src/fusion/types/data/../../messenger.hpp:3,
                 from ../src/fusion/types/data/../../fusion_internal.hpp:3,
                 from ../src/fusion/types/data/receiver.cpp:5:
../src/fusion/types/data/../../../common/flatbuffers/message.hpp: In member function ‘bool StardustXR::Message::Verify(flatbuffers::Verifier&) const’:
../src/fusion/types/data/../../../common/flatbuffers/message.hpp:76:25: error: ‘VerifyNestedFlexBuffer’ is not a member of ‘flexbuffers’
   76 |            flexbuffers::VerifyNestedFlexBuffer(data(), verifier) &&
      |

This issue never happened in 2.0.5.

[dbaileychess]

Looks like you need to include flatbuffers.h, see comment:

flatbuffers/include/flatbuffers/flexbuffers.h

Lines 1879 to 1889 in 1a4c405

	#ifdef FLATBUFFERS_H_
	// This is a verifier utility function that works together with the
	// FlatBuffers verifier, which should only be present if flatbuffer.h
	// has been included (which it typically is in generated code).
	inline bool VerifyNestedFlexBuffer(const flatbuffers::Vector<uint8_t> *nv,
	flatbuffers::Verifier &verifier) {
	if (!nv) return true;
	return verifier.Check(flexbuffers::VerifyBuffer(
	nv->data(), nv->size(), verifier.GetFlexReuseTracker()));
	}
	#endif

[technobaboo]

That's the weird part, this is inside the c++ file that flatc generated by using the command

flatc -c *.fbs --gen-mutable --filename-suffix "" --filename-ext "hpp"

(see https://github.com/StardustXR/libstardustxr/blob/main/src/common/flatbuffers/message.hpp#L4-L8)

It includes the file, but still errors. Weirder still, this only ever occurs on v2.0.6 and not when reverting the flatbuffers version back to 2.0.5

[dbaileychess]

If you remove the #ifdef FLATBUFFERS_H_ part does it work? I don't see anything obvious on why it isn't being linked in.

[aardappel]

Looks like flexbuffers.h somehow got included manually before the generated code. Can you scan where you include flexbuffers.h and see what happens if you remove/move it to below the generated code include?

[technobaboo]

That was the problem indeed, really really stressful... wish there was a fix that would work in flatbuffers instead of my codebase.

…

On Mon, Feb 28, 2022 at 2:41 PM Wouter van Oortmerssen < ***@***.***> wrote: Looks like flexbuffers.h somehow got included manually before the generated code. Can you scan where you include flexbuffers.h and see what happens if you remove/move it to below the generated code include? — Reply to this email directly, view it on GitHub <#7134 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABCU4EEFKBQHFFERNSX6MN3U5PFXFANCNFSM5POHB33Q> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[technobaboo]

Should I close this issue or leave it open as something to fix? I'd count it as a bug with a workaround more than an issue with my codebase as I'm sure many people will have this issue.

…

On Wed, Mar 2, 2022 at 4:33 AM Nova King ***@***.***> wrote: That was the problem indeed, really really stressful... wish there was a fix that would work in flatbuffers instead of my codebase. On Mon, Feb 28, 2022 at 2:41 PM Wouter van Oortmerssen < ***@***.***> wrote: > Looks like flexbuffers.h somehow got included manually before the > generated code. Can you scan where you include flexbuffers.h and see > what happens if you remove/move it to below the generated code include? > > — > Reply to this email directly, view it on GitHub > <#7134 (comment)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/ABCU4EEFKBQHFFERNSX6MN3U5PFXFANCNFSM5POHB33Q> > . > Triage notifications on the go with GitHub Mobile for iOS > <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> > or Android > <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. > > You are receiving this because you authored the thread.Message ID: > ***@***.***> >

[dbaileychess]

Yeah, let's leave this open for now, I'll need to look into how to do this correctly.

[aardappel]

The problem is we ideally want to keep these headers independent.

One solution may be to just add a copy of that function to the generated code.

[technobaboo]

It doesn't look very complicated at all, I think that makes the most sense at first glance

…

On Wed, Mar 2, 2022 at 12:52 PM Wouter van Oortmerssen < ***@***.***> wrote: The problem is we ideally want to keep these headers independent. One solution may be to just add a copy of that function to the generated code. — Reply to this email directly, view it on GitHub <#7134 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABCU4EEH4FQRCOK2LT44UM3U56THLANCNFSM5POHB33Q> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[dbaileychess]

@technobaboo can you try out the linked PR to see if it works for you?

[MalekiRe]

I am having this issue.

[dbaileychess]

This is merged now, so do you have the latest code?

[technobaboo]

This was on an older version of my library, and I did eventually get to test the patch! It works perfectly for me, sorry I never updated this issue to reflect that :S

…

On Thu, Apr 14, 2022 at 1:15 PM Derek Bailey ***@***.***> wrote: This is merged now, so do you have the latest code? — Reply to this email directly, view it on GitHub <#7134 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABCU4EB2T6LBUWVEW5MVR33VFBHDBANCNFSM5POHB33Q> . You are receiving this because you were mentioned.Message ID: ***@***.***>