The main problem is that you've brought very little context - wtf does "outdated socket support" mean to a person who haven't seen your project

Also I'm going ahead and assume that by "update" you mean "refactor"

IOW you have a legacy project with some code that implements communication over sockets so you're migrating this project to use Asio instead but some class doesn't have a function what you expect it to have

Let me remind you that you're talking about that legacy class in your project WE CAN'T ACCESS

I am not smart, but are you forced to use the legacy code, and if so, why?

and you say it has an Underflow() function and you assume we can find a function in Asio that does the same thing as the function mentioned in your legacy class does without even describing it

He left.

I think he didn't like you millenium

Sockets are an implementation detail. "Managing" implementation details is not a software goal

@milleniumbug Man. You did it again. How can you read so well between the lines

@sehe Hey, can I ask

You can try :) (fyi I'm headed for bed soon)

Do you come here often (or on lounge)?

Mostly lounge, but I check this room, especially when I move messages there

I think it should be in the stats of the room, or perhaps my chat user profile

Do you write C++ professionally?

@EnnMichael I like to think so

So

What is that like?

Do you get mad a lot when having to deal with other people being mad?

Being mad?

If you mean "other people writing bad code" - yeah. I hate that

k catch you later

Bye bye

ca someone compile this project and post an exe for win as i have n oexpereince and am getting it wrong for the past 2 days

github.com/TadasBaltrusaitis/OpenFace/wiki/Windows-Installation

← 2 messages moved from Lounge<C++>

The binaries are right there for download.

i know

if i follow the documentation down to the wire i keep getting errors such as **.lib is not a valid win32 application

and then it cant find cl.exe even though my installation has happened twice and works fine

latest error is that landmark.lib is not a valid win32 application

ip extracted the iwn32 binaries into the code folder in the places that have the same titles

created a debug folder

try to build but no joy

i have n oexperience with vs2015 and was hopeful that someone might just be able to build the app and post the exe on github or something like that

← 8 messages moved from Lounge<C++>

@user3504751 The chance that you'll make your own build work if you can't make theirs work are pretty slim

@user3504751 libs are not applications, so there's that.

i dont want my own build , i just want to build what they have done into an exe

they released an older version a while back and havent made it into an exe and the code and binaries are there to do that i thought ?

i thought it just needs someone who uses vs2015 to build the app

with what they have

@user3504751 That sounds like it could make sense. It depends on them though. And on you to be the person using vs2015 I guess

agred

agreed

Hello all

All smiles

so i was thinking , download vs2015 and their project , download the binary for win32 as listed , follow the documentation which effectively says just build it pressing ctrl f5 and i was hoping it would output an exe file of their latest app code , thats where i am right now and tried for the past two days but to no avail , so i was hopeful that soeone who knew what they were doing would be able to just download the code and binary , put them together and build it, post it and happy days ..

I haven't looked at it enough to know, but I'd expect you either download the binaries, or you build them (using the source and the vs2015).

I'm running into a 'expected member name or ';' after declaration specifiers' error when building c++ sources that include wchar.h in the source headers. Specifically the compiler complains at the lines char __reserved[x], where x is a number (4, 44, 36, etc). I'm not familiar with c++…is there a c++ specific wchar.h that should be referenced, or is there a specific order in which wchar.h should be included?

Based on where the error caret points, the compiler wants to see char __reserved; or char __reserved.[x] but I'm not keen on modifying system headers

You should never alter system headers.

Check that the exact version of your toolchain (compiler and library versions) is supported. If not, file a bug tagged with "Build error"

@MonaLisaOverdrive If you can create a minimal example of a code which fails with wchar.h included, we might be able to provide more help.

Ah - didn't notice the change of asker

Toolchain is supported. @sehe, would uploading the relevant source file and header be acceptable?

I suppose. I don't know whether someone has the toolchain to test

right now you could simply improve the context by not saying what you think the compiler expects based on "where the error caret points"

@MonaLisaOverdrive Which compiler/toolchain are you running into issues with?

clang++

Specifically 3.9.0

@sehe, I'll upload the error as well

I have clang 3.8.1 and can try compiling the source code you are having issues with.

gist.github.com is a nice place to upload stuff like this.

Hi Guys. What should I do, if I have base template class, and child template class. And I want to override += for every child, however += should only affect the same child, but I want to put this as a base class abstract method. what should be a signature in those case? Like that: class<K, V> A, class<K, V>B : public A, class<K, V> C: public A, So B+=B should be ok, but B+=A not and B+=C also not allowed?

@Lalaland: Thanks…unfortunately gist lacks file upload support. Here's the pastebin of the error: pastebin.com/UuVaLBW9

@silent_coder Sorry, don't know the answer. You could stackoverflow.com if no one else in here does.

@silent_coder CRTP, or Boost Operator (boost.org/doc/libs/1_62_0/libs/utility/…)

@MonaLisaOverdrive look at the preprocessed source so see what conflicting define there is for __reserved (I guess)

This should get you the bits: make CMakeFiles/sos.dir/__/__/__/coreclr/hosts/unixcoreruncommon/coreruncommon.cpp.i

Source file: pastebin.com/M0U150ge and header: pastebin.com/AkN8eiSc. Unfortunately the file and header are part of a larger project with more files, so I'm not sure if you'll be able to do anything with those links. I'd write something, but again, very little c++ knowledge

@MonaLisaOverdrive what's the platform (OS, standard library and version)

@sehe, will try …thought I addressed all of them by running grep -rn __reserved srcdir/

@sehe Android, bionic, L (I think…will check)

@MonaLisaOverdrive Yeah. That's a good start as well. The compiler is tripping over something excruciatingly simple though. If char __reserved[4]; is not ok, all I can think of is a preprocessor mishap changing the meaning of char (UB), __reserved (IB because meddles with stdlib internals) or 4 (I don't think that is possible).

Either that, or there is a missing ; before that line

There's a newline before and after the include…would that do it?

@MonaLisaOverdrive Are you trying to build coreclr for android? github.com/dotnet/coreclr/issues/952

Thanks @Sehe So idea is to create base class with 3 template params, like <K, V, Item>, and then make childrens as partial specialization like B<K, V> : public A<K, V, B>?

@MonaLisaOverdrive what include? And no. After preprocessing includes obviously don't exist, and newlines (outside literals or single line comments) become meaningless

@sehe rebuild, actually. Built it successfully once, committed all changes and forgot to push to remote before deleting my local repo

And strike.h has the include for wchar.h

@silent_coder If you mean the Item would be the derived type, then yes that is would be CRTP. But it'd be struct Derived<K,V> : Base<K,V,Derived<K,V>>

@sehe yes, that's what I mean. Thanks again.

@MonaLisaOverdrive Well. The good news is it would have worked. Can you retrieve the revision that worked for you earlier (perhaps there have been breaking changes).

@sehe Actually thought of that but sadly, no. I pulled from master the first time and didn't note the label/tag. I tried to pinpoint build using timestamps I posted issues, but no dice

Mmmm. Seems like you're doing all the right steps to isolate the issue. You might be having to file an issue. Are you using the customized version of Bionic? (see the link earlier)

FYI, getting make: *** No rule to make target '/data/data/com.termux/files/pkg/ps/coreclr/src/coreclr/hosts/unixcoreruncommon/‌​coreruncommon.cpp.i'. Stop. when running that make command, likely bc srcdir and obj dir are in two different places. Will try hunting down other __reserved

@sehe There's a customized version of Bionic?

@MonaLisaOverdrive unlikely. It's possible that preprocessed targets aren't always created by CMake

Back in a sec…

> (see the link earlier)

Err…can't find it…

@silent_coder This is somewhat related. Maybe it gives you an idea or a more specific question.

E.g. this comment github.com/dotnet/coreclr/issues/952#issuecomment-100558706

Ah yes. There's a newer open issue…github.com/dotnet/coreclr/issues/1097

Turns out __reserved is in a lot of places…

Cool I hope you make progress soon

@sehe Made changes, rebuilding now. Should know in about an hour if it worked

i had to pop out back now