Bug#881781: stenc: FTBFS on non-Linux: memset, memcpy, strncmp undeclared

[Denys Berkovskyy]

I have uploaded updated version which fixes inclusion of   and
 to debian mentors (RFS bug #882018).

>> The software uses  header, which does not exist on BSD systems.
> 
> Or on the Hurd, which is a different beast altogether.  I suspect we may
> be out of luck on that front unless  suffices for
> stenc's purposes.

I tried using  header on a FreeBSD system and
compilation still failed because of missing type declaration. It looks like
porting stenc to non-linux systems will require more work.

Bug#881781: stenc: FTBFS on non-Linux: memset, memcpy, strncmp undeclared

[Aaron M. Ucko]

Denys Berkovskyy  writes:

> I will prepare a patch for debian and report the issue upstream.

Great, thanks!

> The software uses  header, which does not exist on BSD systems.

Or on the Hurd, which is a different beast altogether.  I suspect we may
be out of luck on that front unless  suffices for
stenc's purposes.

-- 
Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org)
http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu

Bug#881781: stenc: FTBFS on non-Linux: memset, memcpy, strncmp undeclared

[Denys Berkovskyy]

> These declarations are absent because upstream for some reason decided
> to conditionalize the inclusion of  (and , for
> that matter) on OS_LINUX.  These are bog-standard headers that should
> be safe to #include unconditionally.  However, if the upstream
> developers really want to be paranoid, they can capitalize on the
> build system's existing checks and conditionalize on HAVE_STRING_H
> and/or HAVE_STDLIB_H as desired.

I agree, build system’s check is cleaner and more portable solution.
I will prepare a patch for debian and report the issue upstream.

Unfortunately I saw couple more hurd and kfreebsd related problems in build 
logs.
The software uses  header, which does not exist on BSD systems.
I will have a look if it is possible to replace this header with 

or other header and make the software functional on hurd-i386 and kfreebsd-*

Bug#881781: stenc: FTBFS on non-Linux: memset, memcpy, strncmp undeclared

[Aaron M. Ucko]

Source: stenc
Version: 1.0.7-1
Severity: important
Tags: upstream
Justification: fails to build from source
User: debian-h...@lists.debian.org
Usertags: hurd-i386

Builds of stenc for hurd-i386 and kfreebsd-* (admittedly not release
architectures) have been failing:

  main.cpp: In function 'int main(int, char**)':
  main.cpp:72:5: error: 'memset' was not declared in this scope
  main.cpp:76:5: error: 'memcpy' was not declared in this scope
  main.cpp:107:13: error: 'strncmp' was not declared in this scope

These declarations are absent because upstream for some reason decided
to conditionalize the inclusion of  (and , for
that matter) on OS_LINUX.  These are bog-standard headers that should
be safe to #include unconditionally.  However, if the upstream
developers really want to be paranoid, they can capitalize on the
build system's existing checks and conditionalize on HAVE_STRING_H
and/or HAVE_STDLIB_H as desired.

Could you please take a look?

Thanks!

-- 
Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org)
http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu