Re: [Debian-med-packaging] Bug#853568: No idea how to fix abs arguments in nanopolish (original) (raw)

Am Donnerstag, den 31.08.2017, 21:00 -0700 schrieb Walter Landry:

Andreas Tille andreas@an3as.eu wrote:

Hi,

to fix bug #853568 I tried a patch (gcc-7.patch) to fix abs() arguments in nanopolish[1] but I have no idea how to deal with this:

... g++ -o src/hmm/nanopolish_pore_model_set.o -c -g -O2 -fdebug- prefix-map=/build/nanopolish-0.5.0=. -fstack-protector-strong -Wformat -Werror=format-security -g -O3 -std=c++11 -fopenmp -Wdate- t src/common/nanopolish_variant.cpp: In function 'std::vector extract_variants(const string&, const string&)': src/common/nanopolish_variant.cpp:32:69: error: call of overloaded 'abs(std::__cxx11::basic_string::size_type)' is ambiguous      size_t difference = std::abs(reference.size() - haplotype.size());

The result of subtracting two size_t's is still a size_t, which is unsigned.  So you need to cast it to a signed type.  The correct type is ptrdiff_t.

  http://en.cppreference.com/w/cpp/types/ptrdiff_t

The line then becomes

  size_t difference = std::abs(static_cast(reference.size() - haplotype.size()));

Casting the difference may result in undefined behavior. Consider the case

reference.size() == 1 haplotype.size() == 2

then 

reference.size() - haplotype.size() 

will be 0xFFFFFFFF (on 32 bit), and how this is casted to a signed type is implementation dependent (i.e. it is not guaranteed that this simply wraps to -1, it may also raise a trap because of integer overflow).

It would be better to avoid the cast altogether by doing something like

  size_t difference = reference.size() > haplotype.size() ? reference.size() - haplotype.size() : haplotype.size() - reference.size();

or cast both values before doing the subtraction.

Best, Gert

Reply to: