Infinite loop when parsing auguments

[hasindu2008]

When parsing the arguments under each subprogram getopt_long function has been used as follows.

nanopolish_call_methylation.cpp:    for (char c; (c = getopt_long(argc, argv, shortopts, longopts, NULL)) != -1;) {
nanopolish_call_variants.cpp:    for (char c; (c = getopt_long(argc, argv, shortopts, longopts, NULL)) != -1;) {
nanopolish_extract.cpp:    for (char c; (c = getopt_long(argc, argv, shortopts, longopts, NULL)) != -1;) {
nanopolish_getmodel.cpp:    for (char c; (c = getopt_long(argc, argv, shortopts, longopts, NULL)) != -1;) {
nanopolish_index.cpp:    for (char c; (c = getopt_long(argc, argv, shortopts, longopts, NULL)) != -1;) {
nanopolish_methyltrain.cpp:    for (char c; (c = getopt_long(argc, argv, shortopts, longopts, NULL)) != -1;) {
nanopolish_phase_reads.cpp:    for (char c; (c = getopt_long(argc, argv, shortopts, longopts, NULL)) != -1;) {
nanopolish_scorereads.cpp:    for (char c; (c = getopt_long(argc, argv, shortopts, longopts, NULL)) != -1;) {
nanopolish_train_poremodel_from_basecalls.cpp:    for (char c; (c = getopt_long(argc, argv, shortopts, longopts, NULL)) != -1;) {

Variable c has been declared as a char and it is compared with -1. In x86 this works fine as char is signed. However, in architectures such as ARM and PowerPC char is unsigned by default and hence it causes an infinite loop.
See : http://www.network-theory.co.uk/docs/gccintro/gccintro_71.html

A simple fix would be to either include -fsigned-char to the compilation flags in makefile or explicitly declaring c as signed char, for instance

nanopolish_call_methylation.cpp:    for (signed char c; (c = getopt_long(argc, argv, shortopts, longopts, NULL)) != -1;) {

[jts]

Thanks for reporting this, I have added -fsigned-char.