Hello, I'm working on a personal project. Everything works using my own options processing function, but I am having a hard time implementing one specific aspect of my option processing code using the Boost program_options library. My program has a flag (-j) which takes in two doubles as arguments. Currently, I can invoke my program like so: ./foo -j 1.0 2.0 3 where 3 is treated like a positional. Because the -j option must consume several tokens, it seemed natural that I should use multitoken, like so: vector<double> v; po::options_description desc("Allowed options"); desc.add_options() ("jingle,j", po::value<vector<double> >(&v)->multitoken(), "Test thing"); But this causes positional arguments to get eaten up and added to v when -j is the last flag option. So, in my example, v would contain 1.0, 2.0, and 3.0. I suppose I am looking for some way to indicate that the -j flag consumes exactly two tokens. I've looked over the documentation on the Boost website, but I couldn't find anything about this. Any help would be greatly appreciated. Thanks, -Brad
Hello, Brad: What processing is done by your own parsing function? Positional options after multitoken options are bad practice in general. It's not an obvious and flexible approach. Imagine one of your options changes tokens count, all users off your program will be confused using their old command lines. If you like positionals, maybe it's a workaround to use three positional values since your -j tokens count is fixed? Multitoken options are commonly used in cases when you have a variable length list of inputs, i.e. files. If -j is optional, you can place it's values as positional after your current positional option in your invocation scheme. Regards, Dmitriy. On 20/01/2012, Bradley Jensen <bradley.jensen@gmail.com> wrote:
Hello,
I'm working on a personal project. Everything works using my own options processing function, but I am having a hard time implementing one specific aspect of my option processing code using the Boost program_options library.
My program has a flag (-j) which takes in two doubles as arguments. Currently, I can invoke my program like so: ./foo -j 1.0 2.0 3 where 3 is treated like a positional.
Because the -j option must consume several tokens, it seemed natural that I should use multitoken, like so: vector<double> v; po::options_description desc("Allowed options"); desc.add_options() ("jingle,j", po::value<vector<double> >(&v)->multitoken(), "Test thing");
But this causes positional arguments to get eaten up and added to v when -j is the last flag option. So, in my example, v would contain 1.0, 2.0, and 3.0. I suppose I am looking for some way to indicate that the -j flag consumes exactly two tokens.
I've looked over the documentation on the Boost website, but I couldn't find anything about this. Any help would be greatly appreciated.
Thanks, -Brad _______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users
-- С уважением, Матисон Дмитрий.
Thank you for your response. You have convinced me that allowing positional arguments after a multitoken is a bad idea, and so the Boost program_options library will work for my program. Thanks, -Brad On Fri, Jan 20, 2012 at 8:03 AM, Dmitriy Matison <matison.d@gmail.com> wrote:
Hello, Brad:
What processing is done by your own parsing function? Positional options after multitoken options are bad practice in general. It's not an obvious and flexible approach. Imagine one of your options changes tokens count, all users off your program will be confused using their old command lines. If you like positionals, maybe it's a workaround to use three positional values since your -j tokens count is fixed? Multitoken options are commonly used in cases when you have a variable length list of inputs, i.e. files. If -j is optional, you can place it's values as positional after your current positional option in your invocation scheme.
Regards, Dmitriy.
On 20/01/2012, Bradley Jensen <bradley.jensen@gmail.com> wrote:
Hello,
I'm working on a personal project. Everything works using my own options processing function, but I am having a hard time implementing one specific aspect of my option processing code using the Boost program_options library.
My program has a flag (-j) which takes in two doubles as arguments. Currently, I can invoke my program like so: ./foo -j 1.0 2.0 3 where 3 is treated like a positional.
Because the -j option must consume several tokens, it seemed natural that I should use multitoken, like so: vector<double> v; po::options_description desc("Allowed options"); desc.add_options() ("jingle,j", po::value<vector<double> >(&v)->multitoken(), "Test thing");
But this causes positional arguments to get eaten up and added to v when -j is the last flag option. So, in my example, v would contain 1.0, 2.0, and 3.0. I suppose I am looking for some way to indicate that the -j flag consumes exactly two tokens.
I've looked over the documentation on the Boost website, but I couldn't find anything about this. Any help would be greatly appreciated.
Thanks, -Brad _______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users
-- С уважением, Матисон Дмитрий. _______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users
participants (1)
-
Brad Jensen -
Bradley Jensen
-
Dmitriy Matison