Translations bugfix and update

[VlasovVitaly]

Bug fixed

Ammunition type was not translated.

Improved code readability in this piece of file.

Added lots of helping comments for translators.

[BevapDin]

ammo_name should probably do the translation. I see at least 3 further calls to it where the result is not translated:

src/game.cpp:

add_msg(m_info, _("Out of %s!"), ammo_name(tool->ammo_id).c_str());

src/npctalk.cpp:

phrase.replace(fa, l, ammo_name(me->weapon.ammo_type()) );

src/player.cpp:

add_msg(m_info, _("That %s cannot be used on a %s."), used->tname().c_str(),
                       ammo_name(guntype->ammo).c_str());

That would also be kind of consistent with how item::tname, monster::name, etc. return a translated name.

[VlasovVitaly]

I need to check these cases. Fixes (if they need) will in separate PR.

[BevapDin]

You could also use string_format directly here and get rid of the charges_line variable:

string_format( _("Charges: %d"), charges)

[VlasovVitaly]

Yes. Good.

[BevapDin]

Declaring the variables t_max, temp_fmt etc. outside of the if-blocks where they are used makes the impressions there is some connection between the logic in the block and the code afterwards. I expected something like this:

int f_max;
// set f_max depending on various conditions:
if( condition1 ) {
    f_max = value1;
} else {
    f_max = value2;
}
print( "max: %d", f_max );

But here, there is not such connection, each block uses the variable separately. Therefor the variables should be declared in each block (and you can actually make them const there):

if( condition1 ) {
    const int f_max = value1;
    print( "max: %d", f_max );
} else {
    const int f_max = value2;
    print( "max: %d", f_max );
}

Or maybe just move the dump->push_back(iteminfo( line out of the blocks and at the end as it is common to all blocks.

[VlasovVitaly]

Both variants has strengths and weaknesses. I like first one, but this is not obvious behaviour. Second is obvious, but require many lines in every logic block...

Guess, i will choose moving call of dump->push_back() to end of logic. This is not so strange.