feat(net): add {linkSpeed} format replacement

Adds an ethernet link speed ({linkSpeed}) read from
/sys/class/net/<iface>/speed, plus a skip-decimal option on pow_format
to drop a trailing .0. Rebased onto master's store-based network
formatting; existing compact bandwidth args updated for the new
pow_format signature.

include/modules/network.hpp

@@ -54,6 +54,7 @@ class Network : public ALabel {
   const std::string getNetworkState() const;
   void clearIface();
   std::optional<std::pair<unsigned long long, unsigned long long>> readBandwidthUsage();
+  uint32_t readLinkSpeed() const;
 
   int ifid_{-1};
   ip_addr_pref addr_pref_{ip_addr_pref::IPV4};
@@ -94,6 +95,7 @@ class Network : public ALabel {
   uint8_t signal_strength_;
   std::string signal_strength_app_;
   uint32_t route_priority;
+  uint32_t link_speed_{0};
 
   util::SleeperThread thread_;
   util::SleeperThread thread_timer_;

include/util/format.hpp

@@ -5,12 +5,18 @@
 
 class pow_format {
  public:
-  pow_format(long long val, std::string&& unit, bool binary = false, int min_pow_for_decimal = 0)
-      : val_(val), unit_(unit), binary_(binary), min_pow_for_decimal_(min_pow_for_decimal) {};
+  pow_format(long long val, std::string&& unit, bool binary = false, bool skip_decimal = false,
+             int min_pow_for_decimal = 0)
+      : val_(val),
+        unit_(unit),
+        binary_(binary),
+        skip_decimal_(skip_decimal),
+        min_pow_for_decimal_(min_pow_for_decimal) {};
 
   long long val_;
   std::string unit_;
   bool binary_;
+  bool skip_decimal_;
   int min_pow_for_decimal_;
 };
 
@@ -50,17 +56,21 @@ struct formatter<pow_format> {
     const char* units[] = {"", "k", "M", "G", "T", "P", nullptr};
 
     auto base = s.binary_ ? 1024ull : 1000ll;
+    auto div = 1ll;
     auto fraction = (double)s.val_;
 
     int pow;
     for (pow = 0; units[pow + 1] != nullptr && fraction / base >= 1; ++pow) {
       fraction /= base;
+      div *= base;
     }
 
-    auto number_width = 5              // coeff in {:.1f} format
-                        + s.binary_;   // potential 4th digit before the decimal point
-    auto max_width = number_width + 1  // prefix from units array
-                     + s.binary_       // for the 'i' in GiB.
+    auto fixed_precision = (s.skip_decimal_ && ((s.val_ % div) == 0)) ? 0 : 1;
+    auto number_width = 3 + fixed_precision       // coeff in {:.{fixed_precision}f} format
+                        + (fixed_precision != 0)  // float dot
+                        + s.binary_;              // potential digit before the decimal point
+    auto max_width = number_width + 1             // prefix from units array
+                     + s.binary_                  // for the 'i' in GiB.
                      + s.unit_.length();
 
     const char* format;
@@ -71,17 +81,17 @@ struct formatter<pow_format> {
       case '<':
         return fmt::format_to(ctx.out(), "{:<{}}", fmt::format("{}", s), max_width);
       case '=':
-        format = "{coefficient:<{number_width}.1f}{padding}{prefix}{unit}";
+        format = "{coefficient:<{number_width}.{fixed_precision}f}{padding}{prefix}{unit}";
         break;
       case 0:
       default:
         format = pow < s.min_pow_for_decimal_ ? "{coefficient:.0f}{prefix}{unit}"
-                                              : "{coefficient:.1f}{prefix}{unit}";
+                                              : "{coefficient:.{fixed_precision}f}{prefix}{unit}";
         break;
     }
     return fmt::format_to(
         ctx.out(), fmt::runtime(format), fmt::arg("coefficient", fraction),
-        fmt::arg("number_width", number_width),
+        fmt::arg("fixed_precision", fixed_precision), fmt::arg("number_width", number_width),
         fmt::arg("prefix", std::string() + units[pow] + ((s.binary_ && pow) ? "i" : "")),
         fmt::arg("unit", s.unit_),
         fmt::arg("padding", pow         ? ""