Unsorted enums in message specifications do not result in correct
min/max values per the protobuf spec for cpp code:
https://developers.google.com/protocol-buffers/docs/reference/cpp-generated#enum
The google cpp compiler uses the min/max _value_ instead of the position
of the enum entry in the message spec, so this patch updates nanopb to
do the same.
Example input:
```protobuf
syntax = "proto3";
enum Language {
UNKNOWN = 0;
ENGLISH_EN_GB = 12;
ENGLISH_EN_US = 1;
FRENCH_FR_FR = 2;
ITALIAN_IT_IT = 3;
GERMAN_DE_DE = 4;
SPANISH_ES_AR = 13;
SPANISH_ES_ES = 5;
SPANISH_ES_MX = 14;
SWEDISH_SV_SE = 6;
DUTCH_NL_NL = 7;
KOREAN_KO_KR = 8;
JAPANESE_JA_JP = 9;
CHINESE_SIMPLIFIED_ZH_CN = 10;
CHINESE_TRADITIONAL_ZH_TW = 11;
}
```
This would previously result in:
```c
#define _Language_MIN Language_UNKNOWN
#define _Language_MAX Language_CHINESE_TRADITIONAL_ZH_TW
#define _Language_ARRAYSIZE ((Language)(Language_CHINESE_TRADITIONAL_ZH_TW+1))
```
After this patch,
```c
#define _Language_MIN Language_UNKNOWN
#define _Language_MAX Language_SPANISH_ES_MX
#define _Language_ARRAYSIZE ((Language)(Language_SPANISH_ES_MX+1))
```
Add a test `enum_minmax` to cover this case.
