Freeciv

- INFO_POPULATION is confusing. City population (at least 10000 people) as shown in the city dialog is different thing from the city size.
- Do not have 'default:' -case for the switch here. We don't want it to silence compiler warning about missing handling of an enum value here if someone adds a new one

· Got it, I will change it to INFO_CITY_SIZE or maybe INFO_CITIZENS, whichever you might prefer.
· On the second point, ... the `default:` case is intended to catch the missing label for an added INFO_xxx enum. Is a soft warning or a hard exit preferred?

John Robertson wrote:

· On the second point, ... the `default:` case is intended to catch the missing label for an added INFO_xxx enum. Is a soft warning or a hard exit preferred?

Catching it already at compile time (as an hard error with --enable-debug as is normal in development time) is preferred over maybe running on it on run time.

The problem is that the list of labels is built at run-time, not compile time; both in the unchanged current code and the proposed patch. Thus I don't think that there can be a compile time check to assert that the number of elements match between the enum and the label list.

Additionally, the runtime building of the label list in the proposed patch is to ensure that the labels are in the same order as the enums for subsequent algorithmic convenience. The current code does not have an order guarantee.

In addition to removing the 'default' case, you need to have 'i' to have correct enum type instead of being an int. Then the compiler will see if there's no 'case' for some value of 'i'. That's how all the other enum 'switch' cases in freeciv code work, and what CodingStyle document tells you to do.

I get various toolset deprecation errors in the S2_6_2 build. E.g.:
· `QT_DEPRECATED_X("Use default constructor instead") Q_DECL_CONSTEXPR inline QFlags(Zero) noexcept : i(0) {}`
· ` error: 'int QFontMetrics::width(const QString&, int) const' is deprecated: Use QFontMetrics::horizontalAdvance [-Werror=deprecated-declarations]`

I can work through an independent patch of these toolset updates, then produce an unencumbered city_panel patch, ...

or just abandon the city_panel patch for S2_6_2.

What would you recommend?

Your S3_0 patch applied to S2_6 as is.

Qt deprecation warnings are Bug #896335.

Great, I won't worry about it then.

John Robertson wrote:

or just abandon the city_panel patch for S2_6_2.

For the record: S2_6_2 was a branch for emergency 2.6.2.1 release. It won't continue from there, unless we have to do another emergency release before 2.6.3 is ready, and even then only those critical fixes would be cherry-picked to it.
S2_6 is the freeciv-2.6 branch.