[Bug] Field cache ignores orderByIncludeColumn and can reuse the wrong column order

[Aias00]

Fesod version

Current main at 5a9a16b.

Description

The cached FieldCache depends on WriteHolder#orderByIncludeColumn, but ClassUtils.FieldCacheKey contains only the class and include/exclude collections. Two sheets using the same class and include list but different ordering settings can therefore share the wrong cached result.

Location

• fesod-sheet/src/main/java/org/apache/fesod/sheet/util/ClassUtils.java:278-291
• fesod-sheet/src/main/java/org/apache/fesod/sheet/util/ClassUtils.java:379-431
• fesod-sheet/src/main/java/org/apache/fesod/sheet/util/ClassUtils.java:556-573

Reproduction

Use the same model and include list column4, column2, column3 in one writer:

1. First sheet: orderByIncludeColumn(false)
2. Second sheet: orderByIncludeColumn(true)

Observed second-sheet order after cache reuse:

column2, column3, column4

A fresh writer using only orderByIncludeColumn(true) produces the requested order:

column4, column2, column3

Expected behavior

Each sheet should honor its own orderByIncludeColumn setting regardless of a prior cache entry.

Suggested fix

Include orderByIncludeColumn in FieldCacheKey and add a multi-sheet regression test that changes only this setting.

Related existing work

• [Bug] ClassUtils removes the wrong indexFieldMap entry when excluding a column #976/fix: remove indexFieldMap entry by field key, not running counter #977 cover removal of the wrong explicit-index entry.
• [Bug] Column include/exclude drops or keeps the wrong columns for field names like xRealIp #1129/fix: filter columns by java field name for camel case field names #1130 cover camel-case field-name filtering.
• [Enhancement] Split ClassUtils into per-concern classes and de-duplicate the cache-location switch #989/refactor: split ClassUtils into metadata resolvers #1099 refactor ClassUtils but retain the same cache-key semantics.

None covers this cache collision.

Are you willing to submit a PR?

Yes.