GitLab

(cherry picked from commit 7956d4fc)

(cherry picked from commit 3255afcf)

(cherry picked from commit a9094a81)

(cherry picked from commit d2d75b5d)

(cherry picked from commit 97858348)

Thanks to Douglas for pointing out this solution

(cherry picked from commit f44e8c6f)

(cherry picked from commit b6ce28d5)

(cherry picked from commit f1a9cfe0)

(cherry picked from commit c7dc3e5d)

(cherry picked from commit 61f6b32b)

(cherry picked from commit c11845b3)

(cherry picked from commit 7418ec8d)

(cherry picked from commit 49cb3421)

(cherry picked from commit e8357f01)

(cherry picked from commit dce42b36)

(cherry picked from commit f18474b6)

(cherry picked from commit c01d483a)

(cherry picked from commit 29e72c99)

(cherry picked from commit a5df2009)

(cherry picked from commit 8ab6b15c)

(cherry picked from commit 221a10dd)

(cherry picked from commit ab29d869)

(cherry picked from commit 24f1d307)

(cherry picked from commit 371c8f3e)

(cherry picked from commit f31bb7c5)

(cherry picked from commit 3e5dc2c2)

(cherry picked from commit 5549e91a)

(cherry picked from commit 833b6107)

(cherry picked from commit daf36dcb)

(cherry picked from commit b08f575c)

(cherry picked from commit 87c7163a)

(cherry picked from commit 87a5b661)

Avoid reading fld twice in FieldConvert

This MR introduces a small change to avoid loading the fld file twice in FieldConvert. Currently, InputXml imports just the FieldDef to set the expansion, and later InputFld reads the data. However, this is very inefficient, since the whole fld file has to be loaded even if we just want the FieldDef. This MR basically makes InputXml load both the fielddef and the data, and later InputFld just extracts the data.

See merge request !649

(cherry picked from commit 5cfd9d7b)

(cherry picked from commit 7b461fe9)

(cherry picked from commit 9111029d)

(cherry picked from commit deb2b696)

(cherry picked from commit 2333d182)

(cherry picked from commit 1abdbbf9)

(cherry picked from commit 531dd4a6)