Just after XeF<sub>2</sub> was discovered, D. F. Smith published a short communication [Xenon Difluoride][1] *J. Chem. Phys.* 38, 270 (1963) mentioning the electronic structure of your second drawing.  

Smith conducted infrared spectroscopy of XeF<sub>2</sub> and wrote: 

[![enter image description here][2]][2]

Then C. A. Coulson published [The nature of the bonding in xenon fluorides and related molecules][3] *J. Chem. Soc.*, 1964, 1442-1454, which considered the validity of four models (plus some sub-models) of XeF<sub>2</sub> bonding. 

Coulson adds more resonance structures in addition to the two mentioned by Smith, including a purely ionic F- Xe2+ F- resonance structure.  

For a more recent article on the topic, see [The essential role of charge-shift bonding in hypervalent prototype XeF2][4] *Nature Chemistry* 5, 417–422 (2013). ([official link][5])

Overall, ionic contributions do need to be considered.  


  [1]: https://aip.scitation.org/doi/pdf/10.1063/1.1733476
  [2]: https://i.sstatic.net/BeZLR.png
  [3]: https://pubs.rsc.org/en/content/articlelanding/1964/JR/JR9640001442
  [4]: http://jupiter.chem.uoa.gr/thanost/papers/papers7/NatChem_5(2013)417.pdf
  [5]: https://www.nature.com/articles/nchem.1619