Junio C Hamano <gitster@xxxxxxxxx> writes: > All of the above may very well explain why we decided not to use > $(call shellquote) in the past, and it may also explain why it is > safe to start using it again today, but it does not explain why we > want to do so in the first place. > > Having to write > > $(call shellquote,$(VAR)) > > in longhand is much more cumbersome to read, write and merge than > defining a prepackaged VAR_SQ just once and refer it as > > '$(VAR_SQ)' > > everywhere. Having said all that, I would not have minded if this change were more isolated and turned things like CHAINLINTTMP_SQ = $(subst ','\'',$(CHAINLINTTMP)) to use the $(call shellquote) mechanism, if all of us found it easier to read than repeating the same line-noise substitutions for many variables. What was irritating was primarily that the patch went one step further to eliminate the _SQ variant variables, making the resulting sites that use them harder to read and understand.
