[Gretl-devel] Re: off-by-one bug in $system.t1 (and .t2) after var

Am 24.11.20 um 19:46 schrieb Riccardo (Jack) Lucchetti: > On Tue, 24 Nov 2020, Sven Schreiber wrote: > >> Am 21.11.2020 um 11:58 schrieb Sven Schreiber: >>> Am 20.11.2020 um 23:28 schrieb Allin Cottrell: >>>> On Fri, 20 Nov 2020, Allin Cottrell wrote: >>>> >>>> The difference is between equation_system_bundlize() and >>>> gretl_VAR_bundlize(): the former was converting to 1-based t1 and t2, >>>> the latter not. The latter is now fixed in git. >>> >>> OK great! >>> BTW I think some naming harmonization wouldn't hurt, either >>> gretl_system_bundlize (with gretl_) or VAR_bundlize (without >>> gretl_). I >>> guess it could also just be system_bundlize. >> >> I'm trying to understand further the naming convention (or historical >> path dependence). It seems the gretl_ prefix is quite common in >> libgretl, and that actually the various functions starting with >> "equation_system_" are the exception. >> >> My suggestion for harmonization is therefore to add the prefix gretl_ >> also to the systems-related functions. Perhaps "gretl_equation_system_" >> is a bit too long, and actually the "equation" part feels a bit >> redundant IMHO. So what about this replacement rule for libgretl: >> >> *equation_system_* -> *gretl_system_* >> >> OK to apply? > > Of course I see your point and I'm all for anything that improves > code consistency and legibility, but there's this nagging voice I > keep hearing in may head that says "if it ain't broke..." @Sven: Create a new branch, make changes, compile the code and run some/ our test suites ;-) @Jack: This is _exactly_ for what tests are: To verify that things still work as expected after changes. And that's why I frequently plead for much more tests ;-)