Hi Alex, Thanks for pointing out why I should use ‘use’ on those symbols. I have now updated my project.
/Jon > On 7. Feb, 2017, at 17:36, Alexander Burger <a...@software-lab.de> wrote: > > Hi Jon, > >> I have noticed that frequent use of ‘setq’ has been questioned recently on >> this list. In my convConcat.l script there are quite a few setq’s that >> possibly >> could have been coded in a better way. Let me know if you spot some obvious >> candidates. > > Basically there is nothing bad about 'setq'. > > It is just problematic if you use it to modify not-locally-bound symbols (the > term is "free variables") without them being clearly meant as global variables > (by convention with a leading '*'), or from functions meant as local to other > functions (by convention with a leading '_'). > > 'lint' should complain in such cases. > > In case of convConcat.l, you set some globals in the beginning, and local > variables like 'Head', 'Tail', 'Val' and 'C'. Looks all good. > > > If I lint the code, I just get a single warning > > : (lintAll) > -> ((convLink (bnd @Props @After @X @Before))) > > which means "please bind these four symbols" in function 'convLink'. > > This is because 'match' also sets values. You can fix this by calling 'use' on > them: > > (de convLink (F Chunk) > # First checking special case <a name=">foo"> > (use (@Props @After @X @Before) > (if (match '(@Before < a " " n a m e = "\"" > @X "\"" > @After) Chunk) > > ♪♫ Alex > --
