1*22619282SSimon J. Gerraty# $NetBSD: directive-for-errors.mk,v 1.13 2024/07/06 10:14:35 rillig Exp $ 206b9b3e0SSimon J. Gerraty# 306b9b3e0SSimon J. Gerraty# Tests for error handling in .for loops. 406b9b3e0SSimon J. Gerraty 5c1d01b5fSSimon J. Gerraty 606b9b3e0SSimon J. Gerraty# A .for directive must be followed by whitespace, everything else results 706b9b3e0SSimon J. Gerraty# in a parse error. 8c1d01b5fSSimon J. Gerraty# expect+1: Unknown directive "fori" 906b9b3e0SSimon J. Gerraty.fori in 1 2 3 10*22619282SSimon J. Gerraty# expect+1: warning: <> 11c1d01b5fSSimon J. Gerraty. warning <${i}> 12*22619282SSimon J. Gerraty# expect+1: for-less endfor 1306b9b3e0SSimon J. Gerraty.endfor 14c1d01b5fSSimon J. Gerraty 1506b9b3e0SSimon J. Gerraty 1606b9b3e0SSimon J. Gerraty# A slash is not whitespace, therefore this is not parsed as a .for loop. 1706b9b3e0SSimon J. Gerraty# 1806b9b3e0SSimon J. Gerraty# XXX: The error message is misleading though. As of 2020-12-31, it says 19c1d01b5fSSimon J. Gerraty# 'Unknown directive "for"', but that directive is actually known. This is 2006b9b3e0SSimon J. Gerraty# because ForEval does not detect the .for loop as such, so parsing 21b0c40a00SSimon J. Gerraty# continues in ParseLine > ParseDependencyLine > ParseDependency > 22b0c40a00SSimon J. Gerraty# ParseDependencyTargets > ParseErrorNoDependency, and there the directive 2306b9b3e0SSimon J. Gerraty# name is parsed a bit differently. 24c1d01b5fSSimon J. Gerraty# expect+1: Unknown directive "for" 2506b9b3e0SSimon J. Gerraty.for/i in 1 2 3 26*22619282SSimon J. Gerraty# expect+1: warning: <> 27c1d01b5fSSimon J. Gerraty. warning <${i}> 28*22619282SSimon J. Gerraty# expect+1: for-less endfor 2906b9b3e0SSimon J. Gerraty.endfor 3006b9b3e0SSimon J. Gerraty 31c1d01b5fSSimon J. Gerraty 32c1d01b5fSSimon J. Gerraty# Before for.c 1.173 from 2023-05-08, the variable name could be an arbitrary 33c1d01b5fSSimon J. Gerraty# word, it only needed to be separated by whitespace. Even '$' and '\' were 34c1d01b5fSSimon J. Gerraty# valid variable names, which was not useful in practice. 3506b9b3e0SSimon J. Gerraty# 36c1d01b5fSSimon J. Gerraty# The '$$' was not replaced with the values '1' or '3' from the .for loop, 37c1d01b5fSSimon J. Gerraty# instead it was kept as-is, and when the .info directive expanded its 38d5e0a182SSimon J. Gerraty# argument, each '$$' got replaced with a single '$'. The "long 39c1d01b5fSSimon J. Gerraty# expression" ${$} got replaced though, even though this would be a parse 40c1d01b5fSSimon J. Gerraty# error everywhere outside a .for loop. 4106b9b3e0SSimon J. Gerraty${:U\$}= dollar # see whether the "variable" '$' is local 4206b9b3e0SSimon J. Gerraty${:U\\}= backslash # see whether the "variable" '\' is local 43c1d01b5fSSimon J. Gerraty# expect+1: invalid character '$' in .for loop variable name 448d5c8e21SSimon J. Gerraty.for a b $ \ in 1 2 3 4 4506b9b3e0SSimon J. Gerraty. info Dollar $$ ${$} $($) and backslash $\ ${\} $(\). 4606b9b3e0SSimon J. Gerraty.endfor 4706b9b3e0SSimon J. Gerraty 4806b9b3e0SSimon J. Gerraty# If there are no variables, there is no point in expanding the .for loop 49c1d01b5fSSimon J. Gerraty# since this would end up in an endless loop, consuming 0 of the 3 values in 50c1d01b5fSSimon J. Gerraty# each iteration. 51c1d01b5fSSimon J. Gerraty# expect+1: no iteration variables in for 5206b9b3e0SSimon J. Gerraty.for in 1 2 3 5306b9b3e0SSimon J. Gerraty# XXX: This should not be reached. It should be skipped, as already done 5406b9b3e0SSimon J. Gerraty# when the number of values is not a multiple of the number of variables, 5506b9b3e0SSimon J. Gerraty# see below. 5606b9b3e0SSimon J. Gerraty. warning Should not be reached. 5706b9b3e0SSimon J. Gerraty.endfor 5806b9b3e0SSimon J. Gerraty 59c1d01b5fSSimon J. Gerraty 6006b9b3e0SSimon J. Gerraty# There are 3 variables and 5 values. These 5 values cannot be split evenly 6106b9b3e0SSimon J. Gerraty# among the variables, therefore the loop is not expanded at all, it is 62c1d01b5fSSimon J. Gerraty# skipped instead. 63c1d01b5fSSimon J. Gerraty# expect+1: Wrong number of words (5) in .for substitution list with 3 variables 6406b9b3e0SSimon J. Gerraty.for a b c in 1 2 3 4 5 6506b9b3e0SSimon J. Gerraty. warning Should not be reached. 6606b9b3e0SSimon J. Gerraty.endfor 6706b9b3e0SSimon J. Gerraty 68c1d01b5fSSimon J. Gerraty 6906b9b3e0SSimon J. Gerraty# The list of values after the 'in' may be empty, no matter if this emptiness 70d5e0a182SSimon J. Gerraty# comes from an expanded expression or from a syntactically empty line. 7106b9b3e0SSimon J. Gerraty.for i in 7206b9b3e0SSimon J. Gerraty. info Would be reached if there were items to loop over. 7306b9b3e0SSimon J. Gerraty.endfor 7406b9b3e0SSimon J. Gerraty 75c1d01b5fSSimon J. Gerraty 7606b9b3e0SSimon J. Gerraty# A missing 'in' should parse the .for loop but skip the body. 77c1d01b5fSSimon J. Gerraty# expect+1: missing `in' in for 78c1d01b5fSSimon J. Gerraty.for i over k 7906b9b3e0SSimon J. Gerraty# XXX: As of 2020-12-31, this line is reached once. 8006b9b3e0SSimon J. Gerraty. warning Should not be reached. 8106b9b3e0SSimon J. Gerraty.endfor 8206b9b3e0SSimon J. Gerraty 83c1d01b5fSSimon J. Gerraty 8406b9b3e0SSimon J. Gerraty# A malformed modifier should be detected and skip the body of the loop. 8506b9b3e0SSimon J. Gerraty# 8606b9b3e0SSimon J. Gerraty# XXX: As of 2020-12-31, Var_Subst doesn't report any errors, therefore 8706b9b3e0SSimon J. Gerraty# the loop body is expanded as if no error had happened. 88*22619282SSimon J. Gerraty# expect+1: while evaluating "${:U3:Z} 4" with value "3": Unknown modifier "Z" 8906b9b3e0SSimon J. Gerraty.for i in 1 2 ${:U3:Z} 4 90*22619282SSimon J. Gerraty# expect+3: warning: Should not be reached. 91*22619282SSimon J. Gerraty# expect+2: warning: Should not be reached. 92*22619282SSimon J. Gerraty# expect+1: warning: Should not be reached. 9306b9b3e0SSimon J. Gerraty. warning Should not be reached. 9406b9b3e0SSimon J. Gerraty.endfor 95