Re: Multiple rules for same target: what is the order of prerequisites in $^ and what does $< expand to?

[Renaud Pacalet]

On 9/27/23 17:01, Mark Piffer wrote:
> Is there a case (common or rare) where the answer to this question has real 
> world consequences?

Yes, I think. If you don't know how $< is expanded you cannot reliably use it in 
your recipe. I strongly suspect that it is expanded to the first prerequisite of the 
rule with a recipe but is this always the case? I couldn't find this in the manual. 
Moreover, what if the rule with a recipe has no prerequisites? Which one, listed in 
another rule, becomes $<?

> regards,
> Mark
>
> > Am 27.09.2023 um 12:17 schrieb Renaud Pacalet <renaud.pacalet@telecom-paris.fr>:
> >
> > Dear all,
> >
> > I read the manual carefully (especially section 4.11 "Multiple Rules for One 
> > Target") and searched the help-make archives but couldn't find a clear description of 
> > how $^ and $< are computed when there are several rules for the same target and only one 
> > has a recipe.
> >
> > According the manual "All the prerequisites mentioned in all the rules are merged into 
> > one list of prerequisites for the target". But when the recipe gets expanded what is 
> > the order of prerequisites in $^ and, more important, what is $<? Does it depend on the 
> > order of rules in the Makefile(s)? On the type of rules (pattern, static pattern, 
> > explicit...)? On which rule has a recipe? Is it deterministic? etc.
> >
> > I did some tests and came up with the following algorithm for $^ and $< 
> > computation:
> >
> > 0. When the list of prerequisites of a rule is added to $^, the order is 
> > preserved.
> >
> > 1. Initialize $^ with the list of prerequisites of the rule with a recipe, 
> > implicit or not.
> >
> > 2. Consider all other applicable explicit or static pattern rules, in their 
> > order of appearance in the Makefile(s), and for each of them append its list of 
> > prerequisites to $^.
> >
> > 3. Ignore all other implicit rules (which seems to contradict 4.11).
> >
> > 4. Remove duplicates from $^, keeping only the first occurrence.
> >
> > 5. Set $< to the first prerequisite in $^.
> >
> > Can someone confirm?
> >
> > Renaud.

--
Renaud Pacalet
Télécom Paris
Campus SophiaTech
450 Route des Chappes, CS 50193
06904 Biot Sophia Antipolis cedex, FRANCE
Tel : +33 (0) 4 9300 8402
Web : http://www.telecom-paris.fr/