Re: '... | Avoiding implicit rule recursion for rule '%/dir: %'. | ...'

[WaitronCharm]

Thank you for the patch. I was about to try it out and was following the steps 
here: https://git.savannah.gnu.org/cgit/make.git/tree/README.git

Any idea why the build is stuck at '/usr/bin/msgmerge --update  --lang=ga ga.po 
make.pot'? The 'msgmerge' process just hangs and uses zero CPU. Did stop the 
build once and started again, no difference:

$ make
Making all in lib
make[1]: Entering directory '/home/User/make/lib'
make  all-recursive
make[2]: Entering directory '/home/User/make/lib'
make[3]: Entering directory '/home/User/make/lib'
make[3]: Nothing to be done for 'all-am'.
make[3]: Leaving directory '/home/User/make/lib'
make[2]: Leaving directory '/home/User/make/lib'
make[1]: Leaving directory '/home/User/make/lib'
Making all in po
make[1]: Entering directory '/home/User/make/po'
test ! -f ./make.pot || \
  test -z "be.gmo bg.gmo cs.gmo da.gmo de.gmo es.gmo fi.gmo fr.gmo ga.gmo 
gl.gmo he.gmo hr.gmo id.gmo it.gmo ja.gmo ka.gmo ko.gmo lt.gmo nl.gmo pl.gmo 
pt.gmo pt_BR.gmo ro.gmo ru.gmo sr.gmo sv.gmo tr.gmo uk.gmo vi.gmo zh_CN.gmo 
zh_TW.gmo" || make be.gmo bg.gmo cs.gmo da.gmo de.gmo es.gmo fi.gmo fr.gmo 
ga.gmo gl.gmo he.gmo hr.gmo id.gmo it.gmo ja.gmo ka.gmo ko.gmo lt.gmo nl.gmo 
pl.gmo pt.gmo pt_BR.gmo ro.gmo ru.gmo sr.gmo sv.gmo tr.gmo uk.gmo vi.gmo 
zh_CN.gmo zh_TW.gmo
make[2]: Entering directory '/home/User/make/po'
make[2]: 'be.gmo' is up to date.
make[2]: 'bg.gmo' is up to date.
make[2]: 'cs.gmo' is up to date.
make[2]: 'da.gmo' is up to date.
make[2]: 'de.gmo' is up to date.
make[2]: 'es.gmo' is up to date.
make[2]: 'fi.gmo' is up to date.
make[2]: 'fr.gmo' is up to date.
/usr/bin/msgmerge --update  --lang=ga ga.po make.pot
............................................................

Shall I try any other branch / tag, other than master? This is on Cygwin.

Thank you.

On Friday, November 8th, 2024 at 5:41 AM, Dmitry Goncharov 
<dgoncharov@users.sf.net> wrote:

> On Thu, Nov 7, 2024 at 3:49 AM WaitronCharm via Bug reports and
> discussion for GNU make bug-make@gnu.org wrote:
> 
> > > all: hello.tsk
> > > %.tsk: %.o; $(info hello.tsk)
> > > %.o: %.c %.tsk; $(info hello.c)
> > > Wouldn't this result anyway in a cyclic graph due to '%.tsk'?
> 
> 
> This is a legal makefile. Make drops the circular dependency and
> successfully builds hello.tsk.
> 
> > What I meant in my earlier email is that protecting from infinite recursion 
> > should not happen by forbidding recursion completely.
> 
> 
> As i said earlier, the authors probably had other reasons.
> Given that we mostly deal with files like hello.o, rather than
> hello.o.o, one obvious question is why do you even want to do this?
> 
> > Would this be then (1) check out the code, (2) add a new special target, 
> > (3) instead of 'break' simply continue around line 'Avoiding implicit rule 
> > recursion for rule', (4) compile and see the result?
> 
> 
> (5) add tests, (6) update the manual.
> 
> > What I meant by complications is whether the implementation uses the 
> > pattern rule name as some 'identifier' in any structures and multiples of 
> > the same pattern rule name (per target) would cause mess. (New 'identifier' 
> > could change to the combination of pattern rule name and instance no. of 
> > applying that rule).
> 
> 
> No.
> 
> > I am not familiar with the 'make' code and did not look into it yet. Want 
> > to just estimate how much work would this be.
> 
> 
> i posted a diff below that you can use. This diff achieves the
> behavior you desire. You can see, the change is small. The hard part
> here, is not the size of the changeset. The hard part is to convince
> yourself that, with this change, make still does the minimal necessary
> amount of implicit searches.
> 
> regards, Dmitry
> 
> diff --git a/src/file.c b/src/file.c index 6f816c8a..c8017e34 100644
> --- a/src/file.c +++ b/src/file.c @@ -911,6 +911,10 @@ snap_deps
> (void) d2->wait_here = 1; } + f = lookup_file
> 
> (".ALLOW_RECURSIVE_CHAINS"); + if (f != 0 && f->is_target) +
> 
> allow_recursive_chains = 1; + { struct dep *prereqs =
> expand_extra_prereqs (lookup_variable
> (STRING_SIZE_TUPLE(".EXTRA_PREREQS"))); diff --git a/src/implicit.c
> b/src/implicit.c index 134e3fef..2bd2b90e 100644 --- a/src/implicit.c
> +++ b/src/implicit.c @@ -297,7 +297,7 @@ pattern_search (struct file
> file, int archive, / If this rule is in use by a parent
> pattern_search, don't use it here. */ - if (rule->in_use) + if
> 
> (rule->in_use && !allow_recursive_chains) { DBS (DB_IMPLICIT,
> 
> (_("Avoiding implicit rule recursion for rule '%s'.\n"), diff --git
> a/src/main.c b/src/main.c index 78084d09..5333df5f 100644 ---
> a/src/main.c +++ b/src/main.c @@ -193,6 +193,10 @@ static int
> old_builtin_variables_flag; int export_all_variables = 0; +/* Nonzero
> means a chain of implicit rules can have a rule multiple times. / +
> +int allow_recursive_chains; + / Nonzero means keep going even if
> remaking some file fails (-k). */ int keep_going_flag; diff --git
> a/src/makeint.h b/src/makeint.h index 61c78229..27a269ed 100644 ---
> a/src/makeint.h +++ b/src/makeint.h @@ -736,6 +736,7 @@ extern int
> print_version_flag, check_symlink_flag, posix_pedantic; extern int
> not_parallel, second_expansion, clock_skew_detected; extern int
> rebuilding_makefiles, one_shell, output_sync, verify_flag; extern int
> export_all_variables; +extern int allow_recursive_chains; extern
> unsigned long command_count; extern const char *default_shell;