Would timevar be accepted in gnulib?

[Akim Demaille]

Hi!

Bison has adopted, many years ago, timevars, a set of tools that
allow self-profiling.  It does not aim at replacing real profiling
tools, but rather to report in a concise way the cost of various
phases in a program.  For instance with Bison:

Execution times (seconds)
 LALR(1)               :   0.01 ( 3%) usr   0.00 ( 0%) sys   0.00 ( 0%) wall
 IELR(1) Phase 4       :   0.01 ( 3%) usr   0.00 ( 0%) sys   0.00 ( 0%) wall
 outputing parser      :   0.01 ( 3%) usr   0.00 ( 0%) sys   0.00 ( 0%) wall
 running m4            :   0.31 (89%) usr   0.00 ( 0%) sys   0.00 ( 0%) wall
 freeing               :   0.01 ( 3%) usr   0.00 ( 0%) sys   0.00 ( 0%) wall
 TOTAL                 :   0.35             0.01             0.00

timevar was stolen from GCC.  When it was introduced in Bison,
the times were significantly larger, there were less zeros in
the table :)  (real zeros are not displayed, there are more
phases which are measured).

Would there be some interest for this in gnulib?  Of course I
guess there’s work to do for it to be accepted (doc, style, etc.),
but before going that way, I’d like to know where it makes sense.

Here’s how it’s used in Bison:

  timevar_report = trace_flag & trace_time;
  init_timevar ();
  timevar_start (TV_TOTAL);

  timevar_push (TV_READER);
  reader ();
  timevar_pop (TV_READER);

  timevar_push (TV_REDUCE);
  reduce_grammar ();
  timevar_pop (TV_REDUCE);

  timevar_push (TV_SETS);
  derives_compute ();
  nullable_compute ();
  timevar_pop (TV_SETS);

  […]

(Note that phases can be nested: TV_TOTAL includes all the others).

Attached, the current state of the timevar files in Bison.

timevar.c
Description: Binary data

timevar.def
Description: Binary data

timevar.h
Description: Binary data

timevar.m4
Description: Binary data