matthieustigler / lfe Goto Github PK

CRAN team sent a message complaining about new CRAN CHECK errors, and threatening to remove the package by 2023-02-11 if unaddressed.

Dear maintainer,

Please see the problems shown on https://cran.r-project.org/web/checks/check_results_lfe.html.

Please correct before 2023-02-11 to safely retain your package on CRAN.

For further details see https://www.stats.ox.ac.uk/pub/bdr/sprintf.txt

Problems are:

checking whether package ‘lfe’ can be installed ... WARNING
Found the following significant warnings:
  ./lfe.h:121:19: warning: a function declaration without a prototype is deprecated in all versions of C [-Wstrict-prototypes]
  ./lfe.h:122:13: warning: a function declaration without a prototype is deprecated in all versions of C [-Wstrict-prototypes]
  utils.c:195:19: warning: a function declaration without a prototype is deprecated in all versions of C [-Wstrict-prototypes]
  utils.c:208:13: warning: a function declaration without a prototype is deprecated in all versions of C [-Wstrict-prototypes]
See https://www.r-project.org/nosvn/R.check/r-devel-linux-x86_64-debian-clang/lfe-00install.html for details.

and:

checking compiled code ... WARNING
File ‘lfe/libs/lfe.so’:
  Found ‘__sprintf_chk’, possibly from ‘sprintf’ (C)
    Objects: ‘demean.o’, ‘kaczmarz.o’, ‘utils.o’
Compiled code should not call entry points which might terminate R nor write to stdout/stderr instead of to the console, nor use Fortran I/O nor system RNGs nor [v]sprintf.
See ‘Writing portable packages’ in the ‘Writing R Extensions’ manual.`

It seems that felm reports incorrect standard errors for the first stage when estimating a model with multiple instruments. Standard errors are correct only for the first lhs variable but not the rest.