Update configure.ac (and configure)

[eddelbuettel]

This was triggered by the cri de coeur in this tweet. I got the same note from CRAN and fixed it in half a dozen packages via a simple routine update as suggested. Here I only did one more editing step of replace the (old) comment sequence dnl with the newer (more standard) #. As always, the file to look at is configure.ac with configure being machine-generated and not meant for human consumption.

[MatthieuStigler]

thanks a lot for this Dirk, and my apologies that I didn't update lfe's github address to mine (since I merely adopted it), so that I didn't catch your pull request in time :-( In the meanwhile, Kurt Hornik kindly took care of the update and submitted to CRAN, sorry for the duplicate work!

Comparing your version and Kurt's, I see tiny differences (see below). I am happy to include your further edits if you advice so?

Red is Kurt's (on CRAN now), green is Dirks:

3c3
< AC_PREREQ([2.71])
---
> AC_PREREQ([2.69])
10c10
< AC_CONFIG_MACRO_DIRS([tools])
---
> m4_include([tools/ax_pthread.m4])
59c59,60
< AC_RUN_IFELSE([AC_LANG_SOURCE([[#include <semaphore.h>
---
> AC_RUN_IFELSE([AC_LANG_SOURCE([[
> #include <semaphore.h>

[eddelbuettel]

The difference is self-explanatory. KH used 2.71 (on Debian), I did this on Ubuntu where I have 2.69. Thie rest follows.

Providing current git repo addresses and noticing a pull request is indeed preferred over ignoring it for 10+ days expecially when someone made a public cry for help.

But if it is taken care of (as I also noticed via CRANberries) you can just close this. If you need further help let me know (by tagging @eddelbuettel on this site). I usually respond quickly.

[MatthieuStigler]

you are right, sorry about that!

as a consequence of being on the original github, I cannot close the issue myself I believe?