Patches for PR29464, PR29821 & PR29916 - ping*3

[FX Coudert]

Hi Paul,

> These patches are beginning to suffer from bit-rot.  I have already  
> pinged them once.  I propose to commit them to trunk on Saturday  
> morning, unless somebody intervenes with reviews, and to follow  
> with a commit  to 4.2 a week later.

I'm sorry that I'm in no position to do any gfortran-related work  
these days, and really can't review your patches. I did a bit of  
reading anyway, and I think your patches for PR29821 & PR29916 are  
OK. For PR29821, I don't have a definite opinion on what whether we  
should return the result, but I think doing it the way it's done in  
your patch is OK (especially with the comment).

For PR29464, I understand your eagerness to have it commited and also  
agree that reviewing power is not very high these days (I'm partially  
guilty for that) but I frankly don't think a patch 4 days old is  
really something unacceptable :)  Since we've had a few regressions  
on 4.2/4.3 recently (see Joost' recent meta-PR; his code used to  
build perfectly fine, I use it for my day work!), maybe we should  
play a little safer.

FX