bug fixes prompted by recent Coverity scan

[Paul Eggert]

Two things about the recently-proposed bug fixes prompted by the Coverity scan.

First, Coverity itself does the scan for me. They use the glibc source to debug 
their scanner, and they send me their reports as a service to the community. I 
don't control the parameters they use. I found the latest set of four fixes by 
inspecting only the changes in reports from previous glibc. Most of the changes 
were false alarms (this is typical) and I generally left the code alone for the 
false alarms.

Second, there's a lot of bureaucracy involved in getting obvious bugs like these 
fixed. The current distracting thread about ChangeLog style is a symptom of the 
problem. In the old days I would have simply installed the fixes and moved on, 
but now I'm being asked for testcases and bugzilla references and ChangeLog 
style reformattting and whatnot and I don't want to upset anybody's applecart 
nor do I want to bother with all that paperwork (hey, I'm just a volunteer!) so 
I've done nothing to follow up, which is not good.

I can't help contrasting the somewhat offputting responses to my four glibc 
Coverity scan patches with the quite-friendly response I got from the gettext 
project when I filed the gettext-related patch upstream. Here's the entire 
interaction:

I ran "git send-email --to=bug-gettext@gnu.org".

Within a few hours, Daiki Ueno replied "Thanks, applied."

It's a pleasure helping to fix gettext bugs. I wish I could say the same thing 
about glibc.  I realize glibc is a bigger project and needs more bureaucracy 
yadda yadda yadda, but come on guys, we've gone off the deep end.

Anyway, to work through this little problem could you please install the patches 
for BZ#18868, BZ#18871, BZ#18872, and BZ#18873, whenever you have the time? The 
fixes themselves all have reasonable consensus, I think. Please feel free to dot 
all the Is and cross all the Ts that I missed. Thanks.