add missing headers for some system functions #237
Conversation
DerDakon
requested review from
josuah,
leahneukirchen,
mbhangui and
schmonz
as code owners
|
Is there a particular compiler warning level you're using to make these pop out? Something we could include in one of the autobuilds? |
schmonz
left a comment
schmonz
left a comment
There was a problem hiding this comment.
In a different situation I might want to block this on having our autobuilds auto-detect popular merge conflicts. Right now it's more important to me that we start moving again. Thank you for the impetus.
| #include <sys/socket.h> | ||
| #include <netinet/in.h> | ||
| #include <arpa/inet.h> | ||
| #include <unistd.h> |
There was a problem hiding this comment.
In general this change is of course correct and needed, and as usual I'm mildly worried only about the timing of the change in programs which many users are likely to still be heavily patching: this one and smtpd, mainly.
@schmonz gonna schmonz ;-) (and seriously, it's nice to be back here thinking about notqmail, so thank you) -- but I would love to be able to feel that this sort of thing is getting worried about automatically. Then I could find new kinds of things to be mildly annoying about. @josuah, how can we help finish getting the patch-merge-checker into our regular autobuilds?
There was a problem hiding this comment.
As this a bit older: no, I don't think so. At least I can't provoke a warning on my system, most likely because (at least today, maybe not when these patches were originally done) these headers are in one or another indirect way automatically included. So for now I would call it "manual code audit", even if it was not true back then it would be true if I had to do them again.
This has been split out of #65. The commit about headers for
umask()has been omitted asmode_tis not used, so<sys/types.h>isn't actually needed anywhere for this.