Decrement len variable after removing newline, prevents copying the
NUL at the end of a string into the mail. Discovered by dasn.

Move newline printing into outer loop, prevents multiple newlines
f

Decrement len variable after removing newline, prevents copying the
NUL at the end of a string into the mail. Discovered by dasn.

Move newline printing into outer loop, prevents multiple newlines
from appearing if there are multiple comments in a line. Discovered
by okan.

OK okan.

show more ...

Eliminate false positives from <> comments, avoiding mangling
#include <bleh.h> in diffs:
- Only match <> comments from the template. No other text surrounded
by angle brackets is deleted.
- Simplif

Eliminate false positives from <> comments, avoiding mangling
#include <bleh.h> in diffs:
- Only match <> comments from the template. No other text surrounded
by angle brackets is deleted.
- Simplify code by using C strings. We no longer handle binary
data, but I don't think binary data belongs in a bug report anyway.
- Remove dmesg and header comment checks.

"That's a nice approach." deraadt, "seems ok for me" chl.

show more ...

In preparation of eliminating <> comment false positives:
- Move <> comments in template to an array, for easier reference later.
- Rename "comment" variable to avoid shadowing.
- Change some <> comm

In preparation of eliminating <> comment false positives:
- Move <> comments in template to an array, for easier reference later.
- Rename "comment" variable to avoid shadowing.
- Change some <> comments to SENDBUG comments.

OK chl@

show more ...

- Be more careful about pre-existing SIGCHLD handlers (or SIG_IGN) by
temporarily restoring default behaviour. This is not 100% ideal.
But this fixes editor handling in mail... bah, it is really

- Be more careful about pre-existing SIGCHLD handlers (or SIG_IGN) by
temporarily restoring default behaviour. This is not 100% ideal.
But this fixes editor handling in mail... bah, it is really unfortunate
that got broken
- refactor the restoration code as well, to make it simpler
ok ray

show more ...

Don't treat lines in angle brackets <...> as comments in dmesg.

Discovered and tested by deanna.

Remove some unused headers, from Igor Zinovik.

If fork fails with EAGAIN, ignore errno and fail. Looping just
worsens the problem.

cloder@ and deraadt@ agree.

Instead of returning -1 and setting errno to ECHILD if the editor
has a non-zero exit status, just return the exit status on success.
Hard errors still return -1. If the editor is killed, return -1

Instead of returning -1 and setting errno to ECHILD if the editor
has a non-zero exit status, just return the exit status on success.
Hard errors still return -1. If the editor is killed, return -1
and set EINTR.

show more ...

Be more consistent about when dmesg is attached to template.
Specifically, -V.

Simplify some loops.

OK jaredy@ and moritz@.

Restore signals before returning.

Save and restore signals once instead of each time fork fails with
EAGAIN.

Remove extra code that only applies to setuid/setgid programs.

``Slacker! We'll burn your hands if you don't get going!'' deraadt@

Sprinkle const poison, remove unneeded variable.

Use __progname and warn instead of fprintf where appropriate.

De-lint.

Warn if one of the required fields is not filled out. Requested
by millert@.

Don't shadow variables. Enable -Wshadow.

argc does not include __progname after subtracting optind.

Only include newest dmesg.

``Come on, take some risks'' deraadt@.

Attach dmesg to sendbug reports by default. Use -D flag to disable.

OK deraadt@ and millert@.

CC: a copy of the bug report to sender. Requested by Travers Buda.

OK millert@.

Replace some scary pointer code. The new code is O(n^2) for strings
of '&', but gecos fields are usually short and most do not have '&'
so the impact should be minimal. In return we get clearer and

Replace some scary pointer code. The new code is O(n^2) for strings
of '&', but gecos fields are usually short and most do not have '&'
so the impact should be minimal. In return we get clearer and safer
code.

OK moritz@.

show more ...

Remove variable name from prototype.

Simplify gecos parsing code and prevent buffer overflow
when gecos field contains multiple '&' characters with
other characters inbetween.

With and OK ray@