When using clang3.3 on mavericks with libc++, after the
first std::getline-call encountered eof on an ifstream,
all following std::getline calls on the same ifstream will
also get eof, even if ifstream::clear has been called and
the file has been appended in the meantime.
Seeking to the current position after a clear before trying
to read the line fixes this behavior.
This is just a temporary workaround...
* origin/fastpath:
Fix record coercion tolerance of optional fields.
Add NEWS about incompatible local.bro changes, addresses BIT-1047.
Fix minor formatting problem in NEWS.
There were cases where coercing a record value with an uninitialized
field could cause a null pointer dereference even though the field
can validly be unset since it has &optional.
So script parsing is only ~2x slower rather than 20x. Turns out cloning
Vals is particularly slow. Changed to just get a string description of
the Val for initial value and redef tracking.
Snapshotting the work in this branch. I'll merge it again later as we
get closer to the release.
* origin/topic/dnthayer/doc-changes-for-2.2: (29 commits)
Add README files for base/protocols
Fix incorrect uses of reST directives
Fix typos and formatting in the BiFs docs
Fix typos and formatting in the base/utils docs
Fix typos and formatting in the other protocol docs
Fix typos and formatting in the ssl protocol docs
Fix typos and formatting in the http protocol docs
Fix typos and formatting in the ftp protocol docs
Fix typos and formatting in the dns protocol docs
Fix typos and formatting in the dhcp protocol docs
Adjust line numbers to match changes in conn/main.bro
Fix typos and formatting in the conn protocol docs
Update FreeBSD install instructions
Improvements to file analysis docs
Add README files for most Bro frameworks
Fix typos and formatting in various other framework docs
Fix typos and formatting in the software framework docs
Fix typos and formatting in the sumstats docs
Fix typos and formatting in the packet filter docs
Fix typos and formatting in the logging framework docs
...
Sorry for this - I noticed that I named this option quite unfortunately
while writing the documentation.
The patch also removes the dbname configuration option from the sqlite
input reader - it was not used there at all anymore (and I did not notice
that).
Mem leaks and a null pointer deref (not actually expected to occur
unless already in one of the the odd states that warrants an internal
warning/error).
Also dead code.
* topic/robin/internal-errors-merge:
Reverting one of the my internal errors tweaks.
Renaming InternalAnalyzerError to AnalyzerError.
Pass over the InternalError() changes.
Review usage of Reporter::InternalError, addresses BIT-1045.
Fixed reference to wrong field name.
Added documentation of a function arg.
Added a couple references to other parts of the documentation.
Explained how not specifying extraction filename results in automatic
filename generation.
Several other minor clarifications.
This is used to display short summaries for things based on the first
sentence in the comments for it, but wouldn't work well when e.g. a
filename is used there.
Replaced some with InternalWarning or InternalAnalyzerError, the later
being a new method which signals the analyzer to not process further
input. Some usages I just removed if they didn't make sense or clearly
couldn't happen. Also did some minor refactors of related code while
reviewing/exploring ways to get rid of InternalError usages.
Also, for TCP content file write failures there's a new event:
"contents_file_write_failure".
It was getting too bloated and allocated memory in ways that were
difficult to understand how to manage. Separated out primarily in to
new find_file() and open_file()/open_package() functions.
Also renamed other util functions for path-related things.
Doesn't generate any docs, but it's hooked in to all places needed to
gather the necessary stuff w/ significantly less coupling than before.
The gathering now always occurs unconditionally to make documentation
available at runtime and a command line switch (-X) only toggles whether
to output docs to disk (reST format).
Should also improve the treatment of type name aliasing which wasn't a
big problem in practice before, but I think it's more correct now:
there's now a distinct BroType for each alias, but extensible types
(record/enum) will automatically update the types for aliases on redef.
Other misc refactoring of note:
- Removed a redundant/unused way of declaring event types.
- Changed type serialization format/process to preserve type name
information and remove compatibility code (since broccoli will
have be updated anyway).
Using 'char' versus 'int' to store the return value is a problem since
'char' varies in signedness depending on platform/implementation
(works some places, others not).
The code that looks like it was intended to make it apply to a vector
operand couldn't be reached and making it reachable would still require
changes to other methods for it to work so just removing for now.
Integral/floating-point division/modulo by zero in C++ is undefined
behavior, so to prevent such cases in a script from crashing Bro,
they're now reported as an error (with script location information) and
the event handler in which it occurred returns immediately.
