Problem : Out-of-bounds read @ 9514.
Diagnostic : Real issue.
Rationale : PFD_NOTSPECIAL (253) is defined as the maximum not-special
value a prefix can have. But stack (and other) arrays are
defined as having MAXWLEN (250) items.
Resolution : Define MAXWLEN = 254.
Problem : Out-of-bounds read @ 2213.
Diagnostic : Real issue.
Rationale : Error occurs if cmap == ARRAY_SIZE(prt_ps_mbfonts),
but code takes the `if (prt_out_mbyte)` branch. That's it,
if a matching encoding is found but not a matching charset.
In that case, the first matching encoding is used.
Resolution : Remember the value of cmap for the first matching encoding.
Reset cmap to that value if first matching encoding is
going to be used.
It turns out the FreeBSD 10 VM has a symlink for the home directory to
/usr/home. Unfortunately, this breaks the test as arg[0] may not have
the symlink resolved, but the path returned from the exe() call will.
As a result, the comparison fails, even though the result is correct.
Let's fix this by running the absolute path through exe() too, and then
comparing the results.
It turns out that `file(DOWNLOAD ...)` is not very user friendly with
it's error message, and only supports MD5 on v2.8.10 of CMake (the
default for Ubuntu 12.04). If CMake is built without SSL support,
users are left hanging with a message that the hashes don't match.
It turns out that `file(SHA1 ...)` exists in v2.8.10, and we
can use that to compute the hash ourselves. So this splits the hash
checking into a separate step, where we can provide some additional
advice if the SHA1 is the hash for an empty file. Additionally, it also
allows us to drop the MD5 hashes and maintain only SHA1 hashes for our
dependencies.
refs #1045#1051
This was enabled by default a while ago (#1051), and has apparently not
created any issues. The amount of actual code related to it is tiny, so
it has been removed.
The systemlist test currently calls the `echo` command which can potentially
complete before being interrupted, causing random test failures.
Use `yes | xargs` instead. A `yes` invocation that is not piped through `xargs`
can produce a huge amount of lines in a very short time, leading memory
starvation when the result is being converted into a list. `xargs` ensures only
one line of output will be produced while allowing interrupt to be tested.
- Factor out main_msg() in favor of mch_msg() and manual indentation, as to
provide a much closer representation to the actual output of '--help'.
'gcc -E' reveals that main_msg() only consists of 3 printf calls
anyways.
- Factor out for loop used for printing top part of usage text; just
print the text normally.
usage() text:
- Don't print the version; that's what '--version' is for.
- Be consistent about nomenclature, e.g. '<arg>' denotes required
argument, '-h | --help' denotes '-h' and '--help' are equivalent, etc.
- Change some instances of vim{,rc,info} to nvim
Problem : Unitialized scalar variable @ 3239.
Diagnostic : Harmless issue.
Rationale : It's true pos.coladd is not initialized when calling
searchit(). But that's no problem, as coladd is only set in
that function.
Resolution : Initialize variable to 0.
Problem : Argument cannot be negative @ 1165.
Diagnostic : Real issue.
Rationale : len can be assigned a negative value @ 1162;
len is passed as an unsigned argument @ 1165.
Resolution : Refactor variable's types:
- Use ftello instead of ftell to avoid using long.
- Assert ftello result is safely convertible to size_t.
- Introduce variable read_size to avoid using i (int).
Diagnostic : False positive.
Rationale : Coverity thinks we are forgetting to add more char to hold
NULL, but it's not taking into account that two chars from
cntxformat will no be present in the result. In fact, we
can even allocate one byte less than currently done.
Resolution : Add explanatory comment and allocate one less byte.
Marked as "Intentional" at coverity's database.
Problem : Negative array index read @ 909.
Diagnostic : False positive.
Rationale : Suggested error path assigns a negative value to idx at
line 836 (`idx = find_command(ca.cmdchar);`). That's
impossible, as `ca.cmdchar` is set to Ctrl_BSL just two
lines above, so we know that value will be in the table.
Resolution : Assert idx >= 0.
The issue is that libuv decides to use ${PREFIX}/lib64 as the lib
directory instead of ${PREFIX}/lib. Since we're just installing it to
provide access to the static library, let's just force the libdir to be
${PREFIX}/lib.
It appears that commit 393c1c59a27591d705648919b2d7fb921cba37bc (unix:
set non-block mode in uv_{pipe,tcp,udp}_open) has broken Neovim's
drawing under OSX. Let's revert to 1.2.0 until we can figure out what
is happening and get it fixed.
