On 2022-10-24 10:30 AM, NRK wrote:
On Mon, Oct 24, 2022 at 01:10:29PM +0300, Santtu Lakkala wrote:
The dynmaic[sic] version incorrectly passes sizeof(buf), where buf is
char
*, as the size of buffer in the "happy case" leading to unnecessary
hits to
the dynamic path.
Ah yes, the classic. Att
On 2022-10-24 10:01 AM, NRK wrote:
On Mon, Oct 24, 2022 at 09:21:37AM +, Andy Gozas wrote:
• XmbLookupString leaves the ksym unchanged if not filled and
XLookupString
[probably] sets it to NoSymbol (that's what XLookupKeysym does, but
whether
or not XLookupString shares this behavior is unc
On Mon, Oct 24, 2022 at 01:10:29PM +0300, Santtu Lakkala wrote:
> The dynmaic[sic] version incorrectly passes sizeof(buf), where buf is char
> *, as the size of buffer in the "happy case" leading to unnecessary hits to
> the dynamic path.
Ah yes, the classic. Attached ammended version of the dynma
On 24.10.2022 13.01, NRK wrote:> But which approach to take is the
maintainer's call, not mine.
I've attched both fixed-size and dynamic-allocation patch (but
simplified without goto).
The dynmaic[sic] version incorrectly passes sizeof(buf), where buf is
char *, as the size of buffer in the "h
On Mon, Oct 24, 2022 at 09:21:37AM +, Andy Gozas wrote:
> • XmbLookupString leaves the ksym unchanged if not filled and XLookupString
> [probably] sets it to NoSymbol (that's what XLookupKeysym does, but whether
> or not XLookupString shares this behavior is unclear [1]), so we can just
> set i
On 2022-10-24 12:35 AM, NRK wrote:
On Sun, Oct 23, 2022 at 04:18:42PM +, Andy Gozas wrote:
> St relies on an incorrect assumption of how XmbLookupString function
> behaves.
Looking at the XmbLookupString manpage [0] reveals more trouble. It
seems
that `ksym` might be used uninitalized as
