On Wed, Apr 05, 2023 at 09:02:10AM -0700, Tyler Retzlaff wrote: > On Wed, Apr 05, 2023 at 11:56:05AM +0100, Bruce Richardson wrote: > > On Tue, Apr 04, 2023 at 01:07:27PM -0700, Tyler Retzlaff wrote: > > > Windows does not support versioned symbols. Fortunately Windows also > > > doesn't have an exported stable ABI. > > > > > > Export rte_tel_data_add_array_int -> rte_tel_data_add_array_int_24 > > > and rte_tel_data_add_dict_int -> rte_tel_data_add_dict_int_v24 > > > functions. > > > > > > Windows does have a way to achieve similar versioning for symbols but it > > > is not a simple #define so it will be done as a work package later. > > > > > > Signed-off-by: Tyler Retzlaff <roret...@linux.microsoft.com> > > > > Does this require a change in telemetry itself? Can it be done via the > > header file with the versioning macros in it, so it would apply to any > > other versioned functions we have in DPDK? > > i didn't spend a lot of time thinking if the existing macros could be > made to expand in the way needed. there is a way of doing versioning on > windows but it is foreign to how this symbol versioning scheme works so > i plan to investigate it separately after i get unit tests running. > > for now i know what i'm doing is ugly but i need to get protection of > unit tests so i'm doing minimal changes to get to that point. if you're > not comfortable with this going in on a temporary basis i can remove it > from this series and we can work on it as a separated patch set. > > my bar is pretty low here, as long as it doesn't break any existing > linux/gcc/clang etc ok, if msvc is not right i'll take a second pass > and design each stop-gap properly. it already doesn't work so things > aren't made worse. > > let me know if i need to carve this out of the series. > It's not that ugly. :-) If no other clear solution is apparent, I can certainly live with this.
/Bruce
