Hello. As seen in ${URL}, upstream already checks for execinfo.h and compiles the related code conditionally. There's no need to add the same tricks in a separate patch. Please fix.
The cited commit is incomplete and leads to an improperly defined print_backtrace() function. Review of upstream's tree shows that the incorrect definition still exists. Therefore, the Gentoo patch is still required. That said, the first two hunks listed in the patch appear to be redundant with respect to the code and can probably be removed.
> The cited commit is incomplete and leads to an improperly defined > print_backtrace() function. Review of upstream's tree shows that the incorrect > definition still exists. Therefore, the Gentoo patch is still required. I disagree. Configure script sets BACKTRACE depending on execinfo.h presence [1]. Makefile compiles code with print_backtrace() definition iff BACKTRACE is set [2]. [1]: https://git.kernel.org/cgit/network/connman/connman.git/tree/configure.ac?id=61b0b1377a560d81ac9a4b2348aae6852c7eebae#n186 [2]: https://git.kernel.org/cgit/network/connman/connman.git/tree/Makefile.am?id=61b0b1377a560d81ac9a4b2348aae6852c7eebae#n46
All of those commits were before git said connman-1.32-execinfo-assumptions.patch existed. I don't see any problem with those commits, so it makes no sense for the patch to have been created. So, I dug a bit deeper. Turns out that the patch was originally named net-misc/connman/files/connman-1.16-execinfo-assumptions.patch and for some (now unknown) reason I renamed it to net-misc/connman/files/connman-1.32-execinfo-assumptions.patch (Note to self: DON'T DO THAT). In any case, the rename was done as part of the bump to 1.32, so I thought it was a required patch I made as part of that version bump. Since it does appear that's not the case, I'll remove the patch completely in the next bump.
(In reply to nvinson234 from comment #3) > All of those commits were before git said > connman-1.32-execinfo-assumptions.patch existed. I don't see any problem > with those commits, so it makes no sense for the patch to have been created. > > So, I dug a bit deeper. Turns out that the patch was originally named > net-misc/connman/files/connman-1.16-execinfo-assumptions.patch and for some > (now unknown) reason I renamed it to > net-misc/connman/files/connman-1.32-execinfo-assumptions.patch (Note to > self: DON'T DO THAT). In any case, the rename was done as part of the bump > to 1.32, so I thought it was a required patch I made as part of that version > bump. Since it does appear that's not the case, I'll remove the patch > completely in the next bump. minor correction. I didn't rename the patch, I copied it and then patched the patch slightly so it would work again. Nonetheless, it's going to be excised.
fixed with revision bump to 1.33-r1 (https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=6b88e653f4cc941302f16d1bf7a85412657a478b) patch removed from tree with commit (https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=8126e078c240b5145ad96968019b4b3171b744cf)
