Overview
Request 1189062 accepted
- Add lm_sensors-revert-6b5a19b708.patch which reverts an upstream
patch, thus adding back two explicit pointer type-casts. Code
without them used to generate mere warnings until this year but
causes compiler errors with GCC 14 by default.
If you do not like this, an alternative would be to add the
option -Wno-error=incompatible-pointer-types to compiler flags. Or
fix the types properly.
If the request is OK, please forward it to Factory soon so that we can
switch the default compiler. Thanks.
Request History
jamborm created request
- Add lm_sensors-revert-6b5a19b708.patch which reverts an upstream
patch, thus adding back two explicit pointer type-casts. Code
without them used to generate mere warnings until this year but
causes compiler errors with GCC 14 by default.
If you do not like this, an alternative would be to add the
option -Wno-error=incompatible-pointer-types to compiler flags. Or
fix the types properly.
If the request is OK, please forward it to Factory soon so that we can
switch the default compiler. Thanks.
msmeissn accepted request
ok
@jdelvare: review reminder
What's the error with gcc 14 without the revert?
[ 25s] prog/sensord/rrd.c: In function ‘rrdUpdate’: [ 25s] prog/sensord/rrd.c:458:42: error: passing argument 2 of ‘rrd_update’ from incompatible pointer type [-Wincompatible-pointer-types] [ 25s] 458 | if ((ret = rrd_update(3, (char **) /* WEAK */ argv))) { [ 25s] | ^~~~~~~~~~~~~~~~~~~~~~~~~ [ 25s] | | [ 25s] | char ** [ 25s] /usr/include/rrd.h:179:5: note: expected ‘const char **’ but argument is of type ‘char **’ [ 25s] 179 | const char **); [ 25s] | ^~~~~~~~~~~~~I'm confused. The error above is with the cast, so with the revert. If that results in an error then I don't get the point of the revert. I also see that the prototypes of
rrd_create()andrrd_update()have been modified (in package rrdtool), the second argument was changed fromchar **toconst char **by patchrrdtool-1.8.0-gcc14.patch. This was the right way to solve the problem, and once this is solved on the rrdtool side, no change should be needed on the lm-sensors side. I see that Marcus accepted the submission meanwhile. I believe this is a mistake and will have to be reverted.