database compactor tool
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
lsb |
In Progress
|
Medium
|
Unassigned | ||
Mandriva |
In Progress
|
Medium
|
Bug Description
This is listed in the devchk request list because I see it from devchk, but the
tool is probably fine as standalone.
It would be nice to have a way to identify when former devchk runs have
prompted us to push data which makes arch-specific entries that aren't needed.
Here's one example, an anoymous enum from ftw.h:
/* Flags for fourth argument of `nftw'.*/
#if __LSB_VERSION__ >= 12
enum {
FTW_PHYS = 1,
FTW_MOUNT = 2,
FTW_CHDIR = 4,
FTW_DEPTH = 8
};
#endif
we've ended up with this in devchk, which means there are identical records for
each arch for these values:
#if defined __s390x__
CheckEnum(
CheckEnum(
CheckEnum(
CheckEnum(
#elif defined __x86_64__
CheckEnum(
CheckEnum(
CheckEnum(
CheckEnum(
#elif defined __s390__ && !defined __s390x__
CheckEnum(
CheckEnum(
CheckEnum(
CheckEnum(
#elif defined __powerpc64__
CheckEnum(
CheckEnum(
CheckEnum(
CheckEnum(
#elif defined __powerpc__ && !defined __powerpc64__
CheckEnum(
CheckEnum(
CheckEnum(
CheckEnum(
#elif defined __ia64__
CheckEnum(
CheckEnum(
CheckEnum(
CheckEnum(
#elif defined __i386__
CheckEnum(
CheckEnum(
CheckEnum(
CheckEnum(
#else
there is no arch difference here. Seems we should be able to count up cases
where this happens and reduce those down to a single generic entry.
Changed in mandriva: | |
importance: | Unknown → Medium |
status: | Unknown → In Progress |