64 bit tsk_size_t #1618
64 bit tsk_size_t #1618
Conversation
|
@jeromekelleher I think this way is better than lots of casts, and as you point out allows us to encapsulate other complexity like avoiding zero-allocs 👍 |
|
OK, thanks @benjeffery. I think I'll try to cherry pick some of this stuff out so that it'll compile cleanly on 64 bit platforms, and they incorporate the other diffs into the 64 bit switch. |
jeromekelleher
force-pushed
the
size_t_32bit
branch
2 times, most recently
from
d1ff176 to
2406ddc
Compare
|
I think this is basically working @benjeffery, and we're possibly at the point where we should compile a list of follow-up issues and merge. I think this is all the critical stuff. I've we're happy with this set of changes, I think I'll squash down to one commit so that we're cleanly compilable on 32 and 64 bit if we ever need to git bisest. What do you think? |
Codecov Report
@@ Coverage Diff @@
## main #1618 +/- ##
==========================================
+ Coverage 93.68% 93.72% +0.03%
==========================================
Files 27 27
Lines 23306 23318 +12
Branches 1084 1084
==========================================
+ Hits 21834 21854 +20
+ Misses 1438 1430 -8
Partials 34 34
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
jeromekelleher
force-pushed
the
size_t_32bit
branch
from
63edf28 to
20ef148
Compare
jeromekelleher
force-pushed
the
size_t_32bit
branch
from
20ef148 to
643331b
Compare
jeromekelleher
changed the title
jeromekelleher
force-pushed
the
size_t_32bit
branch
from
643331b to
92f0782
Compare
jeromekelleher
force-pushed
the
size_t_32bit
branch
from
92f0782 to
5686314
Compare
jeromekelleher
force-pushed
the
size_t_32bit
branch
from
5686314 to
f7aab2e
Compare
Major change at all levels required to change the size of tsk_size_t to be 64 bit, and to make this compile cleanly on 64 and 32 bit machines. Basic C library changes for 64 bit tsk_size_t Update offset column accessors Fix uint32 in sample_set_sizes Misc size_t casts in library code size_t changes for tests Change tsk_xalloc shims to tsk_size_t Remove problematic consts in lwt code The numpy C API was discarding consts, so they are more trouble than they are worth.
jeromekelleher
force-pushed
the
size_t_32bit
branch
from
f7aab2e to
822ba61
Compare
I started making lots of casts for the various library functions, but then realised that this could actually hide bugs in some cases. So, adding wrappers for the various functions seemed like the best approach. Most of these should probably be macros, but we can figure out the details of this a bit later if we agree on the basic approach.
@benjeffery, what do you think?
Closes #1530
Closes #1527
Closes #1554