| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
|
|
|
|
Mostly by renaming the shadowing variable.
If a shadowing variable was deleted,
the shadowed one can be used instead.
|
|
|
|
|
|
|
|
|
|
%x expects unsigned int and %p expects void *
GCC emits a diagnostic about %p/void* in pedantic mode:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=26542
|
|
|
|
|
|
|
|
|
|
Revert 44a5809224e819bf8bcca4ae9a8c5ea5bf4cefd7 and check range of
`base` with `mrb_assert`.
It actually can only be either 2, 8, 10 or 16 (and never 0), unless
mruby sources are modified. But I don't see a reason to impose further
limits.
|
|
codegen.c.
|
|
'clang-analyzer'"
This reverts commit 29f14e728d319d9fc80702ba78959c1957944767.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
#2258
|
|
|
|
|
|
|
|
take-cheeze-dump_lv
|
|
|
|
Check that `lv`'s length is always `nlocals - 1`.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The new name is better and less confusing, because:
- `mrb_str_append` calls `mrb_str_to_str` and this function doesn't
- `mrb_str_buf_append` _is_ `mrb_str_cat` for `mrb_value` strings
|
|
|
|
|
|
|
|
|
|
|
|
|
|
RRange excl flag use mrb_bool insteard of int
|
|
|
|
|
|
|
|
|
