Precise non-backward change done on weak-pointers#515
Precise non-backward change done on weak-pointers#515bobot wants to merge 2 commits intoocaml:4.03from
Conversation
Changes
Outdated
| Runtime system: | ||
| - PR#7173 GPR#515: Add a public C API for weak arrays and ephemerons. | ||
| (François Bobot and Jacques-Henri Jourdan) | ||
| * GPR#22: The undocumented layout of weak arrays have been changed. |
byterun/caml/weak.h
Outdated
|
|
||
| CAMLextern mlsize_t caml_ephemeron_key_length(value eph); | ||
| /** return the number of key in the ephemeron. The valid key offset goes | ||
| from [0] to the predecessor of the returned value. */ |
There was a problem hiding this comment.
Comments like this usually start with capital letters, no?
Maybe you should add "inclusive" after "value" (if that's what you mean).
byterun/caml/weak.h
Outdated
| caml_ephemeron_unset_key, this function does not prevent the | ||
| incremental GC from erasing the value in its current cycle. The | ||
| values [eph1] and [eph2] must be ephemerons and the offset between | ||
| [off1] and [off1+len] and between [off2] and [off2+offset] must be |
There was a problem hiding this comment.
s/offset/offsets (on line 75)
bobot
force-pushed
the
ephemeron_c_api
branch
3 times, most recently
from
6c467f6 to
4bf1a58
Compare
|
Except for the name used for the exceptions, all the current comments have been taken into account. |
Changes
Outdated
|
|
||
| Runtime system: | ||
| - PR#7173 GPR#515: Add a public C API for weak arrays and ephemerons. | ||
| (François Bobot and Jacques-Henri Jourdan, review by Mark Shinwell) |
There was a problem hiding this comment.
I suspect you broke the careful sorting of the issue list for this and GPR#22 just to spite me :-)
|
@bobot can you make a new GPR for the bug fix? It's too late to add so much code in 4.03. |
|
@damiendoligez Ok. Could I keep only the bug fix in this GPR and move all the new code to another GPR? Just because this one target 4.03. |
bobot
changed the title
|
@damiendoligez this merge request now contains only the prefix of the previous commit. |
|
@damiendoligez could you please reschedule this GPR for 4.03 which now contains only doc related modifications or a bug fix. |
|
Since this pull request targeted |
Add memprof stubs to build and stdlib
5 participants
Fix #7173. There is a discussion in mantis if it should go to 4.03. I'm not convinced, except for the two first commits. I can create a separate GPR for them.