| 1 |
> On 31 Oct 2022, at 10:37, Mickaël Bucas <mbucas@×××××.com> wrote: |
| 2 |
> |
| 3 |
> Le dim. 30 oct. 2022 à 16:48, Sam James <sam@g.o> a écrit : |
| 4 |
>> |
| 5 |
>> Signed-off-by: Sam James <sam@g.o> |
| 6 |
>> --- |
| 7 |
>> eclass/flag-o-matic.eclass | 13 +++++++++++++ |
| 8 |
>> 1 file changed, 13 insertions(+) |
| 9 |
> |
| 10 |
>> +# @FUNCTION: append-time64-flags |
| 11 |
>> +# @DESCRIPTION: |
| 12 |
>> +# Add flags that enable 64-bit time_t. Implies Large File Support |
| 13 |
>> +# (calls append-lfs-flags) automatically. |
| 14 |
|
| 15 |
> Hello Sam |
| 16 |
> |
| 17 |
|
| 18 |
Hi Mickaël, |
| 19 |
|
| 20 |
> As a simple user, can you explain to me or point me to a resource |
| 21 |
> about the link between Large File Support and time64 configuration ? |
| 22 |
> |
| 23 |
|
| 24 |
This is a good question -- the answer is two-fold: |
| 25 |
1. glibc requires LFS to enable time64 |
| 26 |
2. Both can involve an ABI break if an off_t or time_t for example |
| 27 |
are in an exposed structure, so it's best to do them at the same time. |
| 28 |
|
| 29 |
But I've had second thoughts about this and I want to rethink |
| 30 |
if we want to take this approach in the patch I sent. |
| 31 |
|
| 32 |
I think we may want to do a hard-switch for defaults |
| 33 |
instead in new profiles. I need to reflect on it a bit more. |
| 34 |
|
| 35 |
Plans and notes are at https://wiki.gentoo.org/wiki/Project:Toolchain/time64_migration |
| 36 |
if interested! |
| 37 |
|
| 38 |
> Thanks |
| 39 |
> Best regards |
| 40 |
> |
| 41 |
|
| 42 |
Best, |
| 43 |
sam |
Archives