[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[sr #111111] AC_SYS_LARGEFILE system wide control and autoconf releases
From: |
Румен Петров |
Subject: |
[sr #111111] AC_SYS_LARGEFILE system wide control and autoconf releases |
Date: |
Sat, 31 Aug 2024 05:02:17 -0400 (EDT) |
Follow-up Comment #2, sr #111111 (group autoconf):
https://savannah.gnu.org/support/index.php?&func=detailitem&item_id=111111&
My current solution is publicly available -
https://gitlab.com/secsh/pkixssh/-/blob/master/m4/android.m4 . In brief:
ac_cv_sys_largefile_opts='none needed'
ac_cv_sys_file_offset_bits=no
I'm would like to use a stable solution. Documentation states we to expect
additional modifications. This is reason to open this issue.
So use "_cv_" model or ...?
Another clear thing is that Android OS will never correct 64-bits for
32-devices. This support will remain buggy. No surprises - devices with short
life cycle.
Remark: In the past build (NDK) uses header files per API level. This model
does not have issue with "large files" support. If I remember well issue start
with use of so called "unified model" for header files. Now this is the only
model used in build process.
As alternative I could call for other "host" except Android. For instance:
case "$host" in
*-*-linux-android*)
;;
*)
AC_SYS_LARGEFILE
;;
esac
From my point of view such code is not reliable - another macro could use
large file as prerequisite. In such case code above is useless as from
"another macro" 64-bit will be activated.
What could be alternative solution?
* macro redefinition in project code :/
* update macro "large files" to exclude Android :(
* other
_______________________________________________________
Reply to this item at:
<https://savannah.gnu.org/support/?111111>
_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/
signature.asc
Description: PGP signature