On Fri, 28 Mar 2025 20:17:42 +0200 Adrian Bunk <
bunk@debian.org> wrote: Control: tags -1 patch
On Sat, Mar 01, 2025 at 07:21:04AM +0000, Debian Bug Tracking System
wrote:
...
ntpsec (1.2.3+dfsg1-4) unstable; urgency=low
...
* Backport armhf fixes (Closes: 1091303)
...
The problem still seems to be present: https://buildd.debian.org/status/logs.php?pkg=ntpsec&ver=1.2.3%2Bdfsg1-4
cu
Adrian
In 1.2.3+dfsg1-4,
it seems that patch [1] was applied to fix #1091303.
It will be compiled with/without NTP_TIMEX_LONG_LONG
conditionally, but it was not defined by wafhelpers/check_sizeof.py in
1.2.3 yet.
Thus NTP_TIMEX_LONG_LONG was not used at all eventually.
Instead, need to control defining NTP_TIMEX_LONG_LONG additionally.
I've attached PoC patch [2] trying to fix this issue.
(fairly rough approach, not tested well)
I hope that it will help.
[1] 0001-Fix-ntpd-ntp_control.c-to-handle-crazy-timex-slots.patch
[2] fix-bug1091303-ftbfs.patch
Regards,
Index: ntpsec-1.2.3+dfsg1/ntpd/ntp_control.c ===================================================================
--- ntpsec-1.2.3+dfsg1.orig/ntpd/ntp_control.c 2025-04-02 21:32:28.000000000 +0900
+++ ntpsec-1.2.3+dfsg1/ntpd/ntp_control.c 2025-04-02 22:33:28.036759533 +0900
@@ -174,6 +174,13 @@
static const double dbl_zero = 0.0;
#endif
+#if NTP_SIZEOF_TIME_T == 8 && NTP_SIZEOF_LONG == 4
+/* 0001-Fix-ntpd-ntp_control.c-to-handle-crazy-timex-slots.patch checks NTP_TIMEX_LONG_LONG,
+ but it was not defined in 1.2.3 yet.
+*/
+#define NTP_TIMEX_LONG_LONG 1
+#endif
+
enum var_type {v_time,
v_str, v_dbl, v_uli, v_li, v_uint, v_int,
v_u64, v_i64, v_u32, v_i32, v_u8, v_i8, v_bool,
--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)