https://gitlab.synchro.net/main/sbbs/-/commit/8d0dc7e7e95282494c1a2bdb
Modified Files:
src/sbbs3/atcodes.cpp ftpsrvr.cpp js_system.cpp jsdoor.cpp jsexec.cpp mailsrvr.cpp main.cpp sbbs.h services.cpp websrvr.cpp
Log Message:
Track server uptimes using time() again ...
Manual reversion of (most of) commits 0c0cb7c4732 and 5cbc2c2eb39
Since we expose the actual date/time of the server uptime value in JS system.uptime, this "fix" created a new issue (#1109).
Thanks for the report Craig!
Also, the original "fix" didn't really fix anything in the servers:
The original issue (#1093) was only observed in the status bar of SBBSCTRL-Win32, so leave that solution (using xp_fast_timer64()) since it appears to be a Borland-only issue not actually impacting the servers, which are built with just about every toolchain *but* Borland's.
The original code before the above mentioned commits included this old comment which might be a clue:
uptime = time(NULL); /* this must be done *after* setting the timezone */
A *long* time ago, the severs used to be built with Borland C/C++ and we also used to call tzset() during initialization (e.g. see commit f2a7dab6a08cef), so weird stuff used to be happening with the the time() base value changing during initialization and that no longer applies.
---
þ Synchronet þ Vertrauen þ Home of Synchronet þ [vert/cvs/bbs].synchro.net