On Friday 07 January 2005 05:24 pm, Stephen Turner wrote: > On Fri, 7 Jan 2005, Vincent Lefevre wrote: > > Now, you already have a > > > > #include <unistd.h> > > > > (conditionally defined), so there should be nothing wrong to support > > time-related POSIX functions when they are available. > > There is no possibility that I'm going to add any more OS-dependent stuff. > It's a complete nightmare to write and maintain.
Vincent, what's wrong with piping your log file through a little awk sript if you're not happy with timestamps in there and want to massage them a bit before processing? This way, if you want combined statistics from several mirror sites running in different timezones on different OS/server software, *you* fix the timestamps the way *you* want and then analog will just do its thing. What you're proposing involves creating a portable date-analyzing AI, botling it on top of a logfile analyzer, and praying it doesn't quietly screw up your timestamps every time you run the analysis. Dima -- Dimitri Maziuk Programmer/sysadmin BioMagResBank, UW-Madison -- http://www.bmrb.wisc.edu +------------------------------------------------------------------------ | TO UNSUBSCRIBE from this list: | http://lists.meer.net/mailman/listinfo/analog-help | | Usenet version: news://news.gmane.org/gmane.comp.web.analog.general | List archives: http://www.analog.cx/docs/mailing.html#listarchives +------------------------------------------------------------------------