From aaron@lehman.com Thu Mar 17 16:49:11 1994 Return-Path: <aaron@lehman.com> Received: from lehman.com by elsie.nci.nih.gov (4.1/SMI-4.1) id AA28619; Thu, 17 Mar 94 16:49:08 EST Received: from relay.lehman.com by lehman.com (8.6.4/LB 0.1) id QAA19119; Thu, 17 Mar 1994 16:48:58 -0500 Received: from killer.lehman.com by relay.lehman.com (4.1/LB-0.6) id AA04769; Thu, 17 Mar 94 16:48:51 EST Received: from yen.lehman.com by killer.lehman.com (5.0/SMI-SVR4) id AA29096; Thu, 17 Mar 1994 16:48:48 +0500 Date: Thu, 17 Mar 1994 16:48:48 +0500 From: aaron@lehman.com (Aaron Akman) Message-Id: <9403172148.AA29096@killer.lehman.com> Received: by yen.lehman.com (4.1/SMI-4.1) id AA03187; Thu, 17 Mar 94 16:48:45 EST To: ado@elsie.nci.nih.gov Subject: Time Conversion Question Content-Length: 1037 Status: RO

Arthur,

I've downloaded your timezone stuff (got pointers from Arnold Robbins and Don Libes), and have a question:

I'm looking to convert from GMT to one of a few timezones, or from one timezone to another. I'm concerned about setting TZ and calling tzset() because I think it changes some global state associated with the process. If I were interrupted before setting TZ back and re-calling tzset(), I could have an interrupt function that doesn't have the right local TZ setting. I.e.

get_other_local_time(char* tzsetting) { oldsetting = temporarily_set_TZ(tzsetting); tzset();

do_time_stuff_with_alternate_TZ();

/* i could get interrupted here by a signal handler that needs the real local time!! */

set_TZ_back(oldsetting); tzset(); }

Is there a better way? Is it true the effect of tzset() global? I haven't gone through the .c files in depth yet.

It seems like this should be a common problem for a globally distributed system.

Thanks in advance for any advice, Aaron Akman aaron@lehman.com 212 464 3978