11 messages in com.perforce.perforce-userX-Platform UNIX and NT code - CR/LF i...
FromSent OnAttachments
Rick...@vsl.com15 Jun 1999 19:30 
Eric...@Adobe.COM15 Jun 1999 20:12 
Scot...@seaslug.org15 Jun 1999 20:55 
Rick...@vsl.com15 Jun 1999 21:22 
Rick...@vsl.com15 Jun 1999 21:29 
Scot...@seaslug.org15 Jun 1999 22:04 
Nick...@pobox.com16 Jun 1999 02:40 
Nick...@pobox.com16 Jun 1999 02:44 
Nick...@pobox.com16 Jun 1999 02:48 
Eric...@Adobe.COM16 Jun 1999 07:30 
Eric...@Adobe.COM16 Jun 1999 07:30 
Subject:X-Platform UNIX and NT code - CR/LF issues (fwd)
From:Nick...@pobox.com (Nick@pobox.com)
Date:06/16/1999 02:48:30 AM
List:com.perforce.perforce-user

At 1999-06-16 03:12:36+0000, Eric Scouten writes:

An alternative used by another team at Adobe that I just left is to use what we called a "mutex," which is basically an e-mail negotiation for exclusive write access to the sourcebase. In other words, when I have set of changes I want to contribute, I send e-mail requesting the mutex. I then check in my changes, pull them down on the other platform, verify them, make fixes as needed, and send another e-mail releasing the mutex. During the time I hold the mutex, it is understood that no other team member may check in and anyone who syncs to the database runs the risk of getting unfinished code.

A previous employer was developing a big application. Running more than one instance, even on one of the new powerful machines(*), would cause dreadful thrashing. The solution was a token system. One token (a photocopied bank note with suitable annotations) per machine. If you had the token for a machine about your person you could use that machine. Woe betide the developer who inadvertently took a token home!

Nick B

(*) this was about 1989, so those would have been Sparc 4s, I think.