[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
timestamps vs. checksums
From: |
Markus Werle |
Subject: |
timestamps vs. checksums |
Date: |
Wed, 05 Mar 2003 11:46:28 +0100 |
Hi!
Recently I had an email discussion about build tools.
Someone stated that make was flawed due to the fact that it
uses timestamps to decide whether to recompile a file or not.
I disagree with saying that make is flawed.
Actually NFS is flawed, which I encountered today:
make recognized the files had changed (after I waited 6 seconds)
and then compiled the old version provided by the dormant nfs server
(actually invoking "less" always updates the nfs cache, so maybe
make could be improved in this context)
So I start to think about how to improve.
Checking whether a file has changed is just a minor task
among others that make has to address.
I wonder how difficult it would be to change this part of make
(maybe as an addition --use-checksums) when we rely on something
like md5.
Of course this leads to extra bookkeeping of md5 checksum pairs,
but I do not mind having yet another hidden directory created by make.
Any comments?
Is there a reason for not having this feature?
Markus
- timestamps vs. checksums,
Markus Werle <=