chicken-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Chicken-users] Packaging eggs


From: Alaric Snell-Pym
Subject: Re: [Chicken-users] Packaging eggs
Date: Thu, 02 Sep 2010 08:27:30 +0100
User-agent: Mozilla/5.0 (X11; U; NetBSD amd64; en-US; rv:1.9.1.9) Gecko/20100520 Lightning/1.0b2pre Shredder/3.0.4

On 08/31/10 15:12, Felix wrote:
Argh. You're right. An install from a checkout will always set the version
to "unknown". On the other hand, just checking the directory name does not
necessarily mean that the directory name is the version. So, being picky
about it, one could argue that, unless the version comes from henrietta,
it is not known and can not be guaranteed. Which is true.

Aye. An egg that's being built from a "development version" that hasn't
been tagged as a version yet can't really have a meaningful version
number assigned. At best, it could be considered as having the version
number of the most recent tag it's based on, but even then, it'd be a lie.

SVN doesn't make it particularly easy to do this, but git has some
tooling to name a version by getting the most recent tag-name matching a
regexp (eg, v[0-9]+\.[0-9]+ or something) before the commit in question,
and then if that most recent tag isn't point to the actual HEAD
appending the hash of the HEAD, and then if the sources have local
modifications appending "-dirty". Or something like that.

The closest equivalent might be to produce version strings like
"r15231[-dirty]" if building from an SVN checkout!

Where I work we have slightly different development/official build
processes - the development process, which is run if you just "run
make", produces version strings based on the date, time, and username of
the builder (so anyone who leaks one can be BLAMED), while the official
builds are identified by a version number.

ABS

--
Alaric Snell-Pym
http://www.snell-pym.org.uk/alaric/



reply via email to

[Prev in Thread] Current Thread [Next in Thread]