autoconf
[Top][All Lists]
Advanced

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

Re: ChangeLog (usage)


From: Guido Draheim
Subject: Re: ChangeLog (usage)
Date: Thu, 07 Aug 2003 10:26:46 +0200
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3) Gecko/20030313



Frank A. Uepping wrote:
Each directory containing source files should have a ChangeLog.
Should changes to the `GNU build system' source files (configure.ac,
Makefile.am) also be recorded into the ChangeLog?


Yes and No. Each directory and each change - that would be much too much,
and people would soon avoid to read (or write) a changelog. Instead, the
changelog is per each `module` of your of software, and in dramatic
majority of cases just one changelog is sufficient thereby carrying the
change log to a set of related submodules - most usually the src part,
the doc part, and i18n part. That helps also so people write concise
change log entries.

As to the change log entries, it depends on whether you use a form of
file change repository. A good project should have its sources safe
in a cvs - and that one remembers changes all on its own, so one can
leave out most information from a change log entry since a `cvs diff`
will give an impression anyway. It is even possible to use an automatic
script that assembles all `cvs ci -m` messages into a central changelog.

In the latter case, project maintainers would then use another file
to give an overview and describing changes that can not seen alone
from reading diffs and single-line commit-comments. That's needed,
atleast at each release. Some projects however use the (central)
changelog to desribe these, giving a larger explanation and for the
simple changes just dropping a `fixed <filename>` into the log.

So far it seems, that you have the impression to use the changelog
file as a variant of cvs history. Well, go ahead, and in that case
it is a good choice to have a number of changelog files around and
list all changes. However, better think about a cvs repository for
the stuff which yields much of the need for the list-all rule be void.

cheers,
-- guido                                  http://google.de/search?q=guidod
GCS/E/S/P C++/++++$ ULHS L++w- N++@ d(+-) s+a- r+@>+++ y++ 5++X- (geekcode)





reply via email to

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