[Reportbug-maint] reportbug project: Changelog entries versus VCS commit messages

Sandro Tosi matrixhasu at gmail.com
Thu Jun 12 13:40:21 UTC 2008


>> However, documenting minor changes may mean that the larger and more
>> important changes that we want people to know would really be 'lost
>> in the crowd', so to speak.  I would say, use your good judgement :)
>
> This is my concern too. If the package changelog documents
> purely-internal changes as well as user-visible changes, it becomes
> far too verbose for users.

So, what should I write in changelog if I did only internal changes?
nothing? that's wrong! I think you should contact more DDs (I'm not,
yet) and ask what they think: NONE of them ever complained about my
changelog (they usually request to be more verbose than less), and I
do believe they are right.

> Better that the changelog documents only changes that are
> user-visible, so that the binary package contains a changelog that is
> of use to the user.

If a user what to read something, he/she have to do it carefully, and
even internal or "external" changes could be important for him/her,
mainly because reportbug has a widespread user audience, with many
user interested in low-level changes too (or would you discriminate
them?)

>> > for sure there is a need to document them (consider a new
>> > developer stepping in and wanted to know when a file has changed
>> > and why), hence debian/changelog it the best place to do it.
>>
>> For this, I think svn log of the file is the best source.  Also,
>> usually all the files (and the content of those files) that changed
>> for a given fix or a modification is of greater importance.
>
> Yes. The new developer should be using the VCS log history to become
> familiar with the flow of changes to the source, *and* the package
> changelog to see what user-visible changes have been occurring. The
> former is far too much information for end-users, though, so should
> not go into the package changelog.

Again, that's wrong! moreover, if that would be in a license, that
would fail the "desert island" test, hence it would fail DFSG.

*I* will continue with what I think it's the right way to write
debian/changelog (supported by many DDs and moreover by policy), of
course you are free to do what's best for you to get the work done.

That's all IMVHO,
Sandro

-- 
Sandro Tosi (aka morph, Morpheus, matrixhasu)
My website: http://matrixhasu.altervista.org/
Me at Debian: http://wiki.debian.org/SandroTosi



More information about the Reportbug-maint mailing list