[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-ha-dev
Subject:    Re: [Linux-ha-dev] Commit http://hg.linux-ha.org/dev/rev/13dc55dece11
From:       "Andrew Beekhof" <beekhof () gmail ! com>
Date:       2007-07-26 6:00:57
Message-ID: 26ef5e70707252300y473b1ac4pc60b094a8ec9960b () mail ! gmail ! com
[Download RAW message or body]

On 7/25/07, Alan Robertson <alanr@unix.sh> wrote:
> Andrew Beekhof wrote:
> > Alan,
> >
> > is there any chance you could write commit messages that actually relate
> > to the change being made?
> >
> > ie. "Bug 1661 - Use HA_NOARCHDATAHBDIR for the location of crm.dtd as
> > HA_LIBDIR is incorrect on myArch"
>
> I think that particular message got truncated.  It certainly looks
> mangled.  I can't think exactly what myArch would mean, nor why I would
> have typed it (presuming that I did).


It was an example of a commit message that would be a large
improvement on what you wrote:
  "LF bug # 1661 - crm_verify aborts checking and reoccurring
crm_abort errors in logs"

which says nothing about what has been changed and requires one to
jump through various hoops to understand the why.

The first line in http://hg.linux-ha.org/dev/rev/0c8dc61feeb2 was
better, but i'm not sure why the second line was needed.

>
> I don't intend that my messages are obscure to you when you read them,
> any more than you intend for your commit messages to be obscure when I
> read them.
>
> However, the explanation for this particular change is:
>
> --------------------------------------------------------
> Files which aren't architecture related are supposed to go in /usr/lib
> or /usr/lib64.  Instead they're supposed to go in /usr/share/packagename
> - which is HA_NOARCHDATAHBDIR - trying as best I could to follow the
> naming conventions you changed things to.  This commit made our file
> locations consistent with that convention.
>
> I'm not sure if it's part of the LSB or FHS standard or merely a
> convention, but I know that Red Hat requires it for Fedora packages, and
> it is how everyone else does this kind of thing.  [MIBs, perl scripts,
> and a basically everything else follows this convention].
> ---------------------------------------------------------
>
> It also seems that multi-line commit messages aren't ordinarily printed.
>  This is one of many cases where I would have liked to have put a longer
> commit message.
>
> Is there a straightforward way to read the history of a set of changes
> and get the entire commit message - not just the first line?

It cant print what isn't there.

Here is a commit with multiple lines:
   http://hg.linux-ha.org/dev/rev/c7d2cbecb0d4

The one I mentioned does not.

If you want multiple lines, I suggest not supplying -m and letting Hg
start your editor.

> If so, then making a meaningful commit message would become easier, as
> would writing a meaningful set of release notes.
_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
[prev in list] [next in list] [prev in thread] [next in thread]