ChangeLog computation with savannah: "cvs server: nothing known about ymakefile"

ChangeLog computation with savannah: "cvs server: nothing known about ymakefile"

[Steven Tamm]

I got sick of doing ChangeLog computes by hand; so I setup ssh-agent 
and got the whole thing working.   However, if I use rcs2log (1.48) 
from the emacs/src directory without any arguments, I get back the 
following errors:

cvs server: nothing known about ymakefile
cvs server: nothing known about =convexos.h
cvs server: nothing known about =mach2.h
cvs server: nothing known about =sol2-2.h

The knock-off cvs2cl.pl does the same thing so I assume it's a savannah 
problem.  Does anyone else have this problem against savannah using 
ssh?  I really would rather use the menu option than the elisp 
function.

Thanks,
-Steven

Re: ChangeLog computation with savannah: "cvs server: nothing known about ymakefile"

[Stefan Monnier]

> I got sick of doing ChangeLog computes by hand;

What are you referring to?

Are you talking about creating ChangeLog entries out of the CVS log messages?
If so, I recommend that you do it the other way around: write the ChangeLog
entries first and then use C-c C-a in the *VC-log* (or *cvs-commit*) buffer
to fetch them and format them for CVS.

The ChangeLog format is more powerful/flexible so it's the better place to
"do it by hand".  Also Emacs provides facilities to help you write
ChangeLog entries (things like C-x 4 a which also work from a *vc-diff*
buffer), whereas this has not been adapted (yet) to work with a *vc-log*
or *cvs-commit* buffer.


        Stefan

Re: ChangeLog computation with savannah: "cvs server: nothing known about ymakefile"

[Adrian Aichner]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> I got sick of doing ChangeLog computes by hand;
>
> What are you referring to?
>
> Are you talking about creating ChangeLog entries out of the CVS log messages?
> If so, I recommend that you do it the other way around: write the ChangeLog
> entries first and then use C-c C-a in the *VC-log* (or *cvs-commit*) buffer
> to fetch them and format them for CVS.
>
> The ChangeLog format is more powerful/flexible so it's the better place to
> "do it by hand".  Also Emacs provides facilities to help you write
> ChangeLog entries (things like C-x 4 a which also work from a *vc-diff*
> buffer), whereas this has not been adapted (yet) to work with a *vc-log*
> or *cvs-commit* buffer.

The XEmacs package xemacs-devel contains patcher.el which goes quite
far in automating patch creation, regeneration, commital, and
documentation.  It easest the process a great deal by providing
project configurations that will keep track of various parameters of a
project (repository, patch email destination, changelog appearance,
etc.).

-- 
Adrian Aichner
 mailto:adrian@xemacs.org
 http://www.xemacs.org/

Re: ChangeLog computation with savannah: "cvs server: nothing known about ymakefile"

[Steven Tamm]

C-x v a or (vc-update-change-log) is to what I was referring.  It seems 
to work now, and the results it produces are, as is readily apparent, 
useless.  It regenerates the entire changelog, not just finding all the 
changes in the current directory and prepending them to the Log; which 
is what I would expect.  cvs2cl.pl includes a parameter "--accum" which 
does what I think it should.

Doing it "by hand" was doing C-x 4 a, which means I need to go to each 
functional difference and modify the changelog individually.  However, 
if C-x 4 a works from a *vc-diff*, then I can figure it out from there; 
I've only tried it from vc-log or cvs-commit, so I assumed it didn't 
work.

Thanks!
-Steven

On Apr 25, 2004, at 3:25 PM, Stefan Monnier wrote:

>> I got sick of doing ChangeLog computes by hand;
>
> What are you referring to?
>
> Are you talking about creating ChangeLog entries out of the CVS log 
> messages?
> If so, I recommend that you do it the other way around: write the 
> ChangeLog
> entries first and then use C-c C-a in the *VC-log* (or *cvs-commit*) 
> buffer
> to fetch them and format them for CVS.
>
> The ChangeLog format is more powerful/flexible so it's the better 
> place to
> "do it by hand".  Also Emacs provides facilities to help you write
> ChangeLog entries (things like C-x 4 a which also work from a *vc-diff*
> buffer), whereas this has not been adapted (yet) to work with a 
> *vc-log*
> or *cvs-commit* buffer.
>
>
>         Stefan
>
>
> _______________________________________________
> Emacs-devel mailing list
> Emacs-devel@gnu.org
> http://mail.gnu.org/mailman/listinfo/emacs-devel

Re: ChangeLog computation with savannah: "cvs server: nothing known about ymakefile"

[Kai Grossjohann]

Steven Tamm <steventamm@mac.com> writes:

> Doing it "by hand" was doing C-x 4 a, which means I need to go to each
> functional difference and modify the changelog individually.

But don't you have to do that anyway?

I usually hit `=' on the subdir that has changed, then I iterate over
the diff hunks, hit C-x 4 a on each of them to enter my information.
If multiple places are affected by the same change, then this is a
somewhat tedious procedure, though, since you have to navigate to all
of the spots, do C-x 4 a on each of them, and then enter the ChangeLog
info after you have found them all.  There doesn't seem to be a way of
removing the hunks that you have already processed from the diff
buffer.  If there was, then the cvs diff buffer would be a convenient
todo list for the ChangeLog entries, and after hitting C-x 4 a on a
hunk and thereby working on it, you can remove the hunk and thus mark
it as done.

But even if you edit the commit messages, you still have to provide
the same information.  So I don't see how the work would be less.

Kai

Re: ChangeLog computation with savannah: "cvs server: nothing known about ymakefile"

[Stefan Monnier]

> info after you have found them all.  There doesn't seem to be a way of
> removing the hunks that you have already processed from the diff
> buffer.

M-k ?


        Stefan

Re: ChangeLog computation with savannah: "cvs server: nothing known about ymakefile"

[Kai Grossjohann]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> info after you have found them all.  There doesn't seem to be a way of
>> removing the hunks that you have already processed from the diff
>> buffer.
>
> M-k ?

*blush*

Kai