lm_sensors2/doc version.texi

lm_sensors2/doc version.texi

[Jean Delvare]

> Update of /home/cvs/lm_sensors2/doc
> In directory mordac.netroedge.com:/tmp/cvs-serv5300
> 
> Modified Files:
> 	version.texi 
> Log Message:
> (Khali) Removed revision, set version to 2.9.

Due to the order in which I commited the changes, I expect the rule to
build the HTML FAQ to be used on a freshly checked out snapshot (and
thus for the end users of the 2.8.0 release). I see three possibilities
to prevent that.

1* Keep an eye on timestamps and change them as needed. I don't know how
to do that on the other CVS end. Do I need to force a commit?

2* Edit the rules so that the FAQ files are build if and only if they do
not exist, regardless of timestamps. The better for the user, but can
eventually be a burden for us developpers.

3* Ask make to ignore failures on these commands (using a leading dash).
This will show some errors which, even if non-fatal, may confuse the
users.

I'd be in favor of solution #2 but would like to hear any opinion or
other proposal.

-- 
Jean Delvare
http://www.ensicaen.ismra.fr/~delvare/

lm_sensors2/doc version.texi

[Mark Studebaker]

don't worry about it,
doc/Makefile isn't called from the top-level Makefile,
it has to be run manually. There's no reason for the user
to run make in doc/ because we rerun texi and check in
all 3 files after any change.

Jean Delvare wrote:
> 
> > Update of /home/cvs/lm_sensors2/doc
> > In directory mordac.netroedge.com:/tmp/cvs-serv5300
> >
> > Modified Files:
> >       version.texi
> > Log Message:
> > (Khali) Removed revision, set version to 2.9.
> 
> Due to the order in which I commited the changes, I expect the rule to
> build the HTML FAQ to be used on a freshly checked out snapshot (and
> thus for the end users of the 2.8.0 release). I see three possibilities
> to prevent that.
> 
> 1* Keep an eye on timestamps and change them as needed. I don't know how
> to do that on the other CVS end. Do I need to force a commit?
> 
> 2* Edit the rules so that the FAQ files are build if and only if they do
> not exist, regardless of timestamps. The better for the user, but can
> eventually be a burden for us developpers.
> 
> 3* Ask make to ignore failures on these commands (using a leading dash).
> This will show some errors which, even if non-fatal, may confuse the
> users.
> 
> I'd be in favor of solution #2 but would like to hear any opinion or
> other proposal.
> 
> --
> Jean Delvare
> http://www.ensicaen.ismra.fr/~delvare/

lm_sensors2/doc version.texi

[Jean Delvare]

> don't worry about it,
> doc/Makefile isn't called from the top-level Makefile,
> it has to be run manually. There's no reason for the user
> to run make in doc/ because we rerun texi and check in
> all 3 files after any change.

OK, perfect :)

-- 
Jean Delvare
http://www.ensicaen.ismra.fr/~delvare/