[virt-tools-list] [PATCH 3/5] Update gitignore

[Christophe Fergeau]

Hey,

On Tue, Nov 03, 2015 at 09:38:24AM -0200, Eduardo Lima (Etrunko) wrote:
> On 02/11/15 08:36, Christophe Fergeau wrote:
> > On Thu, Oct 29, 2015 at 06:40:38PM -0200, Eduardo Lima (Etrunko) wrote:
> >> Signed-off-by: Eduardo Lima (Etrunko) <etrunko at redhat.com>
> >> ---
> >>  Makefile.am     | 5 +++++
> >>  git.mk          | 2 +-
> >>  m4/.gitignore   | 0
> >>  src/Makefile.am | 8 ++++++++
> >>  4 files changed, 14 insertions(+), 1 deletion(-)
> >>  delete mode 100644 m4/.gitignore
> >>
> >> diff --git a/Makefile.am b/Makefile.am
> >> index d0e9ab4..9646119 100644
> >> --- a/Makefile.am
> >> +++ b/Makefile.am
> >> @@ -48,6 +48,11 @@ MAINTAINERCLEANFILES =				\
> >>  	$(srcdir)/m4/lt~obsolete.m4		\
> >>  	$(NULL)
> >>  
> >> +GITIGNOREFILES =				\
> >> +	AUTHORS					\
> > 
> > This one should be in xxxCLEANFILES (MAINTAINER maybe?)
> > 
> >> +	build-aux/test-driver			\
> > 
> > Not sure about this one, but I suspect it belongs to
> > MAINTAINERCLEANFILES as well as the other build-aux/ files are listed in
> > it.
> > 
> 
> Yes, you can find those in the final Makefile, but test-driver is the
> only one missing. Maybe a bug in automake?

Makefile.am has a MAINTAINERCLEANFILES variable with all these build-aux
files listed except for test-driver, the bug would be there. Now that I
experimented a bit more with this, all these build-aux files are covered
by one of:

 MAINTAINERCLEANFILES =                         \
+       $(GITIGNORE_MAINTAINERCLEANFILES_TOPLEVEL) \
+       $(GITIGNORE_MAINTAINERCLEANFILES_MAKEFILE_IN) \
+       $(GITIGNORE_MAINTAINERCLEANFILES_M4_LIBTOOL)  \
(sorry dunno which one ;)

> 
> >> +	$(NULL)
> >> +
> >>  dist-hook: gen-ChangeLog gen-AUTHORS
> >>  
> >>  # Generate the ChangeLog file (with all entries since the switch to git)
> >> diff --git a/git.mk b/git.mk
> >> index 0b26b23..277e6e3 100644
> >> --- a/git.mk
> >> +++ b/git.mk
> >> @@ -123,7 +123,7 @@ $(srcdir)/.gitignore: Makefile.am $(top_srcdir)/git.mk
> >>  				$(gsettings__enum_file) \
> >>  			; do echo /$$x; done; \
> >>  		fi; \
> >> -		if test -f $(srcdir)/po/Makefile.in.in; then \
> >> +		if test -f $(top_srcdir)/po/Makefile.in.in; then \
> > 
> > Same interrogation as Fidencio here. Are you updating git.mk from
> > 'upstream', or was it changed for a different reason?
> > 
> 
> git.mk assumes that the po/ directory is located under src/, not in the
> root directory, as it happens for virt-viewer. I ended up getting the
> whole lot of untracked files under po/ after running 'make distcheck',
> can you try it and see if it happens for you too?

I haven't been able to reproduce this.

> 
> >>  			for x in \
> >>  				po/Makefile.in.in \
> >>  				po/Makefile.in \
> >> diff --git a/m4/.gitignore b/m4/.gitignore
> >> deleted file mode 100644
> >> index e69de29..0000000
> >> diff --git a/src/Makefile.am b/src/Makefile.am
> >> index 1ebc24e..5ce08bc 100644
> >> --- a/src/Makefile.am
> >> +++ b/src/Makefile.am
> >> @@ -218,4 +218,12 @@ debug_helper_LDFLAGS = $(GLIB2_LIBS) -Wl,--subsystem,windows
> >>  debug_helper_CFLAGS = $(GLIB2_CFLAGS)
> >>  endif
> >>  
> >> +GITIGNOREFILES =	\
> >> +	view/.deps	\
> >> +	view/.dirstamp	\
> >> +	view/.libs	\
> >> +	view/*.lo	\
> >> +	view/*.o	\
> >> +	$(NULL)
> > 
> > This hunk does not look correct, the whole point of git.mk is to avoid
> > listing these explicitly. What happens here is that git.mk does not know
> > it should recurse in view/ (I don't know if it's supposed to guess or
> > not). One way around it was to add a Makefile.am in view/ only
> > containing -include $(top_srcdir)/git.mk (+ the associated changes in
> > configure.ac and src/Makefile.am). I don't know if there's a better way
> > of handling this.
> > 
> 
> Yes, I did not want to create an empty Makefile inside of view/ because
> recursing takes a lot of time and adds up to build times. I also find it
> strange that you need to run make first in order to generate the
> .gitignore files, but I guess this is how it was designed to work in
> first place.
> 
> In the end, I can live with either approach, just want to know which
> should be.

Actually, updating git.mk to the latest version seems to fix this
particular issue with no changes on virt-viewer side
https://github.com/behdad/git.mk

Christophe
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/virt-tools-list/attachments/20151103/4d94542c/attachment.sig>