Bugzilla Bug 210950

Compiling app-editors/emacs-cvs-23.0.60-r2 fails for two days because the file
README.unicode in no longer in the repo, but the ebuild wants to install it.

chmod -R a+r
/var/tmp/portage/app-editors/emacs-cvs-23.0.60-r2/image//usr/share/emacs/23.0.60
/var/tmp/portage/app-editors/emacs-cvs-23.0.60-r2/image//usr/share/emacs/site-lisp
/var/tmp/portage/app-editors/emacs-cvs-23.0.60-r2/image//usr/share/emacs/23.0.60/etc
/var/tmp/portage/app-editors/emacs-cvs-23.0.60-r2/image//usr/share/emacs/23.0.60/lisp
/var/tmp/portage/app-editors/emacs-cvs-23.0.60-r2/image//usr/share/info/emacs-23
thisdir=`/bin/pwd`; \
        cd
/var/tmp/portage/app-editors/emacs-cvs-23.0.60-r2/work/emacs/doc/man; \
        for page in emacs emacsclient etags ctags ; do \
          (cd ${thisdir}; \
           /usr/bin/install -c -m 644
/var/tmp/portage/app-editors/emacs-cvs-23.0.60-r2/work/emacs/doc/man/${page}.1
/var/tmp/portage/app-editors/emacs-cvs-23.0.60-r2/image//usr/share/man/man1/${page}.1;
\
           chmod a+r
/var/tmp/portage/app-editors/emacs-cvs-23.0.60-r2/image//usr/share/man/man1/${page}.1);
\
        done
 * Fixing info documentation ...
 * Fixing manpages ...
dodoc: README.unicode does not exist
 * 
 * ERROR: app-editors/emacs-cvs-23.0.60-r2 failed.
 * Call stack:
 *               ebuild.sh, line   49:  Called src_install
 *             environment, line 3370:  Called die
 * The specific snippet of code:
 *       dodoc README README.unicode BUGS || die "dodoc failed"
 *  The die message:
 *   dodoc failed
 * 
 * If you need support, post the topmost build error, and the call stack if
relevant.
 * A complete build log is located at
'/var/tmp/portage/app-editors/emacs-cvs-23.0.60-r2/temp/build.log'.
 * The ebuild environment file is located at
'/var/tmp/portage/app-editors/emacs-cvs-23.0.60-r2/temp/environment'.
 *

BTW: Is it really sensible to make a merge fail only because of some missing
doc file?  IMO a simple message that it couldn't be found and installed would
suffice here.

(In reply to comment #0)
> BTW: Is it really sensible to make a merge fail only because of some missing
> doc file?  IMO a simple message that it couldn't be found and installed would
> suffice here.

No, it completely sucks but the maintainer thinks otherwise... commented on
this multiple times, last time on Bug 190762

(In reply to comment #1)
> (In reply to comment #0)
> > BTW: Is it really sensible to make a merge fail only because of some missing
> > doc file?  IMO a simple message that it couldn't be found and installed would
> > suffice here.
> No, it completely sucks but the maintainer thinks otherwise... commented on
> this multiple times, last time on Bug 190762

 The comment by ulm there is still valid.  Fixed. Thanks for reporting

I understand the purpose, but why would a simple warning message

  File foo moved at upstream.  Please report that to emacs@gentoo.org.

not suffice?  The missing doc file doesn't make emacs unusable.

At least for temporary files like README.{unicode,multy-tty} killing the merge
makes no sense.  Those files were only there until someone volunteered to merge
the changes into the normal README.

(In reply to comment #3)
> I understand the purpose, but why would a simple warning message
>   File foo moved at upstream.  Please report that to emacs@gentoo.org.
> not suffice?  The missing doc file doesn't make emacs unusable.

 Because no one would read it. :)

> At least for temporary files like README.{unicode,multy-tty} killing the merge
> makes no sense.  Those files were only there until someone volunteered to merge
> the changes into the normal README.

 The live ebuilds are experimental (thus masked) and we want all tiny bits to
fail to notice them early.  So moving it over once they go in pretest phase or
even stable we already have a complete working ebuild without cruft.
 Anything to add, ulm?

*** Bug 211264 has been marked as a duplicate of this bug. ***