#22 "make distclean" doesn't

open
nobody
None
5
2007-05-06
2007-05-06
No

"make distclean" leaves generated files behind. Although it's debatable exactly which files should be removed, and which ones shouldn't, at least most of the files generated in eval/ should be cleaned up.

Discussion

  • Sebastian Biallas

    Logged In: YES
    user_id=3437
    Originator: NO

    There is also "make maintainer-clean" which deletes more files.

     
  • Serge van den Boom

    Logged In: YES
    user_id=604796
    Originator: YES

    I see. That removes at least most of the files I'd expect to be cleaned up with "make distclean", or even "make clean", but shouldn't generated files like /doc/{README,ht.html,ht.info,hthelp.info} also be deleted?
    And also htdoc.c and htdoc.h, although I have to type "make htdoc.h" explicitely for them even to generated. (I'll add a separate report for that).

     
  • Serge van den Boom

    Logged In: YES
    user_id=604796
    Originator: YES

    Oh, never mind about htdoc.c; I see a ticket for that already exists.

     
  • Serge van den Boom

    Logged In: YES
    user_id=604796
    Originator: YES

    Oh, never mind about htdoc.c; I see a ticket for that already exists.

     
  • Sebastian Biallas

    Logged In: YES
    user_id=3437
    Originator: NO

    make clean: clean all binary stuff (Roughly)
    make distclean: Clean everything not needed for distribution (e.g. Makefiles that are generated by configure). The binson and flex generated files are distributed, so that binson/flex is not needed if you just want to compile ht.

    make maintainer-clean should also delete the doc-stuff, you're right. But is this really a problem?

     
  • Serge van den Boom

    Logged In: YES
    user_id=604796
    Originator: YES

    Well, it's not going to cause the end of the world. But it's not how it should be, so it deserves a bug report.
    I was doing a 'cvs diff' to see what I changed, and a whole lot of files unknown to cvs came up, even after 'make distclean'. I though I'd might as well be a nice user and make a small contribution by reporting the problem.

    Sorry about the double post. I pressed reload.

     
  • Sebastian Biallas

    Logged In: YES
    user_id=3437
    Originator: NO

    There are .cvsignore files in CVS which should take care of this.. Which files are problematic? "cvs diff" is absolutely clean for me.

     
  • Serge van den Boom

    Logged In: YES
    user_id=604796
    Originator: YES

    This is what CVS reports after a 'make maintainer-clean'. There's something to say for keeping many of those, but at least I don't think should be left.

    ? Makefile.in
    ? README
    ? aclocal.m4
    ? config.guess
    ? config.h.in
    ? config.sub
    ? configure
    ? depcomp
    ? htdoc.c
    ? htdoc.h
    ? missing
    ? analyser/Makefile.in
    ? asm/Makefile.in
    ? doc/Makefile.in
    ? doc/README
    ? doc/ht.html
    ? doc/ht.info
    ? doc/hthelp.info
    ? eval/Makefile.in
    ? info/Makefile.in
    ? io/djgpp/Makefile.in
    ? io/posix/Makefile.in

     
  • Sebastian Biallas

    Logged In: YES
    user_id=3437
    Originator: NO

    That's odd. These filenames are in ".cvsignore" and so should be ignored by CVS. Are you really using the cvs head branch?

     
  • Serge van den Boom

    Logged In: YES
    user_id=604796
    Originator: YES

    No, "dataport".
    There is no ".cvsignore" file here.

     
  • Sebastian Biallas

    Logged In: YES
    user_id=3437
    Originator: NO

    Please switch to HEAD. dataport was merged.

     
  • Serge van den Boom

    Logged In: YES
    user_id=604796
    Originator: YES

    Ah, much better. Now I just have these files left. They're probably the result from "automake --add-missing --copy".

    ? config.guess
    ? config.sub
    ? depcomp
    ? missing

     
  • Sebastian Biallas

    Logged In: YES
    user_id=3437
    Originator: NO

    Added them to .cvsignore

     

Log in to post a comment.