bug#45186: Emacs master: build broken by spurious message ".../c-by.elc failed to provide feature `semantic/bovine/c-by"

bug#45186: Emacs master: build broken by spurious message ".../c-by.elc failed to provide feature `semantic/bovine/c-by"

[Alan Mackenzie]

Hello, Emacs.

I've just updated a master repo here, and tried a make bootstrap.

I get the error message:

    cedet/semantic/bovine/c.el:35:1: Error: Loading file
    /home/acm/emacs/emacs.git/sub-master-a/lisp/cedet/semantic/bovine/c-by.elc
    failed to provide feature `semantic/bovine/c-by'

.  There is a perfectly good line at the end of this file:

    (provide 'c-by)

.  So what is expecting the compound symbol `semantic/bovine/c-by' to be
provided, and why?

-- 
Alan Mackenzie (Nuremberg, Germany).

bug#45186: Emacs master: build broken by spurious message ".../c-by.elc failed to provide feature `semantic/bovine/c-by"

[Lars Ingebrigtsen]

Alan Mackenzie <acm@muc.de> writes:

> I've just updated a master repo here, and tried a make bootstrap.
>
> I get the error message:
>
>     cedet/semantic/bovine/c.el:35:1: Error: Loading file
>     /home/acm/emacs/emacs.git/sub-master-a/lisp/cedet/semantic/bovine/c-by.elc
>     failed to provide feature `semantic/bovine/c-by'

I tried a "make bootstrap" on the current trunk now (on
Debian/bullseye), and I did not get that error message...

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#45186: Emacs master: build broken by spurious message ".../c-by.elc failed to provide feature `semantic/bovine/c-by"

[Alan Mackenzie]

Hello, Lars.

On Fri, Dec 11, 2020 at 21:46:35 +0100, Lars Ingebrigtsen wrote:
> Alan Mackenzie <acm@muc.de> writes:

> > I've just updated a master repo here, and tried a make bootstrap.

> > I get the error message:

> >     cedet/semantic/bovine/c.el:35:1: Error: Loading file
> >     /home/acm/emacs/emacs.git/sub-master-a/lisp/cedet/semantic/bovine/c-by.elc
> >     failed to provide feature `semantic/bovine/c-by'

> I tried a "make bootstrap" on the current trunk now (on
> Debian/bullseye), and I did not get that error message...

Thanks for looking at it.

I got the error message again this morning (it wasn't just a bad dream),
and I'm trying to think how it could have happened.

The way of calling require

    (require 'semantic/bovine/c-by)

with the path included in the required symbol is totally new to me.  It
just seems like a wrong thing to do.  But it's been working for many
years, obviously, for some value of "working".

I do a git status, and notice amongs my untracked files list:

        lisp/cedet/semantic/wisent/java-tags-wy.el
	lisp/cedet/srecode/srecode-template-wy.el

, so I've deleted these.  It hasn't helped.

I've inserted a

    (provide 'semantic/bovine/c-by)

into c-by.el.  Now I get other errors:

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
In toplevel form:
cedet/semantic/wisent/js-wy.el:58:17: Error: Bad input grammar
make[2]: *** [Makefile:295: cedet/semantic/wisent/js-wy.elc] Error 1
make[2]: *** Waiting for unfinished jobs....

In toplevel form:
cedet/semantic/wisent/javascript.el:31:1: Error: Bad input grammar
make[2]: *** [Makefile:295: cedet/semantic/wisent/javascript.elc] Error
1

In toplevel form:
cedet/semantic/wisent/python-wy.el:58:17: Error: Symbol `SEMICOLON' is
used, but is not defined as a token and has no rules
make[2]: *** [Makefile:295: cedet/semantic/wisent/python-wy.elc] Error 1

In toplevel form:
cedet/semantic/wisent/python.el:37:1: Error: Symbol `SEMICOLON' is used,
but is not defined as a token and has no rules
make[2]: *** [Makefile:295: cedet/semantic/wisent/python.elc] Error 1

In end of data:
cedet/semantic/bovine/c.el:2272:1: Warning: the function
    `semantic-c-by--install-parser' is not known to be defined.
make[2]: Leaving directory '/home/acm/emacs/emacs.git/sub-master-a/lisp'
make[1]: *** [Makefile:318: compile-main] Error 2
make[1]: Leaving directory '/home/acm/emacs/emacs.git/sub-master-a/lisp'
make: *** [Makefile:411: lisp] Error 2
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

.  The word SEMICOLON doesn't appear anywhere in the git log.  I really
don't want to start burrowing through the internals of the cedet things.

I don't understand how Emacs is building for everybody else but not me.
I've tried reconfiguring my Emacs with no options to ./configure.  This
hasn't helped either.

Perhaps this conversation would be better on emacs-devel.  Anyhow thanks
for your patience in reading this miserable post!

> -- 
> (domestic pets only, the antidote for overdose, milk.)
>    bloggy blog: http://lars.ingebrigtsen.no

-- 
Alan Mackenzie (Nuremberg, Germany).

bug#45186: Emacs master: build broken by spurious message ".../c-by.elc failed to provide feature `semantic/bovine/c-by"

[Lars Ingebrigtsen]

Alan Mackenzie <acm@muc.de> writes:

> I do a git status, and notice amongs my untracked files list:
>
>         lisp/cedet/semantic/wisent/java-tags-wy.el
> 	lisp/cedet/srecode/srecode-template-wy.el
>
> , so I've deleted these.  It hasn't helped.

[...]

> I don't understand how Emacs is building for everybody else but not me.
> I've tried reconfiguring my Emacs with no options to ./configure.  This
> hasn't helped either.
>
> Perhaps this conversation would be better on emacs-devel.  Anyhow thanks
> for your patience in reading this miserable post!

:-)

I'd suggest pulling down a clean Emacs tree to see whether the problem
is present there, too...  or perhaps use a "make extraclean" (I think
that's the target that'll remove everything, even more than "bootstrap")
and see whether that helps.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#45186: Emacs master: build broken by spurious message ".../c-by.elc failed to provide feature `semantic/bovine/c-by"

[Alan Mackenzie]

Hello again, Lars.

On Sat, Dec 12, 2020 at 12:58:07 +0100, Lars Ingebrigtsen wrote:
> Alan Mackenzie <acm@muc.de> writes:

> > I do a git status, and notice amongs my untracked files list:

> >         lisp/cedet/semantic/wisent/java-tags-wy.el
> > 	lisp/cedet/srecode/srecode-template-wy.el

> > , so I've deleted these.  It hasn't helped.

[...]

> I'd suggest pulling down a clean Emacs tree to see whether the problem
> is present there, too...  or perhaps use a "make extraclean" (I think
> that's the target that'll remove everything, even more than "bootstrap")
> and see whether that helps.

I tried the extraclean first.  It worked!  Phew!

There must have been something hanging over from cedet in files not
removed by bootstrap.  I hadn't actually used the repo for, perhaps, a
year.  I'll remember the recipe.

Have a good Saturday!

> -- 
> (domestic pets only, the antidote for overdose, milk.)
>    bloggy blog: http://lars.ingebrigtsen.no

-- 
Alan Mackenzie (Nuremberg, Germany).

bug#45186: Emacs master: build broken by spurious message ".../c-by.elc failed to provide feature `semantic/bovine/c-by"

[Alan Mackenzie]

Hello, Lars.

On Sat, Dec 12, 2020 at 12:58:07 +0100, Lars Ingebrigtsen wrote:
> Alan Mackenzie <acm@muc.de> writes:

[ .... ]

> I'd suggest pulling down a clean Emacs tree to see whether the problem
> is present there, too...  or perhaps use a "make extraclean" (I think
> that's the target that'll remove everything, even more than "bootstrap")
> and see whether that helps.

Whoops, forgot about the actual bug.  It feels like there was an actual
bug in make bootstrap here, but it's not obvious what.

Clearly the bug needs to be closed.  What do you think it is?  a
notabug, a wontfix, or what?

> -- 
> (domestic pets only, the antidote for overdose, milk.)
>    bloggy blog: http://lars.ingebrigtsen.no

-- 
Alan Mackenzie (Nuremberg, Germany).

bug#45186: Emacs master: build broken by spurious message ".../c-by.elc failed to provide feature `semantic/bovine/c-by"

[Lars Ingebrigtsen]

Alan Mackenzie <acm@muc.de> writes:

> Whoops, forgot about the actual bug.  It feels like there was an actual
> bug in make bootstrap here, but it's not obvious what.

It is somewhat surprising that "make bootstrap" removes less stuff than
"extraclean", I think?  People understandably get a bit confused about
all these "clean" levels, but I'm not sure what to do about it...

> Clearly the bug needs to be closed.  What do you think it is?  a
> notabug, a wontfix, or what?

Just a "close" with no other status, I think.  :-)

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#45186: Emacs master: build broken by spurious message ".../c-by.elc failed to provide feature `semantic/bovine/c-by"

[Alan Mackenzie]

Hello, Lars.

On Sat, Dec 12, 2020 at 14:17:40 +0100, Lars Ingebrigtsen wrote:
> Alan Mackenzie <acm@muc.de> writes:

> > Whoops, forgot about the actual bug.  It feels like there was an actual
> > bug in make bootstrap here, but it's not obvious what.

> It is somewhat surprising that "make bootstrap" removes less stuff than
> "extraclean", I think?  People understandably get a bit confused about
> all these "clean" levels, but I'm not sure what to do about it...

> > Clearly the bug needs to be closed.  What do you think it is?  a
> > notabug, a wontfix, or what?

> Just a "close" with no other status, I think.  :-)

OK.  This post is closing the bug.

Thanks again for the help.

> -- 
> (domestic pets only, the antidote for overdose, milk.)
>    bloggy blog: http://lars.ingebrigtsen.no

-- 
Alan Mackenzie (Nuremberg, Germany).