A more radical cleanup for make bootstrap?

A more radical cleanup for make bootstrap?

[Gregory Heytings]

[Subject changed so that potentially interested readers do not miss it.]

>> WDYT of simply adding another "full-bootstrap" target, with the meaning 
>> "from the lowest possible ground"?
>
> If other people are so married to the "bootstrap" name that they refuse 
> to consider renaming it, I don't think it's worth spending more time 
> arguing about.  So, sure.
>

Now Stefan M and Po Lu both said (IIUC) that what they'd like is a 
bootstrap that does a less radical cleanup than what bootstrap now does. 
And (again IIUC) Eli and you (and others) agree that bootstrap should 
start with a state that is as clean as possible (which is what extraclean 
does).

So what I'd suggest is to use extraclean in bootstrap, and add either 
another target like "soft-bootstrap" or a configuration variable like 
"SOFT=true" that would make a less radical cleanup than what 
bootstrap-cleanup now does.

Would that be a solution that is acceptable to everyone?

Re: A more radical cleanup for make bootstrap?

[Lars Ingebrigtsen]

Gregory Heytings <gregory@heytings.org> writes:

> So what I'd suggest is to use extraclean in bootstrap, and add either
> another target like "soft-bootstrap" or a configuration variable like
> "SOFT=true" that would make a less radical cleanup than what
> bootstrap-cleanup now does.
>
> Would that be a solution that is acceptable to everyone?

Sounds good to me.

Re: A more radical cleanup for make bootstrap?

[Alan Mackenzie]

Hello, Gregory.

On Mon, Sep 19, 2022 at 19:46:55 +0000, Gregory Heytings wrote:

> [Subject changed so that potentially interested readers do not miss it.]

Thanks!

> >> WDYT of simply adding another "full-bootstrap" target, with the meaning 
> >> "from the lowest possible ground"?

> > If other people are so married to the "bootstrap" name that they refuse 
> > to consider renaming it, I don't think it's worth spending more time 
> > arguing about.  So, sure.


> Now Stefan M and Po Lu both said (IIUC) that what they'd like is a 
> bootstrap that does a less radical cleanup than what bootstrap now does. 
> And (again IIUC) Eli and you (and others) agree that bootstrap should 
> start with a state that is as clean as possible (which is what extraclean 
> does).

The prime quality of the current bootstrap is that it doesn't delete
anything unrecoverable - it only deletes generated files.

> So what I'd suggest is to use extraclean in bootstrap, and add either 
> another target like "soft-bootstrap" or a configuration variable like 
> "SOFT=true" that would make a less radical cleanup than what 
> bootstrap-cleanup now does.

It's not clear from the comments what extraclean does.  The comments
just say it does "delete backup and autosave files, too", without
specifying the files whose backup and autosave files get deleted.

> Would that be a solution that is acceptable to everyone?

If I understand correctly, the proposed bootstrap would irrevocably
delete unrecoverable files.  I don't think this should be done as a side
effect of something else, rather users who want extraclean should have
to say so explicitly, as a safety measure.

Again, if I understand correctly, I would type make bootstrap out of
sheer finger memory and lose lots of files.  I'd be unhappy about that.

Why not, instead, leave bootstrap as it is (possibly removing the call
to ./configure from it) and give the new target a new name, such as
bootstrap-hard?

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: A more radical cleanup for make bootstrap?

[Lars Ingebrigtsen]

Alan Mackenzie <acm@muc.de> writes:

> The prime quality of the current bootstrap is that it doesn't delete
> anything unrecoverable - it only deletes generated files.

But it does not delete all generated files, which is the problem.  Try a
"make bootstrap-clean" then "git clean -n -dfx" (note -- -n for dry run)
to see the ones left behind.

Re: A more radical cleanup for make bootstrap?

[Alan Mackenzie]

Hello, Lars.

On Mon, Sep 19, 2022 at 22:29:09 +0200, Lars Ingebrigtsen wrote:
> Alan Mackenzie <acm@muc.de> writes:

> > The prime quality of the current bootstrap is that it doesn't delete
> > anything unrecoverable - it only deletes generated files.

> But it does not delete all generated files, which is the problem.  Try a
> "make bootstrap-clean" then "git clean -n -dfx" (note -- -n for dry run)
> to see the ones left behind.

If the choice is between failing to delete some generated files, and
deleting unrecoverable files, I'd take the first choice.

Is that the choice we're talking about, or are we talking about
deleting all generated files and no others?

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: A more radical cleanup for make bootstrap?

[Lars Ingebrigtsen]

Alan Mackenzie <acm@muc.de> writes:

> Is that the choice we're talking about, or are we talking about
> deleting all generated files and no others?

The latter.

Re: A more radical cleanup for make bootstrap?

[Stefan Monnier]

> Would that be a solution that is acceptable to everyone?

To be honest, I can live with a "bootstrap" that does a thorough
build from scratch (and cook up my own "minimal bootstrap").
I think it's technically incorrect to call it "bootstrap", but it's not
like it matters.

I was just assuming there are other people in my situation, but I don't
know it to be the case.


        Stefan

Re: A more radical cleanup for make bootstrap?

[Alan Mackenzie]

Hello, Lars.

On Mon, Sep 19, 2022 at 22:51:36 +0200, Lars Ingebrigtsen wrote:
> Alan Mackenzie <acm@muc.de> writes:

> > Is that the choice we're talking about, or are we talking about
> > deleting all generated files and no others?

> The latter.

Oh, good!

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: A more radical cleanup for make bootstrap?

[Gregory Heytings]

>
> It's not clear from the comments what extraclean does.  The comments 
> just say it does "delete backup and autosave files, too", without 
> specifying the files whose backup and autosave files get deleted.
>

extraclean removes all generated files, and also all *~ and #* files.  If 
removing these files is a potential problem, they could perhaps be kept by 
bootstrap (or be kept conditionally to some variable).  But note that *~ 
files are by default not created by Emacs, unless you've set 
vc-make-backup-files to t in your init file.

Re: A more radical cleanup for make bootstrap?

[Eli Zaretskii]

> Date: Mon, 19 Sep 2022 19:46:55 +0000
> From: Gregory Heytings <gregory@heytings.org>
> cc: Po Lu <luangruo@yahoo.com>, Stefan Monnier <monnier@iro.umontreal.ca>, 
>  Stefan Kangas <stefan@marxist.se>, emacs-devel@gnu.org
> 
> So what I'd suggest is to use extraclean in bootstrap, and add either 
> another target like "soft-bootstrap" or a configuration variable like 
> "SOFT=true" that would make a less radical cleanup than what 
> bootstrap-cleanup now does.

What would be the purpose of this "soft-bootstrap"?

Re: A more radical cleanup for make bootstrap?

[Eli Zaretskii]

> Date: Mon, 19 Sep 2022 20:49:23 +0000
> Cc: Gregory Heytings <gregory@heytings.org>, Po Lu <luangruo@yahoo.com>,
>  Stefan Monnier <monnier@iro.umontreal.ca>,
>  Stefan Kangas <stefan@marxist.se>, emacs-devel@gnu.org
> From: Alan Mackenzie <acm@muc.de>
> 
> If the choice is between failing to delete some generated files, and
> deleting unrecoverable files, I'd take the first choice.

What are "unrecoverable files" in this context, please?  Can you give
a couple of examples of such files?

Re: A more radical cleanup for make bootstrap?

[Po Lu]

Eli Zaretskii <eliz@gnu.org> writes:

> What are "unrecoverable files" in this context, please?  Can you give
> a couple of examples of such files?

I think Alan is referring to auto-save files, which make extraclean
deletes.

Re: A more radical cleanup for make bootstrap?

[Alfred M. Szmidt]

   > It's not clear from the comments what extraclean does.  The comments 
   > just say it does "delete backup and autosave files, too", without 
   > specifying the files whose backup and autosave files get deleted.

   extraclean removes all generated files, and also all *~ and #* files.  If 
   removing these files is a potential problem, they could perhaps be kept by 
   bootstrap (or be kept conditionally to some variable).  But note that *~ 
   files are by default not created by Emacs, unless you've set 
   vc-make-backup-files to t in your init file.

vc-make-backup-files only applies for files under version control,
Non-version controlled files get backup files created by default.

So if you have a file, not under version control, its backup files
might get removed.  That extraclean deletes files it does not know
about (*~ and #*) seems like a mis-feature.

Re: A more radical cleanup for make bootstrap?

[Gregory Heytings]

>> So what I'd suggest is to use extraclean in bootstrap, and add either 
>> another target like "soft-bootstrap" or a configuration variable like 
>> "SOFT=true" that would make a less radical cleanup than what 
>> bootstrap-cleanup now does.
>
> What would be the purpose of this "soft-bootstrap"?
>

I don't know exactly.  Stefan M and Po Lu both asked for a "less radical" 
bootstrap, which is what that "soft-bootstrap" would have done.  But now 
Stefan said it's not that important to him.  So I guess there's no need 
for a "soft-bootstrap" anymore, unless someone tells otherwise.

Re: A more radical cleanup for make bootstrap?

[Po Lu]

Gregory Heytings <gregory@heytings.org> writes:

> I don't know exactly.  Stefan M and Po Lu both asked for a "less
> radical" bootstrap, which is what that "soft-bootstrap" would have
> done.  But now Stefan said it's not that important to him.  So I guess
> there's no need for a "soft-bootstrap" anymore, unless someone tells
> otherwise.

What about me?

It generally takes 10-15 minutes for me to build Emacs in ideal
conditions from "make bootstrap", with much of that time taken up by
building temacs.  It takes longer when I am actually working on Emacs,
because that typically happens while my machine is already under load.

Re: A more radical cleanup for make bootstrap?

[Gregory Heytings]

>> extraclean removes all generated files, and also all *~ and #* files. 
>> If removing these files is a potential problem, they could perhaps be 
>> kept by bootstrap (or be kept conditionally to some variable).  But 
>> note that *~ files are by default not created by Emacs, unless you've 
>> set vc-make-backup-files to t in your init file.
>
> vc-make-backup-files only applies for files under version control, 
> Non-version controlled files get backup files created by default.
>
> So if you have a file, not under version control, its backup files might 
> get removed.  That extraclean deletes files it does not know about (*~ 
> and #*) seems like a mis-feature.
>

After thinking a bit more about this, IMO these backup and autosaved files 
can be kept by the new bootstrap-clean, given that it seems unlikely that 
these files have an effect on the build, or at least that it is not more 
likely that they could impact the build more than user-created files 
(which aren't deleted by extraclean either).

Re: A more radical cleanup for make bootstrap?

[Gregory Heytings]

>> I don't know exactly.  Stefan M and Po Lu both asked for a "less 
>> radical" bootstrap, which is what that "soft-bootstrap" would have 
>> done.  But now Stefan said it's not that important to him.  So I guess 
>> there's no need for a "soft-bootstrap" anymore, unless someone tells 
>> otherwise.
>
> What about me?
>

You could be that "someone".

>
> It generally takes 10-15 minutes for me to build Emacs in ideal 
> conditions from "make bootstrap", with much of that time taken up by 
> building temacs.  It takes longer when I am actually working on Emacs, 
> because that typically happens while my machine is already under load.
>

So what you'd need is another cleaning target, is that correct?

Re: A more radical cleanup for make bootstrap?

[Po Lu]

Gregory Heytings <gregory@heytings.org> writes:

> So what you'd need is another cleaning target, is that correct?

Yes, and I think it makes slightly more sense to call it "bootstrap"
than to call it something else.

Re: A more radical cleanup for make bootstrap?

[Gregory Heytings]

>> So what you'd need is another cleaning target, is that correct?
>
> Yes,
>

Okay.  What would that cleaning target do?  Remove everything except 
temacs?

>
> and I think it makes slightly more sense to call it "bootstrap" than to 
> call it something else.
>

I don't think its name matters much.

"What's in a name? That which we call a rose
By any other name would smell as sweet."

Re: A more radical cleanup for make bootstrap?

[Alan Mackenzie]

Hello, Eli.

On Tue, Sep 20, 2022 at 05:35:36 +0300, Eli Zaretskii wrote:
> > Date: Mon, 19 Sep 2022 20:49:23 +0000
> > Cc: Gregory Heytings <gregory@heytings.org>, Po Lu <luangruo@yahoo.com>,
> >  Stefan Monnier <monnier@iro.umontreal.ca>,
> >  Stefan Kangas <stefan@marxist.se>, emacs-devel@gnu.org
> > From: Alan Mackenzie <acm@muc.de>
> > 
> > If the choice is between failing to delete some generated files, and
> > deleting unrecoverable files, I'd take the first choice.

> What are "unrecoverable files" in this context, please?  Can you give
> a couple of examples of such files?

I was thinking of xdisp.c~ and simple.el~.  Also, perhaps, myfile.el.

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: A more radical cleanup for make bootstrap?

[Eli Zaretskii]

> From: Po Lu <luangruo@yahoo.com>
> Cc: Alan Mackenzie <acm@muc.de>,  larsi@gnus.org,  gregory@heytings.org,
>   monnier@iro.umontreal.ca,  stefan@marxist.se,  emacs-devel@gnu.org
> Date: Tue, 20 Sep 2022 10:42:03 +0800
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > What are "unrecoverable files" in this context, please?  Can you give
> > a couple of examples of such files?
> 
> I think Alan is referring to auto-save files, which make extraclean
> deletes.

If those files can never cause problems for a clean build, maybe we
should instead enhance bootstrap-clean and use that?

Re: A more radical cleanup for make bootstrap?

[Eli Zaretskii]

> From: Po Lu <luangruo@yahoo.com>
> Cc: Eli Zaretskii <eliz@gnu.org>,  larsi@gnus.org,
>   monnier@iro.umontreal.ca,  stefan@marxist.se,  emacs-devel@gnu.org
> Date: Tue, 20 Sep 2022 16:20:06 +0800
> 
> Gregory Heytings <gregory@heytings.org> writes:
> 
> > I don't know exactly.  Stefan M and Po Lu both asked for a "less
> > radical" bootstrap, which is what that "soft-bootstrap" would have
> > done.  But now Stefan said it's not that important to him.  So I guess
> > there's no need for a "soft-bootstrap" anymore, unless someone tells
> > otherwise.
> 
> What about me?
> 
> It generally takes 10-15 minutes for me to build Emacs in ideal
> conditions from "make bootstrap", with much of that time taken up by
> building temacs.  It takes longer when I am actually working on Emacs,
> because that typically happens while my machine is already under load.

Why do you bootstrap?

I almost never do that.

Re: A more radical cleanup for make bootstrap?

[Po Lu]

Eli Zaretskii <eliz@gnu.org> writes:

>> It generally takes 10-15 minutes for me to build Emacs in ideal
>> conditions from "make bootstrap", with much of that time taken up by
>> building temacs.  It takes longer when I am actually working on Emacs,
>> because that typically happens while my machine is already under load.

> Why do you bootstrap?

Because the incremental build fails and I cannot get it to work by
eyeballing what elc files to delete, usually due to an obscure
macro-expansion error in tramp.

Re: A more radical cleanup for make bootstrap?

[Eli Zaretskii]

> From: Po Lu <luangruo@yahoo.com>
> Cc: gregory@heytings.org,  larsi@gnus.org,  monnier@iro.umontreal.ca,
>   stefan@marxist.se,  emacs-devel@gnu.org
> Date: Tue, 20 Sep 2022 19:55:20 +0800
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> >> It generally takes 10-15 minutes for me to build Emacs in ideal
> >> conditions from "make bootstrap", with much of that time taken up by
> >> building temacs.  It takes longer when I am actually working on Emacs,
> >> because that typically happens while my machine is already under load.
> 
> > Why do you bootstrap?
> 
> Because the incremental build fails and I cannot get it to work by
> eyeballing what elc files to delete, usually due to an obscure
> macro-expansion error in tramp.

I usually have no problems identifying the offending *.elc files.  And
if that fails (which it does very rarely), then this never fails:

  $ find ./lisp -name "*.elc" -delete
  $ make

Re: A more radical cleanup for make bootstrap?

[Po Lu]

Eli Zaretskii <eliz@gnu.org> writes:

> I usually have no problems identifying the offending *.elc files.

Same, but when the error is a macroexpansion error in Tramp (or
something related to loaddefs.el), I almost universally fail.

> And if that fails (which it does very rarely), then this never fails:
>
>   $ find ./lisp -name "*.elc" -delete
>   $ make

Right, why don't we make "make bootstrap" do that instead?

Re: A more radical cleanup for make bootstrap?

[Eli Zaretskii]

> From: Po Lu <luangruo@yahoo.com>
> Cc: gregory@heytings.org,  larsi@gnus.org,  monnier@iro.umontreal.ca,
>   stefan@marxist.se,  emacs-devel@gnu.org
> Date: Tue, 20 Sep 2022 20:51:20 +0800
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > I usually have no problems identifying the offending *.elc files.
> 
> Same, but when the error is a macroexpansion error in Tramp (or
> something related to loaddefs.el), I almost universally fail.
> 
> > And if that fails (which it does very rarely), then this never fails:
> >
> >   $ find ./lisp -name "*.elc" -delete
> >   $ make
> 
> Right, why don't we make "make bootstrap" do that instead?

Because some (much rarer) changes need other generated files to be
removed as well.  So bootstrap-clean does the above, but it also does
other things, including forcing the rebuild of temacs, which is almost
never needed.  OTOH, if someone wants to test a fresh build, or build
a different configuration, or some other relatively rare jobs, they do
need to remove all those other artifacts.

Re: A more radical cleanup for make bootstrap?

[Lars Ingebrigtsen]

Po Lu <luangruo@yahoo.com> writes:

>>   $ find ./lisp -name "*.elc" -delete
>>   $ make
>
> Right, why don't we make "make bootstrap" do that instead?

If that is something that people need to do a lot, we should instead
introduce a new target like "clean-elc".

Re: A more radical cleanup for make bootstrap?

[Po Lu]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> If that is something that people need to do a lot, we should instead
> introduce a new target like "clean-elc".

Didn't Gerd say deleting all the checked in *.elc files was what "make
bootstrap" was originally intended for, though?

Re: A more radical cleanup for make bootstrap?

[Lars Ingebrigtsen]

Po Lu <luangruo@yahoo.com> writes:

> Didn't Gerd say deleting all the checked in *.elc files was what "make
> bootstrap" was originally intended for, though?

It's developed over the years to do more, and we've been saying "have
you tried 'make bootstrap'?" to people for decades now for all kinds of
reasons.  We shouldn't change that, since that's what people will find
when googling.

Re: A more radical cleanup for make bootstrap?

[Eli Zaretskii]

> From: Po Lu <luangruo@yahoo.com>
> Cc: Eli Zaretskii <eliz@gnu.org>,  gregory@heytings.org,
>   monnier@iro.umontreal.ca,  stefan@marxist.se,  emacs-devel@gnu.org
> Date: Tue, 20 Sep 2022 21:27:18 +0800
> 
> Lars Ingebrigtsen <larsi@gnus.org> writes:
> 
> > If that is something that people need to do a lot, we should instead
> > introduce a new target like "clean-elc".
> 
> Didn't Gerd say deleting all the checked in *.elc files was what "make
> bootstrap" was originally intended for, though?

Originally, only the *.elc files needed to be removed, yes.

Re: A more radical cleanup for make bootstrap?

[Robert Pluim]

>>>>> On Tue, 20 Sep 2022 15:05:50 +0300, Eli Zaretskii <eliz@gnu.org> said:


    Eli> I usually have no problems identifying the offending *.elc files.  And
    Eli> if that fails (which it does very rarely), then this never fails:

    Eli>   $ find ./lisp -name "*.elc" -delete
    Eli>   $ make

I just did that because of a compile failure in ./lisp, it failed
again in ./lisp (but differently), but 'make bootstrap' worked. Go
figure.

I think itʼs going to be just 'make boostrap' for me from here on in.

Robert
-- 

Re: A more radical cleanup for make bootstrap?

[Eli Zaretskii]

> From: Po Lu <luangruo@yahoo.com>
> Cc: gregory@heytings.org,  larsi@gnus.org,  monnier@iro.umontreal.ca,
>  stefan@marxist.se,  emacs-devel@gnu.org
> Date: Tue, 20 Sep 2022 20:51:20 +0800
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > I usually have no problems identifying the offending *.elc files.
> 
> Same, but when the error is a macroexpansion error in Tramp (or
> something related to loaddefs.el), I almost universally fail.

When in doubt, delete all the tramp*.elc files.  If you think it's
loaddefs.el, copy ldefs-boot.el over it and delete loaddefs.elc.

Re: A more radical cleanup for make bootstrap?

[Eli Zaretskii]

> From: Robert Pluim <rpluim@gmail.com>
> Cc: Po Lu <luangruo@yahoo.com>,  gregory@heytings.org,  larsi@gnus.org,
>   monnier@iro.umontreal.ca,  stefan@marxist.se,  emacs-devel@gnu.org
> Date: Tue, 20 Sep 2022 17:43:20 +0200
> 
> >>>>> On Tue, 20 Sep 2022 15:05:50 +0300, Eli Zaretskii <eliz@gnu.org> said:
> 
> 
>     Eli> I usually have no problems identifying the offending *.elc files.  And
>     Eli> if that fails (which it does very rarely), then this never fails:
> 
>     Eli>   $ find ./lisp -name "*.elc" -delete
>     Eli>   $ make
> 
> I just did that because of a compile failure in ./lisp, it failed
> again in ./lisp (but differently)

When that happens, delete temacs and bootstrap-emacs as well.

Re: A more radical cleanup for make bootstrap?

[Michael Welsh Duggan]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Po Lu <luangruo@yahoo.com>
>> Cc: Eli Zaretskii <eliz@gnu.org>,  larsi@gnus.org,
>>   monnier@iro.umontreal.ca,  stefan@marxist.se,  emacs-devel@gnu.org
>> Date: Tue, 20 Sep 2022 16:20:06 +0800
>> 
>> Gregory Heytings <gregory@heytings.org> writes:
>> 
>> > I don't know exactly.  Stefan M and Po Lu both asked for a "less
>> > radical" bootstrap, which is what that "soft-bootstrap" would have
>> > done.  But now Stefan said it's not that important to him.  So I guess
>> > there's no need for a "soft-bootstrap" anymore, unless someone tells
>> > otherwise.
>> 
>> What about me?
>> 
>> It generally takes 10-15 minutes for me to build Emacs in ideal
>> conditions from "make bootstrap", with much of that time taken up by
>> building temacs.  It takes longer when I am actually working on Emacs,
>> because that typically happens while my machine is already under load.
>
> Why do you bootstrap?
>
> I almost never do that.

Personally, I almost always bootstrap.  As I am not actively doing Emacs
development, I am building once every few weeks in order to pick up new
features and bug fixes from the master branch.  There have been many
instances in the past where not bootstrapping causes strange errors when
running Emacs.  It doesn't happen frequently, but it does happen.  It is
worth an extra five minutes (?) or so to me to bootstrap always so I
don't have to deal with noticing that something isn't working correctly,
trying to track down the failure, and then later find out from reading
this list or by checking bug reports that bootstrapping fixes the issue.

-- 
Michael Welsh Duggan
(md5i@md5i.com)

Re: A more radical cleanup for make bootstrap?

[Eli Zaretskii]

> From: Michael Welsh Duggan <mwd@md5i.com>
> Cc: emacs-devel@gnu.org
> Date: Tue, 20 Sep 2022 12:14:32 -0400
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> >> What about me?
> >> 
> >> It generally takes 10-15 minutes for me to build Emacs in ideal
> >> conditions from "make bootstrap", with much of that time taken up by
> >> building temacs.  It takes longer when I am actually working on Emacs,
> >> because that typically happens while my machine is already under load.
> >
> > Why do you bootstrap?
> >
> > I almost never do that.
> 
> Personally, I almost always bootstrap.

That's fine, but my question was specifically to Po Lu (and people who
work on Emacs like he does), who builds Emacs much more frequently
than you do.  Your case is different, and the existing"make bootstrap"
covers it well enough, AFAIU.

Re: A more radical cleanup for make bootstrap?

[Michael Albinus]

Po Lu <luangruo@yahoo.com> writes:

Hi,

>> I usually have no problems identifying the offending *.elc files.
>
> Same, but when the error is a macroexpansion error in Tramp (or
> something related to loaddefs.el), I almost universally fail.

Deleting lisp/net/tramp-loaddefs.el* might help sometimes.

Best regards, Michael.

Re: A more radical cleanup for make bootstrap?

[Po Lu]

Michael Albinus <michael.albinus@gmx.de> writes:

> Deleting lisp/net/tramp-loaddefs.el* might help sometimes.

Thanks, but I already try that and it typically does not work.