Building Emacs using `PKGBUILD`, tests failing when doing `make check`

Haha, no wonder yours is taking so long – you’re building four different versions of Emacs!

I’m not - just not very familiar.

Let me waste some time by perusing the original - and then run the tests :wink:

Before I did this I spent a long time studying the whole makepkg , PKGBUILD stuff. I even built test packages to make sure I was doing it right. This is a helpful place to start, from that I moved on to making tarballs for test scripts and installing them as well, before I moved on to my Emacs build.

… so you are not worried about the failing test, but about the PKGBUILD script not continuing because of it?

that’s why I want to first build it - then run the tests … :wink:

it’s a computer - it is running anyway - it’ll be done eventually, even if now compiling stuff you do not even want
(nor I :sunglasses: )

I would be content with emacs from repos (no compiling) - but what do I know about emacs … :wink:
I decided against using it a long time ago.

1 Like

I know.
I’m just interested in his problem with the tests - not in the slightest in emacs.

… the thing is still compiling, about 1.5 hours later
(a VM with no swap, 4 GB RAM and Intel Core i3-3110M CPU)

that has got to be some powerful editor :stuck_out_tongue_winking_eye:



ps:
it took more than two and a half hours on this machine
compiling a kernel is way faster - and likely even a browser like Firefox
(I will not try to verify - those days are over)

But then again:
Emacs is like a whole OS all in itself - everything you need in one program :nerd_face:



@nonreligious

make check on the basic emacs build took like 10 minutes

the result:
make[3]: Leaving directory '/home/jo/emacs/src/emacs-29.4/test'
make[2]: [Makefile:340: check-doit] Error 2 (ignored)

SUMMARY OF TEST RESULTS
-----------------------
Files examined: 471
Ran 7031 tests, 6849 results as expected, 2 unexpected, 180 skipped
2 files contained unexpected results:
  lisp/progmodes/flymake-tests.log
  lisp/gnus/mml-sec-tests.log
make[2]: *** [Makefile:341: check-doit] Error 1
make[2]: Leaving directory '/home/jo/emacs/src/emacs-29.4/test'
make[1]: *** [Makefile:310: check] Error 2
make[1]: Leaving directory '/home/jo/emacs/src/emacs-29.4/test'
make: *** [Makefile:1105: check] Error 2

and

`make check` on the ...nativecomp build yielded these results:
make[3]: Leaving directory '/home/jo/emacs/src/emacs-29.4-nativecomp/test'
make[2]: [Makefile:340: check-doit] Error 2 (ignored)

SUMMARY OF TEST RESULTS
-----------------------
Files examined: 471
Ran 7198 tests, 7020 results as expected, 1 unexpected, 177 skipped
1 files contained unexpected results:
  lisp/progmodes/flymake-tests.log
make[2]: *** [Makefile:341: check-doit] Error 1
make[2]: Leaving directory '/home/jo/emacs/src/emacs-29.4-nativecomp/test'
make[1]: *** [Makefile:310: check] Error 2
make[1]: Leaving directory '/home/jo/emacs/src/emacs-29.4-nativecomp/test'
make: *** [Makefile:1105: check] Error 2

You misplaced your
make check
function in the PKGBUILD :man_shrugging:

I ran the
make check
after the build was finished - but before install
(I didn’t install the final product … I can provide the log files if you want - they are not large)



But no real errors in either one - haha :sunglasses:

OK, this is now less than helpful – why don’t you read the PKGBUILD documentation? Or look at the PKGBUILD.proto file?

It’s because he’s building 4 different versions of the package at the same time!

1 Like

FWIW, there are certain combinations of configuration options that I want that aren’t available. You wouldn’t understand.

You asked whether this was a problem.
I hope you have have found your answer.

That is your assessment? o.k.

I’ll make sure not to bother you again.
Cheers!

I won’t bother either. Just had my second strike on here; guess this post will result in my third. If that is the case … Goodbye and thanks for all the fish.

1 Like

A strike?
I didn’t know that there even was such a thing.

From what I remember (your post is no longer there)
there wasn’t anything that could be considered even inflammatory by anyone -
and if that, it was towards me.
But I knew, of course, that that was said in jest. :wink:

@nonreligious

What takes so incredibly long in compilation is the configuration option:

--with-native-compilation=aot
(aot = ahead of time … everything …)

I have compiled the thing again with your (@nonreligious) config options

./configure --sysconfdir=/etc --prefix=/usr --libexecdir=/usr/lib --localstatedir=/var --with-cairo --with-harfbuzz --with-libsystemd --with-modules --with-x-toolkit=gtk3 --with-xwidgets --with-imagemagick --with-mailutils --with-native-compilation=aot --with-tree-sitter

and these where the results of the
make check:

SUMMARY OF TEST RESULTS
-----------------------
Files examined: 471
Ran 7198 tests, 7020 results as expected, 2 unexpected, 176 skipped
2 files contained unexpected results:
  src/emacs-tests.log
  lisp/progmodes/flymake-tests.log
make[2]: *** [Makefile:341: check-doit] Error 1
make[2]: Leaving directory '/home/jo/emacs2/emacs-29.4/test'
make[1]: *** [Makefile:310: check] Error 2
make[1]: Leaving directory '/home/jo/emacs2/emacs-29.4/test'
make: *** [Makefile:1105: check] Error 2

… complaining about that missing audit file

… this is all just silly bruised ego stuff, pretending to know things which don’t survive under scrutiny :wink:
(not on your part @BG405 )

1 Like

I have experience with this,

You can skip tests and deleted the whole check()

Arch usually but not always perform tests (and report them and possibly fix them)

1 Like

… he intentionally added it to his build (his PKGBUILD)
It wasn’t there initially.

… and then he wondered about why it failed …
(actually: it didn’t fail - not for me
the check just returned an error re a non existing /var/log/audit.log)

1 Like

I add tests to pratically all my pkgbuilds (and sometimes to others) to see how it goes, not always they all pass

1 Like

of course - good practice

but he apparently was concerned about this one result
which isn’t something to be concerned about …

1 Like
Off-topic rant

If I have a post redacted by the Forum system, then I have committed some sort of infraction for this to happen. This includes things such as “offensive”, “inappropriate”, “spam” etc. and this is now on my record. On some forums: 3 “strikes” and you are out. I’ve made thousands of posts here without this happening; but two in one day is rather disconcerting.

It stings a bit when you are just trying to help somebody and be friendly in the process.

Sorry for this off-topic rant.

Back on track: Why doesn’t the OP do what we suggested and see if this actually works? Post 2 onwards contain decent ideas IMHO.

1 Like

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.