GNU bug report logs - #14704
24.3.50; cl-lib breaks built-in Emacs version

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 14704 in the body.
You can then email your comments to 14704 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Sebastian Wiesner <lunaryorn <at> gmail.com>
To: bug-gnu-emacs <at> gnu.org
Subject: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 17:30:47 +0200

"(require 'cl-lib)" breaks the Emacs version information for
package.el.  In "emacs -Q", "M-x ielm":

ELISP> (package-initialize)
t
ELISP> (package-built-in-p 'emacs (version-to-list "24.1"))
t
ELISP> (assq 'emacs package--builtin-versions)
(emacs 24 3 50)

ELISP> (require 'cl-lib)
cl-lib
ELISP> (package-built-in-p 'emacs (version-to-list "24.1"))
nil
ELISP> (assq 'emacs package--builtin-versions)
(emacs 2 2)

This breaks dependency resolution for packages which depend against a
certain Emacs version.

These incorrect entries in package--builtin-versions come from
"cl-loaddefs.el", which contains two instances of the following line:

(push (purecopy (quote (emacs 2 2))) package--builtin-versions)

These lines are apparently extracted from the package headers of
"cl-macs.el" and "cl-seq.el" which look like the following:

;; Author: Dave Gillespie <daveg <at> synaptics.com>
;; Version: 2.02
;; Keywords: extensions
;; Package: emacs

This meta information is obviously wrong.

This is the 3rd critical package.el bug I discovered within just a week
or so.  Don't you test your code?!

Message #8 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Juanma Barranquero <lekktu <at> gmail.com>
To: Sebastian Wiesner <lunaryorn <at> gmail.com>
Cc: 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 17:48:05 +0200

On Mon, Jun 24, 2013 at 5:30 PM, Sebastian Wiesner <lunaryorn <at> gmail.com> wrote:

> This is the 3rd critical package.el bug I discovered within just a week
> or so.  Don't you test your code?!

If you want well-tested code, stay with Emacs official releases. If
you compile from trunk, expect breakages, untested code and potential
data loss. Caveat user.

Message #11 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Glenn Morris <rgm <at> gnu.org>
To: Sebastian Wiesner <lunaryorn <at> gmail.com>
Cc: 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 13:39:24 -0400

Sebastian Wiesner wrote:

> Don't you test your code?!

I think you've got a point, so I started a discussion:

http://lists.gnu.org/archive/html/emacs-devel/2013-06/msg00995.html

Message #14 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Sebastian Wiesner <lunaryorn <at> gmail.com>
To: Juanma Barranquero <lekktu <at> gmail.com>
Cc: 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 19:51:20 +0200

2013/6/24 Juanma Barranquero <lekktu <at> gmail.com>:
> On Mon, Jun 24, 2013 at 5:30 PM, Sebastian Wiesner <lunaryorn <at> gmail.com> wrote:
>
>> This is the 3rd critical package.el bug I discovered within just a week
>> or so.  Don't you test your code?!
>
> If you want well-tested code, stay with Emacs official releases. If
> you compile from trunk, expect breakages, untested code and potential
> data loss. Caveat user.

Who would test your code then before releases if not those who compile
from trunk?  You developers obviously don't…  and what is worse, you
apparently do even *care*.

These three bugs I found are so blatantly obvious and so severe that
even the most sloppy manual testing, let alone unit testing, would
have revealed them, but all you have to say to such a bad failure is
“Boy, this is trunk, go home to mama, and play with old releases”?!

I am sorry that you do not know about CI, unit testing and software
quality, but I do, and thus I am horrified by the neglect and
ignorance you demonstrate in developing one of the most important
tools of my daily work.

Message #17 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Sebastian Wiesner <lunaryorn <at> gmail.com>
To: Glenn Morris <rgm <at> gnu.org>
Cc: 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 19:52:37 +0200

2013/6/24 Glenn Morris <rgm <at> gnu.org>:
> Sebastian Wiesner wrote:
>
>> Don't you test your code?!
>
> I think you've got a point, so I started a discussion:
>
> http://lists.gnu.org/archive/html/emacs-devel/2013-06/msg00995.html

Thank you for taking my concerns seriously.

Message #20 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Juanma Barranquero <lekktu <at> gmail.com>
To: Sebastian Wiesner <lunaryorn <at> gmail.com>
Cc: 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 19:56:00 +0200

On Mon, Jun 24, 2013 at 7:51 PM, Sebastian Wiesner <lunaryorn <at> gmail.com> wrote:

> Who would test your code then before releases if not those who compile
> from trunk?  You developers obviously don't…  and what is worse, you
> apparently do even *care*.

Yes, users (and developers) test it. Users file bug reports. Sometimes
they offer a patch, often they don't. That's fine and good. Users
don't usually come screaming and demanding. That seems rude, to say
the least.

> These three bugs I found are so blatantly obvious and so severe that
> even the most sloppy manual testing, let alone unit testing, would
> have revealed them, but all you have to say to such a bad failure is
> “Boy, this is trunk, go home to mama, and play with old releases”?!

No, more like "go home to mama and ask her again about good manners".

> I am sorry that you do not know about CI, unit testing and software
> quality, but I do, and thus I am horrified by the neglect and
> ignorance you demonstrate in developing one of the most important
> tools of my daily work.

Not sure if that you is singular or plural. It's wrong in either case.

   J

Message #23 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Sebastian Wiesner <lunaryorn <at> gmail.com>
Cc: lekktu <at> gmail.com, 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 21:16:10 +0300

> Date: Mon, 24 Jun 2013 19:51:20 +0200
> From: Sebastian Wiesner <lunaryorn <at> gmail.com>
> Cc: 14704 <at> debbugs.gnu.org
> 
> 2013/6/24 Juanma Barranquero <lekktu <at> gmail.com>:
> > On Mon, Jun 24, 2013 at 5:30 PM, Sebastian Wiesner <lunaryorn <at> gmail.com> wrote:
> >
> >> This is the 3rd critical package.el bug I discovered within just a week
> >> or so.  Don't you test your code?!
> >
> > If you want well-tested code, stay with Emacs official releases. If
> > you compile from trunk, expect breakages, untested code and potential
> > data loss. Caveat user.
> 
> Who would test your code then before releases if not those who compile
> from trunk?

Each release is preceded by a long pretest period, when we are trying
to collect usage experiences from as many different platforms and
users as possible.

> You developers obviously don't… and what is worse, you apparently do
> even *care*.

I don't think this is a fair comment.  A released version is more
stable than a development snapshot, that's a fact.  It has nothing to
do with caring or not.  It also doesn't only happen due to more
testing: during a pretest, only bugfixes are allowed into the code
base, so potentially destabilizing changes are avoided.

> I am sorry that you do not know about CI, unit testing and software
> quality, but I do, and thus I am horrified by the neglect and
> ignorance you demonstrate in developing one of the most important
> tools of my daily work.

While more testing is always a good thing, some areas in Emacs are
hard to test.  Please consider coming on board and moving Emacs closer
to the CI ideals and testability.  TIA.

Message #26 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Sebastian Wiesner <lunaryorn <at> gmail.com>
To: Juanma Barranquero <lekktu <at> gmail.com>
Cc: 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 20:26:27 +0200

2013/6/24 Juanma Barranquero <lekktu <at> gmail.com>:
> On Mon, Jun 24, 2013 at 7:51 PM, Sebastian Wiesner <lunaryorn <at> gmail.com> wrote:
>
>> Who would test your code then before releases if not those who compile
>> from trunk?  You developers obviously don't…  and what is worse, you
>> apparently do even *care*.
>
> Yes, users (and developers) test it. Users file bug reports. Sometimes
> they offer a patch, often they don't. That's fine and good. Users
> don't usually come screaming and demanding.

If I were a usual user, I would just have gone back to a release as
you said.  After all, your bugs and your lack of software testing are
not my problem in the end.

But instead I did you work: I debugged the problem, found the cause,
and reported it.  I am sorry that I did not provide a patch right
away, but I am unfamiliar with your code base, and have not signed any
copyright papers and no intention whatsoever of doing so.  I doubt
that any patch of mine would have been accepted.

I did neither scream nor demand, but I actually expected you to apply
care and attention to your work.  I write tests for my own code, and I
run them, and I expected you to do so as well.  I see now, that you do
not, which is embarrassing to you and very ignorant towards your
users.

By not writing or running tests for *your* code you have wasted *my*
time.  That is what I call “rude”.

>> These three bugs I found are so blatantly obvious and so severe that
>> even the most sloppy manual testing, let alone unit testing, would
>> have revealed them, but all you have to say to such a bad failure is
>> “Boy, this is trunk, go home to mama, and play with old releases”?!
>
> No, more like "go home to mama and ask her again about good manners".

Fine, let's go home than.  I'll ask ma about good manners, and you'll
ask your high school teacher about software testing.  We'll see who of
us learns more on this trip…

Message #29 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Sebastian Wiesner <lunaryorn <at> gmail.com>
Cc: lekktu <at> gmail.com, 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 21:38:40 +0300

> Date: Mon, 24 Jun 2013 20:26:27 +0200
> From: Sebastian Wiesner <lunaryorn <at> gmail.com>
> Cc: 14704 <at> debbugs.gnu.org
> 
> I did neither scream nor demand, but I actually expected you to apply
> care and attention to your work.  I write tests for my own code, and I
> run them, and I expected you to do so as well.  I see now, that you do
> not, which is embarrassing to you and very ignorant towards your
> users.

Out of fairness, you found 3 bugs in a single 1700-line package, and
made very far-fetching conclusions ("Don't you test your code?!") out
of that.  Emacs weighs in at 400K C lines and 1400K Lisp lines, so
with all due respect, accusing us in not testing our code based on
such a small sample is a bit too much.  If all that code were not
tested at all, it would never have worked at all.

Message #32 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Sebastian Wiesner <lunaryorn <at> gmail.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: lekktu <at> gmail.com, 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 20:40:44 +0200

2013/6/24 Eli Zaretskii <eliz <at> gnu.org>:
>> Date: Mon, 24 Jun 2013 19:51:20 +0200
>> From: Sebastian Wiesner <lunaryorn <at> gmail.com>
>> Cc: 14704 <at> debbugs.gnu.org
>>
>> 2013/6/24 Juanma Barranquero <lekktu <at> gmail.com>:
>> > On Mon, Jun 24, 2013 at 5:30 PM, Sebastian Wiesner <lunaryorn <at> gmail.com> wrote:
>> >
>> >> This is the 3rd critical package.el bug I discovered within just a week
>> >> or so.  Don't you test your code?!
>> >
>> > If you want well-tested code, stay with Emacs official releases. If
>> > you compile from trunk, expect breakages, untested code and potential
>> > data loss. Caveat user.
>>
>> Who would test your code then before releases if not those who compile
>> from trunk?
>
> Each release is preceded by a long pretest period, when we are trying
> to collect usage experiences from as many different platforms and
> users as possible.

>> You developers obviously don't… and what is worse, you apparently do
>> even *care*.
>
> I don't think this is a fair comment.

I know.  I was upset by his comment, and I still am, but I should have
taken a deep breath before replying.  I meant no offense, and I am
sorry.

>> I am sorry that you do not know about CI, unit testing and software
>> quality, but I do, and thus I am horrified by the neglect and
>> ignorance you demonstrate in developing one of the most important
>> tools of my daily work.
>
> While more testing is always a good thing, some areas in Emacs are
> hard to test.

package.el isn't such an area.

> Please consider coming on board and moving Emacs closer to the CI ideals and testability.  TIA.

I am sorry, but I won't, for two reasons: Bazaar, and copyright assignments.

Message #35 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Sebastian Wiesner <lunaryorn <at> gmail.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: lekktu <at> gmail.com, 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 20:57:04 +0200

2013/6/24 Eli Zaretskii <eliz <at> gnu.org>:
>> Date: Mon, 24 Jun 2013 20:26:27 +0200
>> From: Sebastian Wiesner <lunaryorn <at> gmail.com>
>> Cc: 14704 <at> debbugs.gnu.org
>>
>> I did neither scream nor demand, but I actually expected you to apply
>> care and attention to your work.  I write tests for my own code, and I
>> run them, and I expected you to do so as well.  I see now, that you do
>> not, which is embarrassing to you and very ignorant towards your
>> users.
>
> Out of fairness, you found 3 bugs in a single 1700-line package, and
> made very far-fetching conclusions ("Don't you test your code?!") out
> of that.

These three bugs weren't just stupid forgotten corner cases in some
remote and obscure feature that none uses.  They broke an essential
command—which worked well before—in a very obvious and very bad way.

These bugs should have been caught by testing before making a commit.
But they weren't, so in this specific case, I think I can quite
legitimately conclude that these specific changes were not tested at
all, or only very carelessly.

I understand that package.el is being refactored currently, but that
isn't an excuse imho.  On the contrary, every software development
course teaches about making refactorings in a separate branch and
never without unit tests.

But I admit that I should not have generalized this statement, and I
apologize to you and to everyone else whom I offended.

> If all that code were not tested at all, it would never have worked at all.

It is tested, for sure, but probably less by you, and more by your users.

Message #38 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Sebastian Wiesner <lunaryorn <at> gmail.com>
Cc: lekktu <at> gmail.com, 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 22:20:02 +0300

> Date: Mon, 24 Jun 2013 20:57:04 +0200
> From: Sebastian Wiesner <lunaryorn <at> gmail.com>
> Cc: lekktu <at> gmail.com, 14704 <at> debbugs.gnu.org
> 
> > If all that code were not tested at all, it would never have worked at all.
> 
> It is tested, for sure, but probably less by you, and more by your users.

It is tested both by developers and users.  Emacs has a myriad of
features, some of which are not even documented clearly.  No matter
how well and long one tests their own code, they can never test it in
all the possible combinations with the other features and use cases.

As for "less by you", please be careful, because someone might take
offense again.  E.g., I have a large body of unit tests for display
features, such as cursor positioning, which I run by hand every time I
touch any of the related code.

Message #41 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Glenn Morris <rgm <at> gnu.org>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: Sebastian Wiesner <lunaryorn <at> gmail.com>, lekktu <at> gmail.com,
 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 16:02:07 -0400

Eli Zaretskii wrote:

> E.g., I have a large body of unit tests for display features, such as
> cursor positioning, which I run by hand every time I touch any of the
> related code.

Do you feel like adding it to the repo?
Sounds like it would be useful to have in test/, even if not suitable
for test/automated/.

Message #44 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Glenn Morris <rgm <at> gnu.org>
Cc: lunaryorn <at> gmail.com, lekktu <at> gmail.com, 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 23:08:45 +0300

> From: Glenn Morris <rgm <at> gnu.org>
> Cc: Sebastian Wiesner <lunaryorn <at> gmail.com>,  lekktu <at> gmail.com,  14704 <at> debbugs.gnu.org
> Date: Mon, 24 Jun 2013 16:02:07 -0400
> 
> Eli Zaretskii wrote:
> 
> > E.g., I have a large body of unit tests for display features, such as
> > cursor positioning, which I run by hand every time I touch any of the
> > related code.
> 
> Do you feel like adding it to the repo?
> Sounds like it would be useful to have in test/, even if not suitable
> for test/automated/.

Who except myself would want to run, let's see... about 30 tests, by
hand, first in a GUI session, then in a TTY session?

Message #47 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Juanma Barranquero <lekktu <at> gmail.com>
To: Sebastian Wiesner <lunaryorn <at> gmail.com>
Cc: 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 22:18:00 +0200

On Mon, Jun 24, 2013 at 8:26 PM, Sebastian Wiesner <lunaryorn <at> gmail.com> wrote:

> But instead I did you work: I debugged the problem, found the cause,
> and reported it.  I am sorry that I did not provide a patch right
> away,

Nobody has dismissed your bug reports for not having a patch.

> and have not signed any
> copyright papers and no intention whatsoever of doing so.  I doubt
> that any patch of mine would have been accepted.

Until 10 lines or so, they could be accepted if they fix the bug and
are clearly written, as a "tiny change".

> I did neither scream nor demand

" Don't you test your code?! " counts as a scream and a demand in my view.

> By not writing or running tests for *your* code you have wasted *my*
> time.  That is what I call “rude”.

Your definition of "rude" seems quite one-sided.

> Fine, let's go home than.  I'll ask ma about good manners, and you'll
> ask your high school teacher about software testing.  We'll see who of
> us learns more on this trip…

Let's try not to turn this into a high school match...

    J

Message #50 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Dmitry Gutov <dgutov <at> yandex.ru>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: lunaryorn <at> gmail.com, Glenn Morris <rgm <at> gnu.org>, 14704 <at> debbugs.gnu.org,
 lekktu <at> gmail.com
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Tue, 25 Jun 2013 00:21:43 +0400

Eli Zaretskii <eliz <at> gnu.org> writes:

> Who except myself would want to run, let's see... about 30 tests, by
> hand, first in a GUI session, then in a TTY session?

Even if no one else would do that for now (I have no idea), sometimes
things happen, you know. Bus factor and all that.
I'm sure anyone else working on the display engine would find a suite of
tests helpful.

Of course, if you were able to automate them to some degree, that would
be even better.

Message #53 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Dmitry Gutov <dgutov <at> yandex.ru>
Cc: lunaryorn <at> gmail.com, rgm <at> gnu.org, 14704 <at> debbugs.gnu.org, lekktu <at> gmail.com
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 23:25:36 +0300

> From: Dmitry Gutov <dgutov <at> yandex.ru>
> Cc: Glenn Morris <rgm <at> gnu.org>,  lunaryorn <at> gmail.com,  lekktu <at> gmail.com,  14704 <at> debbugs.gnu.org
> Date: Tue, 25 Jun 2013 00:21:43 +0400
> 
> Eli Zaretskii <eliz <at> gnu.org> writes:
> 
> > Who except myself would want to run, let's see... about 30 tests, by
> > hand, first in a GUI session, then in a TTY session?
> 
> Even if no one else would do that for now (I have no idea), sometimes
> things happen, you know. Bus factor and all that.
> I'm sure anyone else working on the display engine would find a suite of
> tests helpful.

Fine.  I will add them one of these days.

> Of course, if you were able to automate them to some degree, that would
> be even better.

We lack a lot of infrastructure for that, like access from Lisp to
data structures manipulated by the display engine.  If I were to write
all that, I'd probably have no time left for development and fixing
bugs.  It would be a fine exercise for someone who'd like to learn the
display engine internals, though, so motivated people are welcome...

Message #56 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Juanma Barranquero <lekktu <at> gmail.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: lunaryorn <at> gmail.com, Glenn Morris <rgm <at> gnu.org>, 14704 <at> debbugs.gnu.org,
 Dmitry Gutov <dgutov <at> yandex.ru>
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 22:30:38 +0200

On Mon, Jun 24, 2013 at 10:25 PM, Eli Zaretskii <eliz <at> gnu.org> wrote:

> We lack a lot of infrastructure for that, like access from Lisp to
> data structures manipulated by the display engine.

What kind of API are you envisioning? Yes, I know that a designing a
precise API is half the work, I'm asking more about the general idea.

    J

Message #59 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Jambunathan K <kjambunathan <at> gmail.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Tue, 25 Jun 2013 02:49:41 +0530

Eli

>> > Who except myself would want to run, let's see... about 30 tests, by
>> > hand, first in a GUI session, then in a TTY session?

> We lack a lot of infrastructure for that, like access from Lisp to
> data structures manipulated by the display engine.  If I were to write
> all that, I'd probably have no time left for development and fixing
> bugs.  It would be a fine exercise for someone who'd like to learn the
> display engine internals, though, so motivated people are welcome...

Will this be a good project for GSoC? I am adding GSoC so that search
engines can pick it up.

Message #62 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: Sebastian Wiesner <lunaryorn <at> gmail.com>
Cc: Juanma Barranquero <lekktu <at> gmail.com>, 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 22:11:13 -0400

> By not writing or running tests for *your* code you have wasted *my*
> time.  That is what I call “rude”.

Im sorry I wasted your time, but do note that all the time I put into
breaking Emacs (and package.el more recently) is purely my own leisure
time, so I'm very happy not to have to follow too strict procedures.


        Stefan "who'd have been happier not touching package.el at all,
                since he never uses it and doesn't even know half of the
                expected use cases, which makes writing tests that much
                more interesting."

Message #65 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: lunaryorn <at> gmail.com, lekktu <at> gmail.com, 14704 <at> debbugs.gnu.org,
 Dmitry Gutov <dgutov <at> yandex.ru>
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Mon, 24 Jun 2013 22:35:25 -0400

> We lack a lot of infrastructure for that, like access from Lisp to
> data structures manipulated by the display engine.

For the tty case at least we could image just use the sequence of bytes
sent to the tty.


        Stefan

Message #68 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Jambunathan K <kjambunathan <at> gmail.com>
Cc: 14704 <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Tue, 25 Jun 2013 05:43:21 +0300

> From: Jambunathan K <kjambunathan <at> gmail.com>
> Cc: 14704 <at> debbugs.gnu.org
> Date: Tue, 25 Jun 2013 02:49:41 +0530
> 
> > We lack a lot of infrastructure for that, like access from Lisp to
> > data structures manipulated by the display engine.  If I were to write
> > all that, I'd probably have no time left for development and fixing
> > bugs.  It would be a fine exercise for someone who'd like to learn the
> > display engine internals, though, so motivated people are welcome...
> 
> Will this be a good project for GSoC?

Could be, I don't know.

Message #71 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Michael Albinus <michael.albinus <at> gmx.de>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: lunaryorn <at> gmail.com, Glenn Morris <rgm <at> gnu.org>, 14704 <at> debbugs.gnu.org,
 lekktu <at> gmail.com
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Tue, 25 Jun 2013 08:17:32 +0200

Eli Zaretskii <eliz <at> gnu.org> writes:

>> > E.g., I have a large body of unit tests for display features, such as
>> > cursor positioning, which I run by hand every time I touch any of the
>> > related code.
>> 
>> Do you feel like adding it to the repo?
>> Sounds like it would be useful to have in test/, even if not suitable
>> for test/automated/.
>
> Who except myself would want to run, let's see... about 30 tests, by
> hand, first in a GUI session, then in a TTY session?

I'm still plagued by display errors on one of my machines @ work. I
would be willing to run the tests in order to catch this annoying
bug. If it helps.

Best regards, Michael.

Message #74 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Stefan Monnier <monnier <at> iro.umontreal.ca>
Cc: lunaryorn <at> gmail.com, lekktu <at> gmail.com, 14704 <at> debbugs.gnu.org,
 dgutov <at> yandex.ru
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Tue, 25 Jun 2013 17:39:05 +0300

> From: Stefan Monnier <monnier <at> iro.umontreal.ca>
> Cc: Dmitry Gutov <dgutov <at> yandex.ru>,  lunaryorn <at> gmail.com,  lekktu <at> gmail.com,  14704 <at> debbugs.gnu.org
> Date: Mon, 24 Jun 2013 22:35:25 -0400
> 
> > We lack a lot of infrastructure for that, like access from Lisp to
> > data structures manipulated by the display engine.
> 
> For the tty case at least we could image just use the sequence of bytes
> sent to the tty.

True, but even TTY output is to some extent dependent on the terminal
capabilities, so parsing such byte streams is not entirely trivial.

Message #77 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Juanma Barranquero <lekktu <at> gmail.com>
Cc: lunaryorn <at> gmail.com, rgm <at> gnu.org, 14704 <at> debbugs.gnu.org, dgutov <at> yandex.ru
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Tue, 25 Jun 2013 17:43:48 +0300

> From: Juanma Barranquero <lekktu <at> gmail.com>
> Date: Mon, 24 Jun 2013 22:30:38 +0200
> Cc: Dmitry Gutov <dgutov <at> yandex.ru>, Glenn Morris <rgm <at> gnu.org>, lunaryorn <at> gmail.com, 
> 	14704 <at> debbugs.gnu.org
> 
> On Mon, Jun 24, 2013 at 10:25 PM, Eli Zaretskii <eliz <at> gnu.org> wrote:
> 
> > We lack a lot of infrastructure for that, like access from Lisp to
> > data structures manipulated by the display engine.
> 
> What kind of API are you envisioning? Yes, I know that a designing a
> precise API is half the work, I'm asking more about the general idea.

First, a caveat: I didn't really put a lot of thinking into this.

That said, what I had in mind was an API that would return a detailed
description of the current glyph matrix for a window.  Another API
should return the position and description (block, bar, hollow, etc.)
of the window's cursor.

Another useful feature would be to be able to produce the full
sequence of calls to display-specific back-end APIs which were
involved in a given redisplay cycle.  This would obviously require
design of some kind of language to describe these sequences.

There will be probably more once we start on this road, but the above
should be a good starting point, IMO.

Message #80 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Michael Albinus <michael.albinus <at> gmx.de>
Cc: lunaryorn <at> gmail.com, rgm <at> gnu.org, 14704 <at> debbugs.gnu.org, lekktu <at> gmail.com
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Tue, 25 Jun 2013 18:00:39 +0300

> From: Michael Albinus <michael.albinus <at> gmx.de>
> Cc: Glenn Morris <rgm <at> gnu.org>,  lunaryorn <at> gmail.com,  lekktu <at> gmail.com,  14704 <at> debbugs.gnu.org
> Date: Tue, 25 Jun 2013 08:17:32 +0200
> 
> Eli Zaretskii <eliz <at> gnu.org> writes:
> 
> >> > E.g., I have a large body of unit tests for display features, such as
> >> > cursor positioning, which I run by hand every time I touch any of the
> >> > related code.
> >> 
> >> Do you feel like adding it to the repo?
> >> Sounds like it would be useful to have in test/, even if not suitable
> >> for test/automated/.
> >
> > Who except myself would want to run, let's see... about 30 tests, by
> > hand, first in a GUI session, then in a TTY session?
> 
> I'm still plagued by display errors on one of my machines @ work. I
> would be willing to run the tests in order to catch this annoying
> bug. If it helps.

Since I don't know what kind of display errors you are having, it's
hard to answer your question.  The tests I was talking about exercise
the code in Emacs which moves and positions cursor triggered by cursor
motion commands, such as C-f, C-n, etc.

Message #83 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Michael Albinus <michael.albinus <at> gmx.de>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: lunaryorn <at> gmail.com, rgm <at> gnu.org, 14704 <at> debbugs.gnu.org, lekktu <at> gmail.com
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Tue, 25 Jun 2013 17:24:49 +0200

[Message part 1 (text/plain, inline)]

Eli Zaretskii <eliz <at> gnu.org> writes:

>> I'm still plagued by display errors on one of my machines @ work. I
>> would be willing to run the tests in order to catch this annoying
>> bug. If it helps.
>
> Since I don't know what kind of display errors you are having, it's
> hard to answer your question.  The tests I was talking about exercise
> the code in Emacs which moves and positions cursor triggered by cursor
> motion commands, such as C-f, C-n, etc.

See the appended screenshots, reading your message in gnus. The first is
opening the message (looks OK), the second one is after entering <space>
in the summary buffer. You see the ">" char in column 1.

It is a fresh Emacs build of today.

Best regards, Michael.

[Screenshot from 2013-06-25 17:17:12.png (image/png, attachment)]

[Screenshot from 2013-06-25 17:17:15.png (image/png, attachment)]

Message #86 received at 14704 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: lunaryorn <at> gmail.com, lekktu <at> gmail.com, 14704 <at> debbugs.gnu.org,
 dgutov <at> yandex.ru
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Tue, 25 Jun 2013 11:50:24 -0400

>> For the tty case at least we could image just use the sequence of bytes
>> sent to the tty.
> True, but even TTY output is to some extent dependent on the terminal
> capabilities, so parsing such byte streams is not entirely trivial.

Definitely not trivial, but we can control the terminal capabilities
(e.g. using a hand-crafted terminfo file).


        Stefan

Message #91 received at 14704-done <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: Sebastian Wiesner <lunaryorn <at> gmail.com>
Cc: 14704-done <at> debbugs.gnu.org
Subject: Re: bug#14704: 24.3.50; cl-lib breaks built-in Emacs version
Date: Tue, 25 Jun 2013 12:27:39 -0400

ELISP> (require 'cl-lib)
> cl-lib
ELISP> (package-built-in-p 'emacs (version-to-list "24.1"))
> nil
ELISP> (assq 'emacs package--builtin-versions)
> (emacs 2 2)

Thanks, should be fixed now,


        Stefan

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.