bug#33748: 27.0.50; harfbuzz branch fails to compile

bug#33748: 27.0.50; harfbuzz branch fails to compile

[Eli Zaretskii]

The compilation fails with this error:

    CC       ftfont.o
  ftfont.c: In function ‘ftfont_shape_by_hb’:
  ftfont.c:2866:3: warning: implicit declaration of function ‘hb_buffer_set_cluster_level’ [-Wimplicit-function-declaration]
     hb_buffer_set_cluster_level (hb_buffer, HB_BUFFER_CLUSTER_LEVEL_MONOTONE_CHARACTERS);
     ^
  ftfont.c:2866:43: error: ‘HB_BUFFER_CLUSTER_LEVEL_MONOTONE_CHARACTERS’ undeclared (first use in this function)
     hb_buffer_set_cluster_level (hb_buffer, HB_BUFFER_CLUSTER_LEVEL_MONOTONE_CHARACTERS);
					     ^
  ftfont.c:2866:43: note: each undeclared identifier is reported only once for each function it appears in
  make[1]: *** [ftfont.o] Error 1

The version of Harfbuzz I have on this machine is 0.9.27.  If this too old?  If so, what minimum version do I need to successfully build the branch?

(Note that this bug report is sent from Emacs built with a different configuration, for obvious reasons.)


In GNU Emacs 27.0.50 (build 35, x86_64-pc-linux-gnu, GTK+ Version 2.24.23)
 of 2018-12-09 built on fencepost.gnu.org
Repository revision: c2fdd50c3cb0b03d2414370c58c1aa2a6ec3311d
Repository branch: master
System Description: Trisquel GNU/Linux Belenos (7.0)

Recent messages:
Saving place in files enabled globally
For information about GNU Emacs and the GNU system, type C-h C-a.

Configured using:
 'configure --enable-checking=yes,glyphs --with-gif=no --with-tiff=no
 --with-jpeg=no --with-modules --enable-check-lisp-object-type
 'CFLAGS=-O0 -g3''

Configured features:
XPM PNG SOUND GSETTINGS GLIB NOTIFY INOTIFY ACL LIBSELINUX GNUTLS
LIBXML2 FREETYPE M17N_FLT LIBOTF XFT ZLIB TOOLKIT_SCROLL_BARS GTK2 X11
XDBE XIM MODULES THREADS

Important settings:
  value of $LANG: en_US.UTF-8
  locale-coding-system: utf-8-unix

Major mode: Lisp Interaction

Minor modes in effect:
  display-time-mode: t
  show-paren-mode: t
  savehist-mode: t
  save-place-mode: t
  tooltip-mode: t
  global-eldoc-mode: t
  eldoc-mode: t
  electric-indent-mode: t
  mouse-wheel-mode: t
  tool-bar-mode: t
  menu-bar-mode: t
  file-name-shadow-mode: t
  global-font-lock-mode: t
  font-lock-mode: t
  auto-composition-mode: t
  auto-encryption-mode: t
  auto-compression-mode: t
  line-number-mode: t

Load-path shadows:
None found.

Features:
(shadow flyspell ispell emacsbug message rmc puny seq byte-opt gv
compile comint ansi-color ring bytecomp byte-compile cconv dired
dired-loaddefs format-spec rfc822 mml easymenu mml-sec password-cache
epa derived epg epg-config gnus-util rmail rmail-loaddefs time-date
mm-decode mm-bodies mm-encode mail-parse rfc2231 mailabbrev gmm-utils
mailheader cl-loaddefs cl-lib sendmail rfc2047 rfc2045 ietf-drums
mm-util mail-prsvr mail-utils term/xterm xterm elec-pair server time
paren cus-start cus-load savehist saveplace mule-util tooltip eldoc
electric uniquify ediff-hook vc-hooks lisp-float-type mwheel
term/x-win x-win term/common-win x-dnd tool-bar dnd fontset image
regexp-opt fringe tabulated-list replace newcomment text-mode
elisp-mode lisp-mode prog-mode register page menu-bar rfn-eshadow
isearch timer select scroll-bar mouse jit-lock font-lock syntax
facemenu font-core term/tty-colors frame cl-generic cham georgian
utf-8-lang misc-lang vietnamese tibetan thai tai-viet lao korean
japanese eucjp-ms cp51932 hebrew greek romanian slovak czech european
ethiopic indian cyrillic chinese composite charscript charprop
case-table epa-hook jka-cmpr-hook help simple abbrev obarray
minibuffer cl-preloaded nadvice loaddefs button faces cus-face
macroexp files text-properties overlay sha1 md5 base64 format env
code-pages mule custom widget hashtable-print-readable backquote
threads inotify dynamic-setting system-font-setting
font-render-setting move-toolbar gtk x-toolkit x multi-tty
make-network-process emacs)

Memory information:
((conses 16 116491 9083)
 (symbols 48 21747 1)
 (strings 32 34083 1784)
 (string-bytes 1 922856)
 (vectors 16 13709)
 (vector-slots 8 470222 9078)
 (floats 8 59 312)
 (intervals 56 222 0)
 (buffers 992 10)
 (heap 1024 12212 901))

bug#33748: 27.0.50; harfbuzz branch fails to compile

[Glenn Morris]

Eli Zaretskii wrote:

>   ftfont.c:2866:3: warning: implicit declaration of function 'hb_buffer_set_cluster_level' [-Wimplicit-function-declaration]
[...]
> The version of Harfbuzz I have on this machine is 0.9.27.  If this too old?  

http://www.manpagez.com/html/harfbuzz/harfbuzz-/harfbuzz-Buffers.php#hb-buffer-set-cluster-level

says "Since: 0.9.42".

(Hooray for good docs!)

bug#33748: 27.0.50; harfbuzz branch fails to compile

[Khaled Hosny]

On Fri, Dec 14, 2018 at 03:27:37PM -0500, Eli Zaretskii wrote:
> 
> The compilation fails with this error:
> 
>     CC       ftfont.o
>   ftfont.c: In function ‘ftfont_shape_by_hb’:
>   ftfont.c:2866:3: warning: implicit declaration of function ‘hb_buffer_set_cluster_level’ [-Wimplicit-function-declaration]
>      hb_buffer_set_cluster_level (hb_buffer, HB_BUFFER_CLUSTER_LEVEL_MONOTONE_CHARACTERS);
>      ^
>   ftfont.c:2866:43: error: ‘HB_BUFFER_CLUSTER_LEVEL_MONOTONE_CHARACTERS’ undeclared (first use in this function)
>      hb_buffer_set_cluster_level (hb_buffer, HB_BUFFER_CLUSTER_LEVEL_MONOTONE_CHARACTERS);
> 					     ^
>   ftfont.c:2866:43: note: each undeclared identifier is reported only once for each function it appears in
>   make[1]: *** [ftfont.o] Error 1
> 
> The version of Harfbuzz I have on this machine is 0.9.27.  If this too old?  If so, what minimum version do I need to successfully build the branch?

Seems so. You can add configure.ac checks for hb_buffer_set_cluster_level
and use it conditionally, or require a newer HarfBuzz version. I prefer
the later, since the behaviour of cluster level 1 is assumed elsewhere
in the code (I don’t recall what exactly).

Regards,
Khaled

bug#33748: 27.0.50; harfbuzz branch fails to compile

[Eli Zaretskii]

> From: Glenn Morris <rgm@gnu.org>
> Cc: 33748@debbugs.gnu.org,  dr.khaled.hosny@gmail.com,  behdad@behdad.org,  far.nasiri.m@gmail.com
> Date: Fri, 14 Dec 2018 15:44:18 -0500
> 
> Eli Zaretskii wrote:
> 
> >   ftfont.c:2866:3: warning: implicit declaration of function 'hb_buffer_set_cluster_level' [-Wimplicit-function-declaration]
> [...]
> > The version of Harfbuzz I have on this machine is 0.9.27.  If this too old?  
> 
> http://www.manpagez.com/html/harfbuzz/harfbuzz-/harfbuzz-Buffers.php#hb-buffer-set-cluster-level
> 
> says "Since: 0.9.42".

Thanks, I reviewed the other symbols, and 0.9.42 seems the earliest
version that supports them.  Still, it would be good to know which
version is considered the oldest we should support, perhaps due to
some algorithmic issues that are not visible just by looking at the
symbol names.  I hope the Harfbuzz people could advise us about that.

bug#33748: 27.0.50; harfbuzz branch fails to compile

[Eli Zaretskii]

> Date: Sat, 15 Dec 2018 00:52:53 +0200
> From: Khaled Hosny <dr.khaled.hosny@gmail.com>
> Cc: bug-gnu-emacs@gnu.org, far.nasiri.m@gmail.com, behdad@behdad.org
> 
> > The version of Harfbuzz I have on this machine is 0.9.27.  If this too old?  If so, what minimum version do I need to successfully build the branch?
> 
> Seems so. You can add configure.ac checks for hb_buffer_set_cluster_level
> and use it conditionally, or require a newer HarfBuzz version. I prefer
> the later, since the behaviour of cluster level 1 is assumed elsewhere
> in the code (I don’t recall what exactly).

Which oldest version of HarfBuzz would you prefer us to require?  Is
0.9.42 OK, or do we need a newer one?

Thanks.

bug#33748: 27.0.50; harfbuzz branch fails to compile

[Eli Zaretskii]

> Date: Sat, 15 Dec 2018 08:59:43 +0200
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: behdad@behdad.org, 33748@debbugs.gnu.org, far.nasiri.m@gmail.com
> 
> Which oldest version of HarfBuzz would you prefer us to require?  Is
> 0.9.42 OK, or do we need a newer one?

Also, if we require at least 0.9.42, then the test for
hb_ft_font_create_referenced, which was introduced in 0.9.38, can be
dropped, together with the #ifdef'ed code that relies on that test,
right?

bug#33748: 27.0.50; harfbuzz branch fails to compile

[Colin Baxter]

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

    >> From: Glenn Morris <rgm@gnu.org> Cc: 33748@debbugs.gnu.org,
    >> dr.khaled.hosny@gmail.com, behdad@behdad.org,
    >> far.nasiri.m@gmail.com Date: Fri, 14 Dec 2018 15:44:18 -0500
    >> 
    >> Eli Zaretskii wrote:
    >> 
    >> > ftfont.c:2866:3: warning: implicit declaration of function >
    >> 'hb_buffer_set_cluster_level' [-Wimplicit-function-declaration]
    >> [...]  > The version of Harfbuzz I have on this machine is
    >> 0.9.27.  If this too old?
    >> 
    >> http://www.manpagez.com/html/harfbuzz/harfbuzz-/harfbuzz-Buffers.php#hb-buffer-set-cluster-level
    >> 
    >> says "Since: 0.9.42".

    > Thanks, I reviewed the other symbols, and 0.9.42 seems the
    > earliest version that supports them.  Still, it would be good to
    > know which version is considered the oldest we should support,
    > perhaps due to some algorithmic issues that are not visible just
    > by looking at the symbol names.  I hope the Harfbuzz people could
    > advise us about that.

I have Harfbuzz 0.9.35 on Debian 8.11 and I get similar compilation
errors.

Best wishes,

bug#33748: 27.0.50; harfbuzz branch fails to compile

[Khaled Hosny]

On Sat, Dec 15, 2018 at 08:59:43AM +0200, Eli Zaretskii wrote:
> > Date: Sat, 15 Dec 2018 00:52:53 +0200
> > From: Khaled Hosny <dr.khaled.hosny@gmail.com>
> > Cc: bug-gnu-emacs@gnu.org, far.nasiri.m@gmail.com, behdad@behdad.org
> > 
> > > The version of Harfbuzz I have on this machine is 0.9.27.  If this too old?  If so, what minimum version do I need to successfully build the branch?
> > 
> > Seems so. You can add configure.ac checks for hb_buffer_set_cluster_level
> > and use it conditionally, or require a newer HarfBuzz version. I prefer
> > the later, since the behaviour of cluster level 1 is assumed elsewhere
> > in the code (I don’t recall what exactly).
> 
> Which oldest version of HarfBuzz would you prefer us to require?  Is
> 0.9.42 OK, or do we need a newer one?

I’d prefer the latest version of course, there have been lots of bug
fixes and new features since then. If one is building his own HarfBuzz,
then definitely the latest version, otherwise the one that comes with
the system is good enough (Emacs wouldn’t be disadvantaged anymore than
any HarfBuzz-using application on the said system), as long as it has
the APIs we use.

Regards,
Khaled

bug#33748: 27.0.50; harfbuzz branch fails to compile

[Khaled Hosny]

On Sat, Dec 15, 2018 at 09:13:52AM +0200, Eli Zaretskii wrote:
> > Date: Sat, 15 Dec 2018 08:59:43 +0200
> > From: Eli Zaretskii <eliz@gnu.org>
> > Cc: behdad@behdad.org, 33748@debbugs.gnu.org, far.nasiri.m@gmail.com
> > 
> > Which oldest version of HarfBuzz would you prefer us to require?  Is
> > 0.9.42 OK, or do we need a newer one?
> 
> Also, if we require at least 0.9.42, then the test for
> hb_ft_font_create_referenced, which was introduced in 0.9.38, can be
> dropped, together with the #ifdef'ed code that relies on that test,
> right?

Right.

bug#33748: 27.0.50; harfbuzz branch fails to compile

[Eli Zaretskii]

> Date: Sat, 15 Dec 2018 12:32:41 +0200
> From: Khaled Hosny <dr.khaled.hosny@gmail.com>
> Cc: behdad@behdad.org, 33748@debbugs.gnu.org, far.nasiri.m@gmail.com
> 
> > Also, if we require at least 0.9.42, then the test for
> > hb_ft_font_create_referenced, which was introduced in 0.9.38, can be
> > dropped, together with the #ifdef'ed code that relies on that test,
> > right?
> 
> Right.

OK, I see that Paul already did that, so we are fine.

bug#33748: 27.0.50; harfbuzz branch fails to compile

[Colin Baxter]

>>>>> Colin Baxter <m43cap@yandex.com> writes:

    > I have Harfbuzz 0.9.35 on Debian 8.11 and I get similar
    > compilation errors.

To add to the success-chorus, moving to another machine, debian 9.6 with
harfbuzz 1.4.2, I can make emacs-27.0.50 with HARFBUZZ as a configured
feature. 

bug#33748: 27.0.50; harfbuzz branch fails to compile

[Eli Zaretskii]

> From: Colin Baxter <m43cap@yandex.com>
> Cc: dr.khaled.hosny@gmail.com, far.nasiri.m@gmail.com,  behdad@behdad.org,  33748@debbugs.gnu.org
> Cc: 
> Date: Sat, 15 Dec 2018 11:24:10 +0000
> 
> To add to the success-chorus, moving to another machine, debian 9.6 with
> harfbuzz 1.4.2, I can make emacs-27.0.50 with HARFBUZZ as a configured
> feature. 

Thanks.

bug#33748: 27.0.50; harfbuzz branch fails to compile

[Glenn Morris]

Khaled Hosny wrote:

>> Which oldest version of HarfBuzz would you prefer us to require?  Is
>> 0.9.42 OK, or do we need a newer one?
>
> I'd prefer the latest version of course, there have been lots of bug
> fixes and new features since then.

Right, but that doesn't answer the question.
What is the minimum version configure should accept?
I did test 0.9.42 and it worked, but eg perhaps 1.0 is a more sensible cut-off.

bug#33748: 27.0.50; harfbuzz branch fails to compile

[Eli Zaretskii]

> From: Glenn Morris <rgm@gnu.org>
> Cc: Eli Zaretskii <eliz@gnu.org>,  behdad@behdad.org,  33748@debbugs.gnu.org,  far.nasiri.m@gmail.com
> Date: Sat, 15 Dec 2018 12:50:59 -0500
> 
> Khaled Hosny wrote:
> 
> >> Which oldest version of HarfBuzz would you prefer us to require?  Is
> >> 0.9.42 OK, or do we need a newer one?
> >
> > I'd prefer the latest version of course, there have been lots of bug
> > fixes and new features since then.
> 
> Right, but that doesn't answer the question.

I think it does:

>                                    If one is building his own HarfBuzz,
> then definitely the latest version, otherwise the one that comes with
> the system is good enough (Emacs wouldn’t be disadvantaged anymore than
> any HarfBuzz-using application on the said system), as long as it has
> the APIs we use.

The last part, in my interpretation, means 0.9.42 is "good enough",
since it has the APIs used by the code.

bug#33748: 27.0.50; harfbuzz branch fails to compile

[Glenn Morris]

Eli Zaretskii wrote:

> The last part, in my interpretation, means 0.9.42 is "good enough",
> since it has the APIs used by the code.

I feel like I'm flogging a dead horse, but to me there's a difference
between "good enough" and "the minimum version that we want to support
going forward". Is anyone going to be testing 0.9.42 on a regular basis?

It looks 0.9.42 and 1.0 were released on the same day, so requiring the
former instead of the latter makes little sense to me.

Debian stable has 1.4.2, RHEL 7 has 1.7.5, Ubuntu 16.04 has 1.0.1,
Ubuntu 18.04 has 1.7.2.

So personally I would set the minimum to 1.4.2 (Jan 2017).

And that's all I have to say about that. :)