Build failure caused by a .tit file

Build failure caused by a .tit file

[Göktuğ Kayaalp]

Hi,

for the past couple of months I've been unable to build from the git
repo altho following the instructions from ‘INSTALL.REPO’.  My recipe is

    ./autogen.sh all
    ./configure --enable-silent-rules --with-x-toolkit=gtk3 --with-modules
    --with-file-notification=yes --with-mailutils --with-x
    make

and I end up with:

,----
| ...
| make[4]: Entering directory '/igk/checkouts/GNUEmacs/leim'
|   GEN      ../lisp/leim/quail/CCDOSPY.el
| Converting /igk/checkouts/GNUEmacs/leim/CXTERM-DIC/CCDOSPY.tit to quail-package...
| Loading macroexp.elc...
| TIT dictionary doesn’t have body part
| Makefile:106: recipe for target '../lisp/leim/quail/CCDOSPY.el' failed
| make[4]: *** [../lisp/leim/quail/CCDOSPY.el] Error 255
| make[4]: Leaving directory '/igk/checkouts/GNUEmacs/leim'
| Makefile:355: recipe for target 'leim' failed
| make[3]: *** [leim] Error 2
| make[3]: Leaving directory '/igk/checkouts/GNUEmacs/lisp'
| Makefile:734: recipe for target '../lisp/loaddefs.el' failed
| make[2]: *** [../lisp/loaddefs.el] Error 2
| make[2]: Leaving directory '/igk/checkouts/GNUEmacs/src'
| Makefile:416: recipe for target 'src' failed
| make[1]: *** [src] Error 2
| make[1]: Leaving directory '/igk/checkouts/GNUEmacs'
| Makefile:1097: recipe for target 'bootstrap' failed
| make: *** [bootstrap] Error 2
`----

This happens both on the master branch (edc63bf) and on emacs-25.1 tag.
I've also tried ‘make bootstrap’ with no luck, ending up here again.
I don't have trouble building from the latest release tarball.

Thanks in advance,
-gk.

Re: Build failure caused by a .tit file

[良ϖ]

Well, have you tried to figure out if it's a big tit file?

On 17/04/2017, Göktuğ Kayaalp <self@gkayaalp.com> wrote:
> Hi,
>
> for the past couple of months I've been unable to build from the git
> repo altho following the instructions from ‘INSTALL.REPO’.  My recipe is
>
>     ./autogen.sh all
>     ./configure --enable-silent-rules --with-x-toolkit=gtk3 --with-modules
>     --with-file-notification=yes --with-mailutils --with-x
>     make
>
> and I end up with:
>
> ,----
> | ...
> | make[4]: Entering directory '/igk/checkouts/GNUEmacs/leim'
> |   GEN      ../lisp/leim/quail/CCDOSPY.el
> | Converting /igk/checkouts/GNUEmacs/leim/CXTERM-DIC/CCDOSPY.tit to
> quail-package...
> | Loading macroexp.elc...
> | TIT dictionary doesn’t have body part
> | Makefile:106: recipe for target '../lisp/leim/quail/CCDOSPY.el' failed
> | make[4]: *** [../lisp/leim/quail/CCDOSPY.el] Error 255
> | make[4]: Leaving directory '/igk/checkouts/GNUEmacs/leim'
> | Makefile:355: recipe for target 'leim' failed
> | make[3]: *** [leim] Error 2
> | make[3]: Leaving directory '/igk/checkouts/GNUEmacs/lisp'
> | Makefile:734: recipe for target '../lisp/loaddefs.el' failed
> | make[2]: *** [../lisp/loaddefs.el] Error 2
> | make[2]: Leaving directory '/igk/checkouts/GNUEmacs/src'
> | Makefile:416: recipe for target 'src' failed
> | make[1]: *** [src] Error 2
> | make[1]: Leaving directory '/igk/checkouts/GNUEmacs'
> | Makefile:1097: recipe for target 'bootstrap' failed
> | make: *** [bootstrap] Error 2
> `----
>
> This happens both on the master branch (edc63bf) and on emacs-25.1 tag.
> I've also tried ‘make bootstrap’ with no luck, ending up here again.
> I don't have trouble building from the latest release tarball.
>
> Thanks in advance,
> -gk.
>
>

Re: Build failure caused by a .tit file

[Göktuğ Kayaalp]

On 2017-04-17 14:41 GMT, 良ϖ <p.de.boisset@gmail.com> wrote:
> Well, have you tried to figure out if it's a big tit file?

It has 565 lines and ~10000 characters.

Re: Build failure caused by a .tit file

[Eli Zaretskii]

> From: Göktuğ Kayaalp <self@gkayaalp.com>
> Date: Mon, 17 Apr 2017 17:03:58 +0300
> 
> for the past couple of months I've been unable to build from the git
> repo altho following the instructions from ‘INSTALL.REPO’.

This is not a good place to report failures to build Emacs.

> and I end up with:
> 
> ,----
> | ...
> | make[4]: Entering directory '/igk/checkouts/GNUEmacs/leim'
> |   GEN      ../lisp/leim/quail/CCDOSPY.el
> | Converting /igk/checkouts/GNUEmacs/leim/CXTERM-DIC/CCDOSPY.tit to quail-package...
> | Loading macroexp.elc...
> | TIT dictionary doesn’t have body part

This error message comes from this snippet in titdic-cnv.el:

	  ;; Decode the buffer contents from the encoding specified by a
	  ;; value of the key "ENCODE:".
	  (if (not (search-forward "\nBEGIN" nil t))
	      (error "TIT dictionary doesn't have body part"))

Is it really true that CCDOSPY.tit you have has no line that begins
with "BEGIN"?  Otherwise, why doesn't the search-forward call succeed?

> I've also tried ‘make bootstrap’ with no luck, ending up here again.

Did you try "make distclean" before building?

Re: Build failure caused by a .tit file

[Göktuğ Kayaalp]

On 2017-04-17 17:45 +0300, Eli Zaretskii <eliz@gnu.org> wrote:
> This is not a good place to report failures to build Emacs.

I'm sorry.  Should I rather post it to emacs-devel or report bug?

> This error message comes from this snippet in titdic-cnv.el:
>
> 	  ;; Decode the buffer contents from the encoding specified by a
> 	  ;; value of the key "ENCODE:".
> 	  (if (not (search-forward "\nBEGIN" nil t))
> 	      (error "TIT dictionary doesn't have body part"))
>
> Is it really true that CCDOSPY.tit you have has no line that begins
> with "BEGIN"?  Otherwise, why doesn't the search-forward call succeed?

The line 55-57 reads

# the following line must not be removed
BEGINDICTIONARY
#

and manually running

M-: (search-forward "\nBEGIN" nil t)

returns 1850.  The snippet you've included when run via M-:, works as
expected, i.e., no errors, moves to line 56, after first ‘N’.

> Did you try "make distclean" before building?

Yes, I tried starting from scratch both via the build tool and fetching
a clean copy of the repository.  FWIW the file's coding is ‘euc-china’
and it has 565 lines (another reply was asking whether the file was a
big one).

Re: Build failure caused by a .tit file

[Eli Zaretskii]

> From: Göktuğ Kayaalp <self@gkayaalp.com>
> Cc: help-gnu-emacs@gnu.org
> Date: Mon, 17 Apr 2017 18:55:30 +0300
> 
> > This error message comes from this snippet in titdic-cnv.el:
> >
> > 	  ;; Decode the buffer contents from the encoding specified by a
> > 	  ;; value of the key "ENCODE:".
> > 	  (if (not (search-forward "\nBEGIN" nil t))
> > 	      (error "TIT dictionary doesn't have body part"))
> >
> > Is it really true that CCDOSPY.tit you have has no line that begins
> > with "BEGIN"?  Otherwise, why doesn't the search-forward call succeed?
> 
> The line 55-57 reads
> 
> # the following line must not be removed
> BEGINDICTIONARY
> #
> 
> and manually running
> 
> M-: (search-forward "\nBEGIN" nil t)
> 
> returns 1850.  The snippet you've included when run via M-:, works as
> expected, i.e., no errors, moves to line 56, after first ‘N’.

So if you manually invoke titdic-convert on CCDOSPY.tit, it does work?
And you do all this in the same Emacs binary, probably called
bootstrap-emacs, which runs the failing command during the build?

Re: Build failure caused by a .tit file

[Göktuğ Kayaalp]

On 2017-04-17 19:21 +0300, Eli Zaretskii <eliz@gnu.org> wrote:
>> From: Göktuğ Kayaalp <self@gkayaalp.com>
>> Cc: help-gnu-emacs@gnu.org
>> Date: Mon, 17 Apr 2017 18:55:30 +0300
>> 
>> > This error message comes from this snippet in titdic-cnv.el:
>> >
>> > 	  ;; Decode the buffer contents from the encoding specified by a
>> > 	  ;; value of the key "ENCODE:".
>> > 	  (if (not (search-forward "\nBEGIN" nil t))
>> > 	      (error "TIT dictionary doesn't have body part"))
>> >
>> > Is it really true that CCDOSPY.tit you have has no line that begins
>> > with "BEGIN"?  Otherwise, why doesn't the search-forward call succeed?
>> 
>> The line 55-57 reads
>> 
>> # the following line must not be removed
>> BEGINDICTIONARY
>> #
>> 
>> and manually running
>> 
>> M-: (search-forward "\nBEGIN" nil t)
>> 
>> returns 1850.  The snippet you've included when run via M-:, works as
>> expected, i.e., no errors, moves to line 56, after first ‘N’.
>
> So if you manually invoke titdic-convert on CCDOSPY.tit, it does work?
> And you do all this in the same Emacs binary, probably called
> bootstrap-emacs, which runs the failing command during the build?

I've run this commands from an instance of Emacs 25.1, not from the
version that I'm trying to build from the repo.  bootstrap-emacs aborts
with ‘Fatal error 6’ when I run it, dumping a backtrace and core.  I’ve
both tried to run it normally and with the command line

src/bootstrap-emacs -file leim/CXTERM-DIC/CCDOSPY.tit -eval '(titdic-convert)'

same results.  I also tried using titdic-convert with my usual emacs
instance (25.1), which successfully created a ‘CCDOSPY.el’ file in the
current directory.  Find below the backtrace I get from bootstrap-emacs:

,----
| Fatal error 6: Aborted
| Backtrace:
| src/bootstrap-emacs[0x50d4c9]
| src/bootstrap-emacs[0x4f3f37]
| src/bootstrap-emacs[0x50d573]
| src/bootstrap-emacs[0x4aa614]
| src/bootstrap-emacs[0x4ae45f]
| src/bootstrap-emacs[0x441985]
| src/bootstrap-emacs[0x452de6]
| src/bootstrap-emacs[0x426b07]
| src/bootstrap-emacs[0x5680ac]
| src/bootstrap-emacs[0x56842d]
| src/bootstrap-emacs[0x5681a0]
| src/bootstrap-emacs[0x56b9a6]
| src/bootstrap-emacs[0x5681a0]
| src/bootstrap-emacs[0x56b7ed]
| src/bootstrap-emacs[0x5681a0]
| src/bootstrap-emacs[0x568475]
| src/bootstrap-emacs[0x5681a0]
| src/bootstrap-emacs[0x568885]
| src/bootstrap-emacs[0x567b50]
| src/bootstrap-emacs[0x567ef2]
| src/bootstrap-emacs[0x56842d]
| src/bootstrap-emacs[0x5681a0]
| src/bootstrap-emacs[0x56b9a6]
| src/bootstrap-emacs[0x5681a0]
| src/bootstrap-emacs[0x56b53d]
| src/bootstrap-emacs[0x5681a0]
| src/bootstrap-emacs[0x568885]
| src/bootstrap-emacs[0x567b50]
| src/bootstrap-emacs[0x567ef2]
| src/bootstrap-emacs[0x56842d]
| src/bootstrap-emacs[0x5681a0]
| src/bootstrap-emacs[0x568885]
| src/bootstrap-emacs[0x568acb]
| src/bootstrap-emacs[0x56a72c]
| src/bootstrap-emacs[0x568b56]
| src/bootstrap-emacs[0x5a20f3]
| src/bootstrap-emacs[0x567b50]
| src/bootstrap-emacs[0x567ef2]
| src/bootstrap-emacs[0x56b7ed]
| src/bootstrap-emacs[0x5681a0]
| src/bootstrap-emacs[0x56aff6]
| ...
| Aborted (core dumped)
`----

--gk.

Re: Build failure caused by a .tit file

[Eli Zaretskii]

> From: Göktuğ Kayaalp <self@gkayaalp.com>
> Cc: help-gnu-emacs@gnu.org
> Date: Tue, 18 Apr 2017 02:08:34 +0300
> 
> I've run this commands from an instance of Emacs 25.1, not from the
> version that I'm trying to build from the repo.  bootstrap-emacs aborts
> with ‘Fatal error 6’ when I run it, dumping a backtrace and core.  I’ve
> both tried to run it normally and with the command line
> 
> src/bootstrap-emacs -file leim/CXTERM-DIC/CCDOSPY.tit -eval '(titdic-convert)'
> 
> same results.

What happens if you add -Q to the bootstrap-emacs command line?  The
above loads your init files, which could be the reason for the abort.

> Find below the backtrace I get from bootstrap-emacs:
> 
> ,----
> | Fatal error 6: Aborted
> | Backtrace:
> | src/bootstrap-emacs[0x50d4c9]

This cannot be interpreted except on your system (see the Emacs manual
for how to do that).  If using -Q doesn't stop the aborts, it is best
to run these commands under GDB and show the backtrace from there.

Re: Build failure caused by a .tit file

[Göktuğ Kayaalp]

On 2017-04-18 05:43 +0300, Eli Zaretskii <eliz@gnu.org> wrote:
>> From: Göktuğ Kayaalp <self@gkayaalp.com>
>> Cc: help-gnu-emacs@gnu.org
>> Date: Tue, 18 Apr 2017 02:08:34 +0300
>> 
>> src/bootstrap-emacs -file leim/CXTERM-DIC/CCDOSPY.tit -eval
>> '(titdic-convert)'
>
> What happens if you add -Q to the bootstrap-emacs command line?  The
> above loads your init files, which could be the reason for the abort.

I actually ran it with -Q too, just pasted in the wrong command line.
Sorry.  The same results anyways.  The program aborts whathever the
arguments are (even when there are none), but I got it to run adding
‘-batch’ argument.  Here is what I got from some example sessions:

,----
| g@xi:~/co/GNUEmacs$ src/bootstrap-emacs -Q -batch -eval "(call-interactively #'titdic-convert)"
| TIT dictionary file: leim/CXTERM-DIC/CCDOSPY.tit
| Loading macroexp.elc...
| TIT dictionary doesn’t have body part
`----

Here it prompted me for the file name.

,----
| g@xi:~/co/GNUEmacs$ src/bootstrap-emacs -Q -batch -eval "(progn (find-file \"leim/CXTERM-DIC/CCDOSPY.tit\") (message \"%d\" (search-forward \"\\nBEGIN\")))"
| 1850
`----

Here it does indeed find the ‘\nBEGIN’ string.  The following patch
fixes this error, but IDK if sanely:

diff --git a/lisp/international/titdic-cnv.el b/lisp/international/titdic-cnv.el
index 130bc74..6a593e2 100644
--- a/lisp/international/titdic-cnv.el
+++ b/lisp/international/titdic-cnv.el
@@ -474,7 +474,6 @@ titdic-convert
     (with-temp-file  (tit-make-quail-package-file-name filename dirname)
       (let ((standard-output (current-buffer)))
 	(with-temp-buffer
-	  (set-buffer-multibyte nil)
 	  ;; Here we must use `raw-text' instead of `no-conversion' to
 	  ;; enable auto-decoding of eol format (CRLF->LF).
 	  (let ((coding-system-for-read 'raw-text))

And here is what happens now:

,----
| g@xi:~/co/GNUEmacs$ src/bootstrap-emacs -Q -batch -eval '(titdic-convert "leim/CXTERM-DIC/CCDOSPY.tit")'
| Decoding with coding system euc-china...
| Processing header part...
| Formatting translation rules...
`----

Exits 0, file created.  The build also *continues* with this patch
applied.  That ‘set-buffer-multibyte’ line was there since 2002 though
(see commit c7211fed), so maybe some semantics changed recently about
regexps?

>> Find below the backtrace I get from bootstrap-emacs:
>> 
>> ,----
>> | Fatal error 6: Aborted
>> | Backtrace:
>> | src/bootstrap-emacs[0x50d4c9]
>
> This cannot be interpreted except on your system (see the Emacs manual
> for how to do that).  If using -Q doesn't stop the aborts, it is best
> to run these commands under GDB and show the backtrace from there.

Find below a copy of a gdb session.  I do not know much about debugging
C programs, but I'll try and do further debugging in my spare time if
need be.  I guess without -batch it tried to load the GUI which wasn't
properly lined at that time or maybe isn't included in bootstrap-emacs
at all?

,----
| g@xi:~/co/GNUEmacs$ gdb --args src/bootstrap-emacs -Q -file leim/CXTERM-DIC/CCDOSPY.tit -eval '(titdic-convert)'
| GNU gdb (Ubuntu 7.11.1-0ubuntu1~16.04) 7.11.1
| Copyright (C) 2016 Free Software Foundation, Inc.
| License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
| This is free software: you are free to change and redistribute it.
| There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
| and "show warranty" for details.
| This GDB was configured as "x86_64-linux-gnu".
| Type "show configuration" for configuration details.
| For bug reporting instructions, please see:
| <http://www.gnu.org/software/gdb/bugs/>.
| Find the GDB manual and other documentation resources online at:
| <http://www.gnu.org/software/gdb/documentation/>.
| For help, type "help".
| Type "apropos word" to search for commands related to "word"...
| Reading symbols from src/bootstrap-emacs...done.
| (gdb) run
| Starting program: /igk/checkouts/GNUEmacs/src/bootstrap-emacs -Q -file leim/CXTERM-DIC/CCDOSPY.tit -eval \(titdic-convert\)
| [Thread debugging using libthread_db enabled]
| Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
| [New Thread 0x7fffe57b2700 (LWP 13095)]
| [New Thread 0x7fffe4869700 (LWP 13096)]
| [New Thread 0x7fffdfdb4700 (LWP 13097)]
| Fatal error 6: Aborted
| Backtrace:
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x50d4c9]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x4f3f37]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x50d573]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x4aa614]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x4ae45f]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x441985]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x452de6]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x426b07]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x5680ac]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x56842d]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x5681a0]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x56b9a6]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x5681a0]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x56b7ed]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x5681a0]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x568475]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x5681a0]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x568885]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x567b50]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x567ef2]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x56842d]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x5681a0]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x56b9a6]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x5681a0]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x56b53d]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x5681a0]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x568885]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x567b50]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x567ef2]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x56842d]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x5681a0]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x568885]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x568acb]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x56a72c]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x568b56]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x5a20f3]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x567b50]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x567ef2]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x56b7ed]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x5681a0]
| /igk/checkouts/GNUEmacs/src/bootstrap-emacs[0x56aff6]
| ...
| 
| Thread 1 "bootstrap-emacs" received signal SIGABRT, Aborted.
| 0x00007ffff0d59269 in raise (sig=sig@entry=6)
|     at ../sysdeps/unix/sysv/linux/pt-raise.c:35
| 35	../sysdeps/unix/sysv/linux/pt-raise.c: No such file or directory.
| (gdb) bt
| #0  0x00007ffff0d59269 in raise (sig=sig@entry=6)
|     at ../sysdeps/unix/sysv/linux/pt-raise.c:35
| #1  0x00000000004f3f04 in terminate_due_to_signal (sig=sig@entry=6, 
|     backtrace_limit=backtrace_limit@entry=40) at emacs.c:393
| #2  0x000000000050d573 in emacs_abort () at sysdep.c:2371
| #3  0x00000000004aa614 in bidi_initialize () at bidi.c:1097
| #4  0x00000000004ae45f in bidi_init_it (charpos=charpos@entry=1, 
|     bytepos=1, frame_window_p=<optimized out>, 
|     bidi_it=bidi_it@entry=0x7fffffffaf90) at bidi.c:1139
| #5  0x0000000000441985 in init_iterator (it=0x7fffffffa5c0, w=0xc80688, 
|     charpos=1, bytepos=<optimized out>, row=<optimized out>, 
|     base_face_id=DEFAULT_FACE_ID) at xdisp.c:2982
| #6  0x0000000000452de6 in resize_mini_window (w=0xc80688, 
|     exact_p=exact_p@entry=true) at xdisp.c:11066
| #7  0x0000000000426b07 in do_switch_frame (frame=31925957, 
|     track=<optimized out>, for_deletion=<optimized out>, 
|     norecord=<optimized out>) at frame.c:1229
| #8  0x00000000005680ac in eval_sub (form=<optimized out>) at eval.c:2221
| #9  0x000000000056842d in Fprogn (body=27113891) at eval.c:449
| #10 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #11 0x000000000056b9a6 in Funwind_protect (args=27113635) at eval.c:1185
| #12 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #13 0x000000000056b7ed in Fprogn (body=27495923) at eval.c:449
| #14 Flet (args=27547491) at eval.c:963
| #15 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #16 0x0000000000568475 in Fprogn (body=27597507) at eval.c:449
| #17 Fif (args=<optimized out>) at eval.c:407
| #18 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #19 0x0000000000568885 in Fprogn (body=27792595) at eval.c:449
| #20 funcall_lambda (fun=fun@entry=27792947, nargs=nargs@entry=1, 
|     arg_vector=arg_vector@entry=0x7fffffffbe40) at eval.c:3013
| #21 0x0000000000567b50 in apply_lambda (fun=27792947, 
|     args=<optimized out>, count=count@entry=41) at eval.c:2879
| #22 0x0000000000567ef2 in eval_sub (form=<optimized out>) at eval.c:2293
| #23 0x000000000056842d in Fprogn (body=17257859) at eval.c:449
| #24 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #25 0x000000000056b9a6 in Funwind_protect (args=17222675) at eval.c:1185
| #26 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #27 0x000000000056b53d in Fprogn (body=17221587) at eval.c:449
| #28 FletX (args=17299395) at eval.c:898
| #29 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #30 0x0000000000568885 in Fprogn (body=17299203) at eval.c:449
| #31 funcall_lambda (fun=17298915, fun@entry=17298899, nargs=nargs@entry=1, 
|     arg_vector=arg_vector@entry=0x7fffffffc2d0) at eval.c:3013
| #32 0x0000000000567b50 in apply_lambda (fun=17298899, 
|     args=<optimized out>, count=count@entry=34) at eval.c:2879
| #33 0x0000000000567ef2 in eval_sub (form=<optimized out>) at eval.c:2293
| #34 0x000000000056842d in Fprogn (body=27892675) at eval.c:449
| #35 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #36 0x0000000000568885 in Fprogn (body=28054339) at eval.c:449
| #37 funcall_lambda (fun=28110403, nargs=nargs@entry=1, 
|     arg_vector=arg_vector@entry=0x7fffffffc678) at eval.c:3013
| #38 0x0000000000568acb in Ffuncall (nargs=nargs@entry=2, 
|     args=args@entry=0x7fffffffc670) at eval.c:2756
| #39 0x000000000056a72c in Fapply (nargs=2, args=0x7fffffffc670)
|     at eval.c:2330
| #40 0x0000000000568b56 in Ffuncall (nargs=<optimized out>, 
|     args=args@entry=0x7fffffffc668) at eval.c:2742
| #41 0x00000000005a20f3 in exec_byte_code (bytestr=<optimized out>, 
|     vector=22990725, maxdepth=<optimized out>, 
|     args_template=<optimized out>, nargs=nargs@entry=281474976710657, 
|     args=<optimized out>, args@entry=0x7fffffffc668) at bytecode.c:641
| #42 0x000000000056884f in funcall_lambda (fun=140737488340733, 
|     fun@entry=19308213, nargs=281474976710657, nargs@entry=1, 
|     arg_vector=0x7fffffffc668, arg_vector@entry=0x7fffffffc7f0)
|     at eval.c:2942
| #43 0x0000000000567b50 in apply_lambda (fun=19308213, 
|     args=<optimized out>, count=count@entry=29) at eval.c:2879
| #44 0x0000000000567ef2 in eval_sub (form=<optimized out>) at eval.c:2293
| #45 0x000000000056b7ed in Fprogn (body=22217139) at eval.c:449
| #46 Flet (args=22217267) at eval.c:963
| #47 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #48 0x000000000056aff6 in Fsetq (args=<optimized out>) at eval.c:511
| #49 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #50 0x000000000056b53d in Fprogn (body=25416915) at eval.c:449
| #51 FletX (args=25347251) at eval.c:898
| #52 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #53 0x0000000000568885 in Fprogn (body=25341763) at eval.c:449
| #54 funcall_lambda (fun=25364227, fun@entry=25364211, nargs=nargs@entry=1, 
|     arg_vector=arg_vector@entry=0x7fffffffcd00) at eval.c:3013
| #55 0x0000000000567b50 in apply_lambda (fun=25364211, 
|     args=<optimized out>, count=count@entry=22) at eval.c:2879
| #56 0x0000000000567ef2 in eval_sub (form=<optimized out>) at eval.c:2293
| #57 0x000000000056aff6 in Fsetq (args=<optimized out>) at eval.c:511
| #58 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #59 0x000000000056aff6 in Fsetq (args=<optimized out>) at eval.c:511
| #60 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #61 0x000000000056842d in Fprogn (body=25321651) at eval.c:449
| #62 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #63 0x00000000005683bb in For (args=25321475) at eval.c:364
| #64 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #65 0x000000000056842d in Fprogn (body=25353171) at eval.c:449
| #66 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #67 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #68 0x0000000000568885 in Fprogn (body=25352195) at eval.c:449
| #69 funcall_lambda (fun=25464211, fun@entry=25464131, nargs=nargs@entry=0, 
|     arg_vector=arg_vector@entry=0x7fffffffd3c0) at eval.c:3013
| #70 0x0000000000567b50 in apply_lambda (fun=25464131, 
|     args=<optimized out>, count=count@entry=14) at eval.c:2879
| #71 0x0000000000567ef2 in eval_sub (form=<optimized out>) at eval.c:2293
| #72 0x0000000000568475 in Fprogn (body=26393955) at eval.c:449
| #73 Fif (args=<optimized out>) at eval.c:407
| #74 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #75 0x0000000000568885 in Fprogn (body=26789411) at eval.c:449
| #76 funcall_lambda (fun=25272643, fun@entry=25272627, nargs=nargs@entry=0, 
|     arg_vector=arg_vector@entry=0x7fffffffd650) at eval.c:3013
| #77 0x0000000000567b50 in apply_lambda (fun=25272627, 
|     args=<optimized out>, count=count@entry=11) at eval.c:2879
| #78 0x0000000000567ef2 in eval_sub (form=<optimized out>) at eval.c:2293
| #79 0x000000000056b9a6 in Funwind_protect (args=26399843) at eval.c:1185
| #80 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #81 0x000000000056b7ed in Fprogn (body=26347235) at eval.c:449
| #82 Flet (args=26347123) at eval.c:963
| #83 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #84 0x0000000000568475 in Fprogn (body=26346979) at eval.c:449
| #85 Fif (args=<optimized out>) at eval.c:407
| #86 0x00000000005681a0 in eval_sub (form=<optimized out>) at eval.c:2171
| #87 0x0000000000568885 in Fprogn (body=26009539) at eval.c:449
| #88 funcall_lambda (fun=26324227, fun@entry=26324243, nargs=nargs@entry=0, 
|     arg_vector=arg_vector@entry=0x7fffffffdb10) at eval.c:3013
| #89 0x0000000000567b50 in apply_lambda (fun=26324243, 
|     args=<optimized out>, count=count@entry=4) at eval.c:2879
| #90 0x0000000000567ef2 in eval_sub (form=form@entry=20487219)
|     at eval.c:2293
| #91 0x000000000056bd20 in Feval (form=20487219, lexical=<optimized out>)
|     at eval.c:2040
| #92 0x000000000056708e in internal_condition_case (
|     bfun=bfun@entry=0x4f4300 <top_level_2>, handlers=handlers@entry=19872, 
|     hfun=hfun@entry=0x4f98a0 <cmd_error>) at eval.c:1324
| #93 0x00000000004f6a4c in top_level_1 (ignore=ignore@entry=0)
|     at keyboard.c:1129
| #94 0x000000000056702c in internal_catch (tag=tag@entry=48432, 
|     func=func@entry=0x4f69f0 <top_level_1>, arg=arg@entry=0) at eval.c:1091
| #95 0x00000000004f4298 in command_loop () at keyboard.c:1090
| #96 0x00000000004f9486 in recursive_edit_1 () at keyboard.c:697
| #97 0x00000000004f97cb in Frecursive_edit () at keyboard.c:768
| #98 0x0000000000419a5e in main (argc=6, argv=0x7fffffffdee8)
|     at emacs.c:1687
`----

--gk.

Re: Build failure caused by a .tit file

[Göktuğ Kayaalp]

On 2017-04-18 15:41 +0300, Göktuğ Kayaalp <self@gkayaalp.com> wrote:
> diff --git a/lisp/international/titdic-cnv.el b/lisp/international/titdic-cnv.el
> index 130bc74..6a593e2 100644
> --- a/lisp/international/titdic-cnv.el
> +++ b/lisp/international/titdic-cnv.el
> @@ -474,7 +474,6 @@ titdic-convert
>      (with-temp-file  (tit-make-quail-package-file-name filename dirname)
>        (let ((standard-output (current-buffer)))
>  	(with-temp-buffer
> -	  (set-buffer-multibyte nil)
>  	  ;; Here we must use `raw-text' instead of `no-conversion' to
>  	  ;; enable auto-decoding of eol format (CRLF->LF).
>  	  (let ((coding-system-for-read 'raw-text))

A better version of this one is as follows, which removes the
complementary set-buffer-multibyte call:

diff --git a/lisp/international/titdic-cnv.el b/lisp/international/titdic-cnv.el
index 130bc74..aa6e4f1 100644
--- a/lisp/international/titdic-cnv.el
+++ b/lisp/international/titdic-cnv.el
@@ -474,7 +474,6 @@ titdic-convert
     (with-temp-file  (tit-make-quail-package-file-name filename dirname)
       (let ((standard-output (current-buffer)))
 	(with-temp-buffer
-	  (set-buffer-multibyte nil)
 	  ;; Here we must use `raw-text' instead of `no-conversion' to
 	  ;; enable auto-decoding of eol format (CRLF->LF).
 	  (let ((coding-system-for-read 'raw-text))
@@ -504,7 +503,6 @@ titdic-convert
 		  (coding-system-change-eol-conversion coding-system 'unix))
 	    (remove-text-properties (point-min) (point-max) '(charset nil)))
 
-	  (set-buffer-multibyte t)
 	  ;; Set point the starting position of the body part.
 	  (goto-char (point-min))
 	  (if (not (search-forward "\nBEGIN" nil t))

BTW I did manage to successfully build with this patch applied.  Maybe
that second call should happen before the first call to the search
forward, i.e. just after the let form in the first hunk.

Now I get this weird warning, I can report that elsewhere if not
expected:

Error (initialization): Creation of the default fontsets failed: (error
Invalid script or charset name: old-italic)

This is from a build of master as of yesterday (edc63bf).  I don't get
this with 25.1.

--gk.

Re: Build failure caused by a .tit file

[Eli Zaretskii]

> From: Göktuğ Kayaalp <self@gkayaalp.com>
> Cc: help-gnu-emacs@gnu.org
> Date: Tue, 18 Apr 2017 16:16:12 +0300
> 
> A better version of this one is as follows, which removes the
> complementary set-buffer-multibyte call:

That's not the right fix.

> Now I get this weird warning, I can report that elsewhere if not
> expected:
> 
> Error (initialization): Creation of the default fontsets failed: (error
> Invalid script or charset name: old-italic)

You didn't fix the problem, it is still there.  The changes you made
just make the same problem appear somewhere else.

The old-italic script is defined in lisp/international/charscript.el,
do you see it there?

This all looks like you are trying to build the master branch with
some of the Lisp files from a very different version.  Is your
repository clean?  Did you try to clone it anew?

Re: Build failure caused by a .tit file

[Eli Zaretskii]

> From: Göktuğ Kayaalp <self@gkayaalp.com>
> Cc: help-gnu-emacs@gnu.org
> Date: Tue, 18 Apr 2017 15:41:45 +0300
> 
> Here it does indeed find the ‘\nBEGIN’ string.  The following patch
> fixes this error, but IDK if sanely:
> 
> diff --git a/lisp/international/titdic-cnv.el b/lisp/international/titdic-cnv.el
> index 130bc74..6a593e2 100644
> --- a/lisp/international/titdic-cnv.el
> +++ b/lisp/international/titdic-cnv.el
> @@ -474,7 +474,6 @@ titdic-convert
>      (with-temp-file  (tit-make-quail-package-file-name filename dirname)
>        (let ((standard-output (current-buffer)))
>  	(with-temp-buffer
> -	  (set-buffer-multibyte nil)
>  	  ;; Here we must use `raw-text' instead of `no-conversion' to
>  	  ;; enable auto-decoding of eol format (CRLF->LF).
>  	  (let ((coding-system-for-read 'raw-text))

Doesn't look right at all.

> | Thread 1 "bootstrap-emacs" received signal SIGABRT, Aborted.
> | 0x00007ffff0d59269 in raise (sig=sig@entry=6)
> |     at ../sysdeps/unix/sysv/linux/pt-raise.c:35
> | 35	../sysdeps/unix/sysv/linux/pt-raise.c: No such file or directory.
> | (gdb) bt
> | #0  0x00007ffff0d59269 in raise (sig=sig@entry=6)
> |     at ../sysdeps/unix/sysv/linux/pt-raise.c:35
> | #1  0x00000000004f3f04 in terminate_due_to_signal (sig=sig@entry=6, 
> |     backtrace_limit=backtrace_limit@entry=40) at emacs.c:393
> | #2  0x000000000050d573 in emacs_abort () at sysdep.c:2371
> | #3  0x00000000004aa614 in bidi_initialize () at bidi.c:1097
> | #4  0x00000000004ae45f in bidi_init_it (charpos=charpos@entry=1, 
> |     bytepos=1, frame_window_p=<optimized out>, 
> |     bidi_it=bidi_it@entry=0x7fffffffaf90) at bidi.c:1139

Does the file lisp/international/uni-bidi.el exist? is it non-empty?

If it does exist, do you have EMACSDATA or EMACSPATH variables set in
the environment?  Do you run Emacs from a shell or from inside another
Emacs session?

The backtrace indicates that something goes terribly wrong during
Emacs startup, and it doesn't find some of the files it needs for
that.

(And why are we still discussing this on help-gnu-emacs?)

Re: Build failure caused by a .tit file

[Göktuğ Kayaalp]

On 2017-04-18 17:56 +0300, Eli Zaretskii <eliz@gnu.org> wrote:
> Does the file lisp/international/uni-bidi.el exist? is it non-empty?

Yes.

> If it does exist, do you have EMACSDATA or EMACSPATH variables set in
> the environment?  Do you run Emacs from a shell or from inside another
> Emacs session?

Within a shell within Emacs.  Those variables are not set.

> The backtrace indicates that something goes terribly wrong during
> Emacs startup, and it doesn't find some of the files it needs for
> that.

Responding also to the other message of yours, my local copy of the repo
does not contain any extraneous files I know of, I just update it via
‘git pull’ every now and then.  I have a branch that I build my stable
versions from, but this troublesome build is on master.

> (And why are we still discussing this on help-gnu-emacs?)

I'll retry on the weekend from a pristine copy of the repo that I'll
fetch from the upsteam, and report the results.  Should I switch to
emacs-devel?

Re: Build failure caused by a .tit file

[Eli Zaretskii]

> From: Göktuğ Kayaalp <self@gkayaalp.com>
> Cc: help-gnu-emacs@gnu.org
> Date: Tue, 18 Apr 2017 19:19:59 +0300
> 
> > If it does exist, do you have EMACSDATA or EMACSPATH variables set in
> > the environment?  Do you run Emacs from a shell or from inside another
> > Emacs session?
> 
> Within a shell within Emacs.  Those variables are not set.

Try from a shell outside Emacs.

> Responding also to the other message of yours, my local copy of the repo
> does not contain any extraneous files I know of, I just update it via
> ‘git pull’ every now and then.  I have a branch that I build my stable
> versions from, but this troublesome build is on master.

What does "git status" say in the repo where you build?

One more thing to try: remove all the *.elc files in the source tree,
and try building again.

> > (And why are we still discussing this on help-gnu-emacs?)
> 
> I'll retry on the weekend from a pristine copy of the repo that I'll
> fetch from the upsteam, and report the results.  Should I switch to
> emacs-devel?

No, please use "M-x report-emacs-bug RET", which will send the report
to the bug tracker and also to the bug-gnu-emacs mailing list.  Then
let's take it from there (assuming building a fresh clone still
crashes or errors out).  The right place to discuss Emacs bugs is via
the bug tracker.  All the core maintainers subscribe to the bug list.

Thanks.