Adding : back to fl in global-aliases causes a display glitch

[cpitclaudel]

Here's a repro:

''(font-lock-test)
;; ^ here
;; Local Variables:
;; nameless-global-aliases:
;;    (("fl:" . "font-lock"))
;; End:

And here's how it displays for me:

Note the spurious colon overlapped with the f of fl:.

[Malabarba]

I've been getting loads of display glitches here, and I suspect it might be a real display bug in the master branch. Am I correct in assuming you're using the Emacs master branch?

[cpitclaudel]

Spot on. Time for a git bisect probably. You or me?

[Malabarba]

Oh dear, bisecting a display bug? I think I'm gonna leave that to you. :-)

If it helps I can tell you my build is from June 1st. So it's at least as old as that.

[cpitclaudel]

Ok :) I won't have time until a few days from now, though.

[Malabarba]

Yeah. It's a bug that needs visual confirmation, and probably needs an entire make before each testing step. It's probably going to take a while to hunt down. :-(

[cpitclaudel]

Hmm. I think can use a screenshot to automate the bisecting; that'll make for a fun challenge. I already use screenshots for regression testing in company-coq.

Last time I had to do such bisecting was for https://debbugs.gnu.org/cgi/bugreport.cgi?bug=21028 , and the three hours it took to pinpoint the commit are nothing compared to the multiple months that it took/will take to get it fixed :)

[Malabarba]

Hmm. I think can use a screenshot to automate the bisecting; that'll make for a fun challenge. I already use screenshots for regression testing in company-coq.

Now that sounds like a lot more fun. :-)

[cpitclaudel]

git bisect run ... is working on it right now. 1500 revisions to go, at about 15 minutes per bisect loop should be a few hours :)

[Malabarba]

😱

[cpitclaudel]

Ok, found it :) 🎉 after 130k lines of make output and 4 hours of automatic bisecting.

237244bbd5ce753bcdf79634561de515bd76c687 is the first bad commit
commit 237244bbd5ce753bcdf79634561de515bd76c687
Author: Paul Eggert <eggert@cs.ucla.edu>
Date:   Mon May 30 16:09:25 2016 -0700

    Omit IF_LINT code that no longer seems needed

    Nowadays GCC is smarter, or the Emacs code has mutated, or both,
    and now is as good a time as any to remove uses of IF_LINT that
    now seem to be unnecessary.
    * lib-src/emacsclient.c (set_local_socket):
    * lib-src/movemail.c (main) [MAIL_USE_MAILLOCK && HAVE_TOUCHLOCK]:
    * src/buffer.c (fix_start_end_in_overlays, fix_overlays_before):
    * src/casefiddle.c (casify_region):
    * src/charset.c (load_charset_map):
    * src/coding.c (decode_coding_object, encode_coding_object):
    * src/data.c (Fmake_variable_buffer_local, Fmake_local_variable)
    (cons_to_unsigned, cons_to_signed):
    * src/frame.c (make_frame, x_set_frame_parameters):
    * src/keyboard.c (read_event_from_main_queue):
    * src/regex.c (regex_compile):
    * src/syntax.c (back_comment):
    * src/window.c (Frecenter):
    * src/xfaces.c (Fx_list_fonts):
    Remove IF_LINT that no longer seems necessary.
    * src/image.c (png_load_body, jpeg_load_body): Simplify use of IF_LINT.
    * src/keyboard.c (read_char): Use IF_LINT (volatile) rather than
    a pragma dance to pacify GCC -Wclobbered.
    * src/xdisp.c (x_produce_glyphs): Rewrite to avoid need for IF_LINT.

    * src/xterm.c (x_connection_closed): Now _Noreturn, which should
    mean we do not need IF_LINT any more.
    (x_io_error_quitter): Now _Noreturn.  Put an 'assume (false)’
    at the end, to forestall warnings from older compilers.

:040000 040000 cc239faed6bf69466195aff82bbea45f8ae688de 9852e21ced17d2245500ec938cb32f30a00bc953 M  lib-src
:040000 040000 58ffe214536a59dcbca8e332fc3b7cf7a065e23d 1dcc927806ec3029d5de7f41a07d443c6013a3f6 M  src
bisect run success

With that commit I see ; with the one just before I see .

Can you help make a small repro that doesn't depend on nameless? That would be very nice. Then I can write the report :)

Here's the full trace:

# bad: [01030eed9395f5004e7d0721394697d1ca90cc2f] ; Spelling fixes
# good: [866501efe0fdc0c29448e0aaf8696eb0a3c8fcd6] * lisp/ChangeLog: Remove a duplicate
git bisect start 'HEAD' 'emacs-24.5'
# good: [02586bdee33aad75b6495f0949271c27d44d34a0] Revert my change to gnus-art.el
git bisect good 02586bdee33aad75b6495f0949271c27d44d34a0
# skip: [a0b7842700dfa3c730a1aed5fefa222c0f84605f] ; * lisp/emacs-lisp/autoload.el update-directory-autoloads): Fix typo in previous.
git bisect skip a0b7842700dfa3c730a1aed5fefa222c0f84605f
# skip: [4e559f22a00b38e067cde7863635c96d8790e4b3] * lisp/progmodes/project.el (project--read-regexp): Quote the identifier.
git bisect skip 4e559f22a00b38e067cde7863635c96d8790e4b3
# good: [bec57a486a2a40d7c770dab72a34cf6a4d17a5d0] Some final fixes in file notification before merging with master
git bisect good bec57a486a2a40d7c770dab72a34cf6a4d17a5d0
# good: [a4d882cd09507fa1f891984fc7435923de3566fe] Correct old cell name unbinding when renaming cell.
git bisect good a4d882cd09507fa1f891984fc7435923de3566fe
# good: [35d58020b6e173b137429786d174ffdb8894a812] Remove trailing blank lines in shr
git bisect good 35d58020b6e173b137429786d174ffdb8894a812
# good: [36906806ccfc0e53f1d8c365ab0d7151288b7833] Improve display of tex-verbatim and Info quoted
git bisect good 36906806ccfc0e53f1d8c365ab0d7151288b7833
# bad: [2f523c15801366d269cb4a526906e018f8cd713f] Make python.el work in Emacs 24
git bisect bad 2f523c15801366d269cb4a526906e018f8cd713f
# bad: [01030eed9395f5004e7d0721394697d1ca90cc2f] ; Spelling fixes
git bisect bad 01030eed9395f5004e7d0721394697d1ca90cc2f
# good: [21e87ece97e90f5500bd49b84dca08d97bd4f155] Fix Bug#23614.
git bisect good 21e87ece97e90f5500bd49b84dca08d97bd4f155
# skip: [4e182dce20bd09e38a69bdf637a3bf2a8bbfdc58] Port to platforms where rlim_max < 0
git bisect skip 4e182dce20bd09e38a69bdf637a3bf2a8bbfdc58
# good: [6ca36676c96954466bc0f1266115dff641697424] Repair damage from emacs-25 merge
git bisect good 6ca36676c96954466bc0f1266115dff641697424
# good: [062757dca01cff127583438ae16237af55b6253c] Sync with Tramp repository.
git bisect good 062757dca01cff127583438ae16237af55b6253c
# skip: [09b72fc38aee1032b6d2c8476bcc22a2e7ec5642] * lisp/emacs-lisp/cl-macs.el (cl-defstruct): Add docstring to accessors.
git bisect skip 09b72fc38aee1032b6d2c8476bcc22a2e7ec5642
# skip: [b342815c0a9af8d94d4290d17882de73f6fd9373] Improve define-function omitted-arg documentation
git bisect skip b342815c0a9af8d94d4290d17882de73f6fd9373
# good: [5cdebe7d8b692113c3577a6ded81cb149fe6f042] * src/puresize.h (BASE_PURESIZE): Increase it by a little bit.
git bisect good 5cdebe7d8b692113c3577a6ded81cb149fe6f042
# good: [4ce68f8f0f1dd1aa518e6e682b2c2cef1f49d953] * src/puresize.h (BASE_PURESIZE): Increase it some more.
git bisect good 4ce68f8f0f1dd1aa518e6e682b2c2cef1f49d953
# good: [66cd9187e396abfa7220d6a8f8d1a7064ef20b1e] Don’t document declare-function internals
git bisect good 66cd9187e396abfa7220d6a8f8d1a7064ef20b1e
# good: [1535aaf2873ddc6bcf113261800fc137a611e661] Fix Bug#23631 for Tramp
git bisect good 1535aaf2873ddc6bcf113261800fc137a611e661
# good: [cb379cbb7fc617f897e7dbc3ce45bf99ea3d1f87] Port --enable-gcc-warnings to Cygwin, FreeBSD
git bisect good cb379cbb7fc617f897e7dbc3ce45bf99ea3d1f87
# bad: [1992075d990ecf780225f446ec3f616bf821d89e] * src/conf_post.h (SYSTEM_PURESIZE_EXTRA) [CYGWIN]: Increase.
git bisect bad 1992075d990ecf780225f446ec3f616bf821d89e
# bad: [237244bbd5ce753bcdf79634561de515bd76c687] Omit IF_LINT code that no longer seems needed
git bisect bad 237244bbd5ce753bcdf79634561de515bd76c687
# first bad commit: [237244bbd5ce753bcdf79634561de515bd76c687] Omit IF_LINT code that no longer seems needed

[Malabarba]

Oh dear, it had to be a large commit. ¬¬

Can you help make a small repro that doesn't depend on nameless? That would be very nice. Then I can write the report :)

Absolutely. That I can do. :-) I'll write something later today (or tomorrow at the latest).

[Malabarba]

Also, is the bad commit in the emacs-25 branch? It would be a shame if a whole stable release got out that's incompatible with nameless. :-(

[cpitclaudel]

Absolutely. That I can do. :-) I'll write something later today (or tomorrow at the latest).

Yay, thanks!

Also, is the bad commit in the emacs-25 branch? It would be a shame if a whole stable release got out that's incompatible with nameless. :-(

No worries, it's only on master :) Phew.

[Malabarba]

Here you go. To reproduce simply:

1. Switch to a new buffer in fundamental-mode.
2. Insert some text into it.
3. Evaluate: (compose-region (point-min) (point-max) '(?b (Br . Bl) ?d)).

Previously, you should see bd. Instead, you just see the two letter overlayed on top of each other.

[cpitclaudel]

Thanks! Indeed, works perfect. I'll open the report now :)

[cpitclaudel]

https://lists.gnu.org/archive/html/bug-gnu-emacs/2016-07/msg00572.html

[cpitclaudel]

Confirmed fixed by 4ba2946369cb19dfeb258839add0658c742c12a1

[Malabarba]

Great! ☺