malc [Tue, 30 Aug 2016 17:37:52 +0000 (30 20:37 +0300)]
Phy. vs Str. equality strikes again
malc [Tue, 30 Aug 2016 16:28:44 +0000 (30 19:28 +0300)]
Real coordinates
Addresses https://github.com/moosotc/llpp/issues/34 also too ugly to live
malc [Mon, 29 Aug 2016 21:23:34 +0000 (30 00:23 +0300)]
Remove redundant entry
malc [Mon, 29 Aug 2016 21:06:08 +0000 (30 00:06 +0300)]
Fix visibility conditions
malc [Sun, 28 Aug 2016 15:59:14 +0000 (28 18:59 +0300)]
Types...
malc [Sun, 28 Aug 2016 15:28:56 +0000 (28 18:28 +0300)]
Uhm... shell is hard too
malc [Fri, 26 Aug 2016 15:06:43 +0000 (26 18:06 +0300)]
Fix vertical bar scrolling when maxfit is false
malc [Fri, 26 Aug 2016 00:05:11 +0000 (26 03:05 +0300)]
Smart(er) mouse wheel zooming
Maybe some day should be enabled as a config option.
malc [Thu, 25 Aug 2016 23:16:15 +0000 (26 02:16 +0300)]
Stop gap listview scrollbars solution
malc [Thu, 25 Aug 2016 22:24:56 +0000 (26 01:24 +0300)]
Fix horizontal scrollbar
malc [Thu, 25 Aug 2016 21:59:04 +0000 (26 00:59 +0300)]
Hide veritcal scroll bar when not needed
If the document fits the window vertically. This also matches
horizontal behavior.
malc [Thu, 25 Aug 2016 21:55:04 +0000 (26 00:55 +0300)]
Begin scrollbar overhaul
This is "wrong", but the old approach was very ad hoc and caused major
pains.
Random points:
a. Behavior with this patch lends itself to an en vogue way of fading
out bars
b. Scrollbars are hard - https://bugzilla.mozilla.org/show_bug.cgi?id=
1248118
c. Perhaps view and bar areas should be covered by different view
ports. This would be the "right" thing to do, but will probably
require substantial changes all over the place
malc [Thu, 25 Aug 2016 16:23:37 +0000 (25 19:23 +0300)]
Use centered zooming pervasively
malc [Thu, 25 Aug 2016 16:06:41 +0000 (25 19:06 +0300)]
Fix visible area calculation
malc [Thu, 25 Aug 2016 14:25:34 +0000 (25 17:25 +0300)]
Be greedy
malc [Thu, 25 Aug 2016 14:17:46 +0000 (25 17:17 +0300)]
Address #33
Takeaways:
a. scrollbars ought to be re-engineered
b. scrollbar doesn't end with an 'e'
malc [Thu, 25 Aug 2016 01:28:03 +0000 (25 04:28 +0300)]
Fix selection zooming
malc [Thu, 25 Aug 2016 01:05:57 +0000 (25 04:05 +0300)]
Fix placement of list view selection rectangle
malc [Thu, 25 Aug 2016 00:35:11 +0000 (25 03:35 +0300)]
Prevent scroll bar width double accounting
malc [Wed, 24 Aug 2016 10:11:09 +0000 (24 13:11 +0300)]
Do not pivotzoom when there are pending geometry changes
This is a sledgehammer approach but nothing better springs to mind at
the moment.
malc [Wed, 24 Aug 2016 01:00:49 +0000 (24 04:00 +0300)]
Fix vertical positioning with pivot zooming
Wrong calculations inside represent went unnoticed for far too long
due to the fact that by default window dimensions are the same
(900x900).
malc [Tue, 23 Aug 2016 21:24:14 +0000 (24 00:24 +0300)]
Fix rect zooming
And probably a lot more.
malc [Tue, 23 Aug 2016 21:15:18 +0000 (24 00:15 +0300)]
Zoom centering
Needs more work for (at least) mouse wheel I suppose.
malc [Sat, 20 Aug 2016 23:42:00 +0000 (21 02:42 +0300)]
Simplify
malc [Thu, 18 Aug 2016 23:02:47 +0000 (19 02:02 +0300)]
Use correct x coordinate when layouting
Sigh...
malc [Thu, 18 Aug 2016 19:10:53 +0000 (18 22:10 +0300)]
Use hypot instead of open coding it
Given that other C99 features crept in anyways.
malc [Sun, 7 Aug 2016 00:06:25 +0000 (7 03:06 +0300)]
Cosmetics
malc [Sat, 30 Jul 2016 17:06:21 +0000 (30 20:06 +0300)]
Morf gotoy into gotoxy
When old code needed to adjust horizontal position it just merrily set
the state.x to the needed value and called gotoy, that was wrong in
case of throttling - gotoy might fail, so that state.y wasn't updated
but state.x was uncodnitionally set and the result was in inconsistent
behavior - one could end up in a situation where some tiles were
rendered - immediately evicted from the cache and then scheduled to be
rendered again.
malc [Thu, 28 Jul 2016 01:39:23 +0000 (28 04:39 +0300)]
Suppress redisplay on invalidation
We can end up in a situation when throttling is active, redisplay is
scheduled (if post-processing fails) and then invalidate is called -
resulting in an empty layout and consequently flickering.
This is probably the third time when post-processing intricacies
lead to nasty visual issues.
Reproducer:
mkfifo /tmp/fifo
llpp some.pdf -remote /tmp/fifo &
echo 'keys ctrl-0' >/tmp/fifo
for i in $(seq 1 100); do
echo 'keys ctrl-+' >/tmp/fifo;
sleep 0.01;
done;
malc [Tue, 26 Jul 2016 22:25:18 +0000 (27 01:25 +0300)]
Cosmetics
malc [Tue, 26 Jul 2016 21:48:30 +0000 (27 00:48 +0300)]
Fix logcurrently (angle was forgotten)
malc [Fri, 15 Jul 2016 09:58:00 +0000 (15 12:58 +0300)]
Just lock, do not warn buyers
malc [Thu, 14 Jul 2016 20:21:59 +0000 (14 23:21 +0300)]
Drop timed lock inside post-processing
This introduced nasty usability regression - any failure to take the
lock during post-processing results in a "freeze" so to speak. Hence
indicate the failure due to magical -1 post-processing return code and
schedule re-display if seen from the OCaml side, meaning the display
and post-processing will be triggered on any future event (most likely
the notification from the rendering/loading side that will likely
release the lock and allow forward progress)
Things work fine with original offender (visualFAQ.pdf -
http://tug.ctan.org/info/visualFAQ/visualFAQ.pdf) as does scrolling
around (mainly horizontally) in sufficiently large and zoomed in
image, such as:
https://upload.wikimedia.org/wikipedia/commons/c/c4/Vasily_Surikov_-_%D0%91%D0%BE%D1%8F%D1%80%D1%8B%D0%BD%D1%8F_%D0%9C%D0%BE%D1%80%D0%BE%D0%B7%D0%BE%D0%B2%D0%B0_-_Google_Art_Project.jpg)
malc [Thu, 14 Jul 2016 12:57:06 +0000 (14 15:57 +0300)]
Robustify
malc [Thu, 14 Jul 2016 02:01:12 +0000 (14 05:01 +0300)]
Sync with upstream
malc [Tue, 12 Jul 2016 23:58:42 +0000 (13 02:58 +0300)]
Give ml_postprocess a chance to succeed
Admittedly this is fugly, but no clean solution immediately springs to
mind.
(Register continuation that will be called upon unlock and will notify
caml code to retry? Use cloned contexts? Something else?)
All other trylocks are just as ugly, but probably much less noticeable.
malc [Tue, 12 Jul 2016 22:30:33 +0000 (13 01:30 +0300)]
Fix locking
malc [Tue, 12 Jul 2016 21:46:16 +0000 (13 00:46 +0300)]
Add clang guard
malc [Tue, 12 Jul 2016 17:26:58 +0000 (12 20:26 +0300)]
Update ignore list
Tramp auxiliary files.
malc [Tue, 12 Jul 2016 17:26:20 +0000 (12 20:26 +0300)]
Sync with upstream
Also fix some issues reported by valgrind.
malc [Tue, 12 Jul 2016 11:44:22 +0000 (12 14:44 +0300)]
Revert a bunch of syncs
Things are rather sad. Segfaults, compilation problems, etc, on top of
that Tor is on vaccation. So let's take it slow and stick to a
slightly outdated mupdf, but at least it works (that's not to say that
those are mupdf problems)
malc [Tue, 12 Jul 2016 11:11:54 +0000 (12 14:11 +0300)]
Safety net
null pointer dereference with "Frank Frazetta - Fantasy Illustrated #2.pdf"
malc [Tue, 12 Jul 2016 10:39:23 +0000 (12 13:39 +0300)]
"Fix" trimctm
I don't have any idea what's going on here. Details:
mupdf
834243dbfc8ab3e9269b7a789d61e3fe0f8e71f7 - good
e2a85460370a3110b89fb838bc924c73f87f5b3c - bad
Most things work with either of those, except for:
http://www.gly.uga.edu/railsback/11111misc/SizeofThings36.pdf
malc [Sat, 9 Jul 2016 15:08:14 +0000 (9 18:08 +0300)]
Sync with upstream
malc [Fri, 8 Jul 2016 20:32:08 +0000 (8 23:32 +0300)]
Sync with upstream
malc [Wed, 6 Jul 2016 18:15:19 +0000 (6 21:15 +0300)]
Refactor
rotate is initialized the same way in both branches, so initialize it
once.
malc [Wed, 6 Jul 2016 17:50:36 +0000 (6 20:50 +0300)]
Sync with upstream
The trimctm stuff and the whole poling deep into fitz internals is
pretty fragile and disgusting, maybe this should be revisited.
Raison d'ĂȘtre for some of those "optimizations" was speed,
DDI0406B_arm_architecture_reference_manual_errata_markup_8_0.pdf
and
ECMA-376, Second Edition, Part 1 - Fundamentals And Markup Language Reference.pdf
[maybe things aren't as atrocious now (speed wise) as they were when some
of those "hacks" were introduced]
malc [Sat, 2 Jul 2016 19:45:25 +0000 (2 22:45 +0300)]
Consistency
malc [Sat, 2 Jul 2016 19:42:31 +0000 (2 22:42 +0300)]
Cosmetics
malc [Wed, 29 Jun 2016 11:29:17 +0000 (29 14:29 +0300)]
Abstract keyboard handling a bit more
malc [Wed, 29 Jun 2016 09:37:38 +0000 (29 12:37 +0300)]
Cosmetics
malc [Sat, 25 Jun 2016 17:50:02 +0000 (25 20:50 +0300)]
Kill stray space
malc [Sat, 25 Jun 2016 17:46:48 +0000 (25 20:46 +0300)]
Change splitatspace to return a pair
malc [Sat, 25 Jun 2016 12:54:06 +0000 (25 15:54 +0300)]
Simplify
malc [Fri, 24 Jun 2016 19:34:01 +0000 (24 22:34 +0300)]
Add remote `keys' command
malc [Fri, 24 Jun 2016 13:18:07 +0000 (24 16:18 +0300)]
Remove diagnostic pragmas around fitz/mupdf headers
Apparently the issues with warnings were fixed upstream.
malc [Thu, 23 Jun 2016 09:21:58 +0000 (23 12:21 +0300)]
Demacrofication
Replace MIN/MAX usage with calls to fz_min/max[i].
malc [Thu, 23 Jun 2016 08:20:29 +0000 (23 11:20 +0300)]
Sync with upstream
malc [Wed, 22 Jun 2016 15:17:19 +0000 (22 18:17 +0300)]
Fix typo
Right... Coud those sort of typos thinkos be avoided? No idea...
malc [Sat, 18 Jun 2016 09:20:43 +0000 (18 12:20 +0300)]
Sync with upstream
malc [Fri, 10 Jun 2016 17:33:41 +0000 (10 20:33 +0300)]
Cosmetics
We only need to build libs part of mupdf.
malc [Mon, 30 May 2016 21:34:04 +0000 (31 00:34 +0300)]
Refactor
malc [Mon, 30 May 2016 21:15:12 +0000 (31 00:15 +0300)]
Fix native builds
Thanks to Tuncer Ayaz for heads up. (Also this can use some refactoring).
malc [Thu, 26 May 2016 11:06:32 +0000 (26 14:06 +0300)]
Make hlint happy
malc [Wed, 25 May 2016 21:06:16 +0000 (26 00:06 +0300)]
Dedup
malc [Wed, 25 May 2016 11:30:17 +0000 (25 14:30 +0300)]
Cosmetics
Coincidentally fixes dyslexia.
malc [Tue, 24 May 2016 22:09:26 +0000 (25 01:09 +0300)]
Add fb2
malc [Tue, 24 May 2016 18:04:23 +0000 (24 21:04 +0300)]
Quoting
Strangely enough this works better.
malc [Tue, 24 May 2016 15:28:38 +0000 (24 18:28 +0300)]
Sync with upstream
malc [Tue, 24 May 2016 11:59:01 +0000 (24 14:59 +0300)]
More dedup
malc [Tue, 24 May 2016 11:49:26 +0000 (24 14:49 +0300)]
Avid duplication
malc [Tue, 24 May 2016 11:39:28 +0000 (24 14:39 +0300)]
Merge branch 'ghennequin-master'
malc [Tue, 24 May 2016 11:38:41 +0000 (24 14:38 +0300)]
Merge branch 'master' of https://github.com/ghennequin/llpp into ghennequin-master
Guillaume Hennequin [Tue, 24 May 2016 10:57:53 +0000 (24 11:57 +0100)]
Added more precise building instructions for mupdf in [BUILDING]
malc [Sat, 14 May 2016 20:29:33 +0000 (14 23:29 +0300)]
Trim more stray whitespace
malc [Sat, 14 May 2016 20:20:49 +0000 (14 23:20 +0300)]
Kill stray space
malc [Sat, 7 May 2016 18:41:44 +0000 (7 21:41 +0300)]
Add fz_var annotations where appropriate
Thanks to Tuncer Ayaz for heads up.
Worth metnioning that the warnings only trigger when source is
compiled with optimizations (those were set to 0 (i.e. -O0) over here
so were absent when the source was compiled locally)
malc [Tue, 3 May 2016 14:27:26 +0000 (3 17:27 +0300)]
Use proper matrix
Takes trimming into account among other things.
malc [Tue, 3 May 2016 14:13:35 +0000 (3 17:13 +0300)]
Remove pointless code
malc [Tue, 3 May 2016 13:29:00 +0000 (3 16:29 +0300)]
Fix hovering over final char in a region of interest
Thanks for heads up to Tor Andersson.
malc [Tue, 3 May 2016 12:04:20 +0000 (3 15:04 +0300)]
Add svg
malc [Fri, 29 Apr 2016 02:39:10 +0000 (29 05:39 +0300)]
Compress more
malc [Fri, 29 Apr 2016 02:36:49 +0000 (29 05:36 +0300)]
Compress
malc [Thu, 28 Apr 2016 18:42:48 +0000 (28 21:42 +0300)]
Merge branch 'tuncer-mupdf-release-native'
Tuncer Ayaz [Thu, 28 Apr 2016 18:26:11 +0000 (28 20:26 +0200)]
Add target build/llpp.murel.native
This expects a mupdf tree that's been built in release mode and is
needed because -march=native objects can use instructions that are not
supported by all CPUs.
malc [Wed, 27 Apr 2016 22:37:47 +0000 (28 01:37 +0300)]
Silence the compiler
w.r.t. lablGL
malc [Sat, 2 Apr 2016 22:47:48 +0000 (3 01:47 +0300)]
Cosmetics
malc [Fri, 1 Apr 2016 19:14:55 +0000 (1 22:14 +0300)]
Roll own ASCII lowercase functionality
4.03 deprecates lowercase functions, and lowercase_ascii is not
available on prior compilers, roll our own.
cf. https://sympa.inria.fr/sympa/arc/caml-list/2016-04/msg00005.html
malc [Mon, 28 Mar 2016 13:47:03 +0000 (28 16:47 +0300)]
Collapse extended parameters before exiting
Something has to be done with this mess.
malc [Mon, 28 Mar 2016 09:54:13 +0000 (28 12:54 +0300)]
Wikipedia mode
malc [Thu, 24 Mar 2016 20:42:27 +0000 (24 23:42 +0300)]
mupdf can handle bmps on its own now
malc [Thu, 24 Mar 2016 18:18:17 +0000 (24 21:18 +0300)]
Sync with upstream
moosotc [Sun, 20 Mar 2016 09:27:52 +0000 (20 13:27 +0400)]
Merge pull request #14 from nojb/fix-osx-x11-socket-name
OS X: recognize X11 unix socket path in $DISPLAY
malc [Mon, 14 Mar 2016 21:49:03 +0000 (15 00:49 +0300)]
Indentation
malc [Mon, 14 Mar 2016 17:45:38 +0000 (14 20:45 +0300)]
Sync with upstream
malc [Mon, 14 Mar 2016 13:47:21 +0000 (14 16:47 +0300)]
Shrug
malc [Mon, 29 Feb 2016 23:11:53 +0000 (1 02:11 +0300)]
Allow specifying output path name via command line
malc [Mon, 29 Feb 2016 06:19:16 +0000 (29 09:19 +0300)]
Use command line instead of clipboard to specify urls
malc [Mon, 29 Feb 2016 06:17:16 +0000 (29 09:17 +0300)]
Enable css hack for shake builds
malc [Mon, 29 Feb 2016 06:16:32 +0000 (29 09:16 +0300)]
Update