JasonWoof Got questions, comments, patches, etc.? Contact Jason Woofenden
st.git
9 years agoFix disabling of bold and fastblink
Michael Forney [Sat, 26 Jul 2014 07:48:15 +0000 (00:48 -0700)]
Fix disabling of bold and fastblink

According to ECMA-48¹ 8.3.117, an attribute value of 21 is "doubly
underlined", while 22 is "normal colour or normal intensity (neither
bold nor faint)".

Additionally, 25 is "steady (not blinking)", which likely means neither
slow blink nor fast blink.

¹: http://www.ecma-international.org/publications/files/ECMA-ST/Ecma-048.pdf

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoAdd info about Backspace and Delete to the FAQ
Wolfgang Corcoran-Mathe [Fri, 25 Jul 2014 17:32:29 +0000 (13:32 -0400)]
Add info about Backspace and Delete to the FAQ

Here is a modest attempt at cleaning it up a little bit. I changed a
few phrases that seemed awkward, but I think the content is the same.

--
Wolfgang Corcoran-Mathe

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoAdd info about Backspace and Delete to the FAQ
Roberto E. Vargas Caballero [Thu, 24 Jul 2014 17:56:58 +0000 (19:56 +0200)]
Add info about Backspace and Delete to the FAQ

9 years agoAdd information about librt and OpenBSD to the FAQ
Roberto E. Vargas Caballero [Sun, 20 Jul 2014 08:31:16 +0000 (10:31 +0200)]
Add information about librt and OpenBSD to the FAQ

9 years agoAdd information about librt and OpenBSD to the FAQ
Roberto E. Vargas Caballero [Sun, 20 Jul 2014 08:31:16 +0000 (10:31 +0200)]
Add information about librt and OpenBSD to the FAQ

9 years agoAdd information about librt and OpenBSD to the FAQ
Roberto E. Vargas Caballero [Sun, 20 Jul 2014 08:31:16 +0000 (10:31 +0200)]
Add information about librt and OpenBSD to the FAQ

9 years agoVery minor grammar fixes in FAQ
Wolfgang Corcoran-Mathe [Sun, 20 Jul 2014 15:05:39 +0000 (11:05 -0400)]
Very minor grammar fixes in FAQ

Self-explanatory.

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoFix st with input method.
Weng Xuetian [Wed, 16 Jul 2014 22:36:46 +0000 (18:36 -0400)]
Fix st with input method.

XFilterEvent need to be called against every event, otherwise it would
missing some message in the xim protocol and misbehave on some im server.

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoChanged inconsistent indent
Eric Pruitt [Mon, 14 Jul 2014 14:05:07 +0000 (09:05 -0500)]
Changed inconsistent indent

- A line was indented using spaces despite the rest of the code using
  tabs.

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoAdd 8 bit version of DCS, APC, PM, OSC
Roberto E. Vargas Caballero [Fri, 20 Jun 2014 07:51:18 +0000 (09:51 +0200)]
Add 8 bit version of DCS, APC, PM, OSC

DCS, APC, PM, OSC version for 7 bits environments already was implemented
in st.  This patch adds the 8 bit version of it.

9 years agoAdd 8 bit version of HTS
Roberto E. Vargas Caballero [Fri, 20 Jun 2014 07:51:18 +0000 (09:51 +0200)]
Add 8 bit version of HTS

HTS version for 7 bits environments already was implemented in st.
This patch adds the 8 bit version of it.

9 years agoAdd 8 bit version of NEL
Roberto E. Vargas Caballero [Fri, 20 Jun 2014 07:51:18 +0000 (09:51 +0200)]
Add 8 bit version of NEL

NEL version for 7 bits environments already was implemented in st.
This patch adds the 8 bit version of it.

9 years agoAdd 8 bit version of DECID
Roberto E. Vargas Caballero [Fri, 20 Jun 2014 07:51:18 +0000 (09:51 +0200)]
Add 8 bit version of DECID

DECID version for 7 bits environments already was implemented in st.
This patch adds the 8 bit version of it.

9 years agoRender faint attribute
Anders Eurenius [Sat, 21 Jun 2014 22:10:59 +0000 (00:10 +0200)]
Render faint attribute

Faint text is implemented by allocating a new color at one-half
intensity of each of the r, g, b components, or if the text bold at the
same time, it is not made lighter.

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>
Signed-off-by: Christoph Lohmann <20h@r-36.net>

9 years agoRemove all strcmp and strlen calls on Glyph.c[]
Ivan Delalande [Fri, 27 Jun 2014 07:15:56 +0000 (09:15 +0200)]
Remove all strcmp and strlen calls on Glyph.c[]

There were a few occurrences of strcmp and strlen being called on Glyph.c[],
which is not always null-terminated (this actually depends on the last values in
the buffer s in ttyread()). This patch replace all the calls to strcmp with a
test on c[0] directly or a call to tlinelen, and the one to strlen with utf8len.
I also took the opportunity to refactor getsel and tdumpline.

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoMaking surf compile again with rt.
Christoph Lohmann [Mon, 30 Jun 2014 15:33:43 +0000 (17:33 +0200)]
Making surf compile again with rt.

The clock_gettime patch needs librt to be linked to surf.

Thanks _odie for the hint.

9 years agoRender struck-out attribute
Anders Eurenius [Sat, 21 Jun 2014 18:32:34 +0000 (20:32 +0200)]
Render struck-out attribute

Implement crossed-out text with an XftDrawRect call, similar to how
underline is implemented. The line is drawn at 2/3 of the font ascent,
which seems to work nicely in practice.

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoRender invisible attribute
Anders Eurenius [Sat, 21 Jun 2014 18:30:22 +0000 (20:30 +0200)]
Render invisible attribute

Implement invisible mode by setting the foreground color to be the same
as the background color. Not rendering anything would also be an
alternative, but this seems less likely to cause surprises in
conjunction with any hacks.

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoReorder and extend glyph attributes
Anders Eurenius [Sat, 21 Jun 2014 18:29:36 +0000 (20:29 +0200)]
Reorder and extend glyph attributes

Faint, invisible, struck and fast blink are added as glyph attributes.
Since there's an edit here, let's take the opportunity to reorder them
so that they correspond to the two's power of the corresponding escape
code. (just for neatness, let's hope that property never gets used for
anything.)

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoRemove CEIL macro
Roberto E. Vargas Caballero [Thu, 26 Jun 2014 10:37:06 +0000 (12:37 +0200)]
Remove CEIL macro

This macro was not correct in some cases, and it was used only in
one place, where we did'nt get any benefit in performance of in size,
so the macro is removed and ceilf is used instead of it. The only
function needed from math.h is ceilf, so this patch defines the
prototype of it instead of including math.h.

9 years agoFixed wrong nanosecond factor 10E6.
Ivan Delalande [Thu, 26 Jun 2014 02:30:43 +0000 (04:30 +0200)]
Fixed wrong nanosecond factor 10E6.

Commit 5edeec1 introduced a wrong factor for nanosecond computation, the correct
value is 1E6. Time and timeout values are 10 times less than they should be and
this cause high CPU usage.

Reported by pyroh on IRC. Thanks!

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoUse monotonic clock to prevent timing issues
FRIGN [Mon, 23 Jun 2014 14:51:51 +0000 (16:51 +0200)]
Use monotonic clock to prevent timing issues

This patch replaces the gettimeofday()/timeval-system with
uses of clock_gettime() with a monolithic clock and timespec-structs.
gettimeofday() is not accurate and prone to jumps and POSIX.1-2008
marks it as obsolete. Read more here [0].

The patch should speak for itself and decreases the binary
size for me by almost 200K(!).

[0]: http://blog.habets.pp.se/2010/09/gettimeofday-should-never-be-used-to-measure-time

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoFix typo in config.def.h
Eric Pruitt [Wed, 18 Jun 2014 13:39:57 +0000 (08:39 -0500)]
Fix typo in config.def.h

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoupdate size hints on zoom
Troy Sankey [Wed, 18 Jun 2014 17:41:26 +0000 (10:41 -0700)]
update size hints on zoom

On font zooming (i.e. xzoom()), window size hints are not updated.  This
patch does that.

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoRefactor the innermost loop of the xdraws function
Silvan Jegen [Fri, 6 Jun 2014 20:09:22 +0000 (22:09 +0200)]
Refactor the innermost loop of the xdraws function

Signed-off-by: Silvan Jegen <s.jegen@gmail.com>
Signed-off-by: Christoph Lohmann <20h@r-36.net>

9 years agoFAQ: update typo and escape code for smkx
Hiltjo Posthuma [Mon, 9 Jun 2014 15:19:15 +0000 (17:19 +0200)]
FAQ: update typo and escape code for smkx

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoStyle police.
Christoph Lohmann [Sat, 7 Jun 2014 11:49:04 +0000 (13:49 +0200)]
Style police.

9 years agoSimplify tdeftrans
Roberto E. Vargas Caballero [Sat, 7 Jun 2014 11:23:45 +0000 (13:23 +0200)]
Simplify tdeftrans

9 years agoRemove unnecessary typedef
FRIGN [Fri, 6 Jun 2014 12:08:51 +0000 (14:08 +0200)]
Remove unnecessary typedef

This should also fix compiling-errors on OpenBSD,
as reported by Nils R. Thanks!

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoRefactor selsnap SNAP_WORD.
Colona [Thu, 5 Jun 2014 04:32:01 +0000 (06:32 +0200)]
Refactor selsnap SNAP_WORD.

Refactor the SNAP_WORD part in selsnap, and fix a bug that caused the word
delimiters to be ignored if it was at the very beginning or end of a wrapped
line.

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoFix rectangular selection.
Colona [Wed, 4 Jun 2014 04:47:55 +0000 (21:47 -0700)]
Fix rectangular selection.

selsort computes the wrong normalized coordinates when rectangular
selection is enabled, causing rectangular selection to only work
when going toward either the top left corner, or the bottom right
one.

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agotiny cleanup
Hiltjo Posthuma [Tue, 3 Jun 2014 15:55:53 +0000 (17:55 +0200)]
tiny cleanup

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoMake selection consistent over line breaks.
Colona [Tue, 3 Jun 2014 02:34:58 +0000 (04:34 +0200)]
Make selection consistent over line breaks.

Currently, selection is expanded to the end of the line over line breaks only in
regular selection mode, when the line in not empty and when going down and/or
right. This covers all the cases including word selection mode, with the
exception of rectangular selection because it would make this mode too rigid.
This adjustment is made in selsort so I renamed it to selnormalize to better
reflect what it does now.

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoFixing color and refactor xsetcolorname.
Christoph Lohmann [Sun, 1 Jun 2014 15:08:16 +0000 (17:08 +0200)]
Fixing color and refactor xsetcolorname.

By  the recommendation of FRIGN I refactored xsetcolorname to remove the
unnecessary r, g, b variables when allocating a new  color.  Colors  are
now freed and set to the new color. A die() should not happen here. Oth‐
erwise it is easy for applications to kill st. St should be resilent  to
malicious input.

Second  this  patch  standardises  the  naming  of  »color«. There is no
»colour« here. Maybe in some parts of the world.

9 years agoRevert "Refactor xsetcolorname()"
Christoph Lohmann [Sun, 1 Jun 2014 14:52:19 +0000 (16:52 +0200)]
Revert "Refactor xsetcolorname()"

This reverts commit a32c5f5726f514b49bd396f27aab0e78c40126d3.

9 years agoRefactor xsetcolorname()
FRIGN [Sun, 25 May 2014 10:13:22 +0000 (12:13 +0200)]
Refactor xsetcolorname()

I mainly improved the slightly off algorithm used to load colours in the 256-colour-space and
removed unnecessary local values (r,g,b,colour).
"colour" is not necessary as a punchbag for XftColorAlloc[Value,Name], as they don't mess with
the result-adress until they are absolutely sure everything worked out[0].

Being at it, I changed the error-returns for AllocValue to dies (just like in xloadcols()), as
a failure is most likely an OOM-situation you better catch early.
In case of an invalid name everything stays the same.

[0]: http://www.opensource.apple.com/source/X11libs/X11libs-40/libXft/libXft-2.1.13/src/xftcolor.c

Signed-off-by: Christoph Lohmann <20h@r-36.net>

9 years agoDon't report release events for mouse wheel
Balazs Kezes [Sat, 31 May 2014 21:24:58 +0000 (22:24 +0100)]
Don't report release events for mouse wheel

Signed-off-by: Christoph Lohmann <20h@r-36.net>

9 years agoFixing italic bold.
Christoph Lohmann [Sun, 1 Jun 2014 13:54:28 +0000 (15:54 +0200)]
Fixing italic bold.

Thanks Felipe Spychalski <spychalski@gmail.com> for the patch!

9 years agoFixed copying empty lines inside selection.
Alexander [Mon, 26 May 2014 05:23:56 +0000 (09:23 +0400)]
Fixed copying empty lines inside selection.

The code was assuming that empty lines have implicit wrap-around attribute.

Signed-off-by: Roberto E. Vargas Caballero <k0ga@shike2.com>

9 years agoFix some more LICENSE changes.
Christoph Lohmann [Sat, 24 May 2014 19:27:54 +0000 (21:27 +0200)]
Fix some more LICENSE changes.

9 years agoAdd FRIGN to License, remove trailing whitespace
FRIGN [Sat, 24 May 2014 19:07:53 +0000 (21:07 +0200)]
Add FRIGN to License, remove trailing whitespace

Signed-off-by: Christoph Lohmann <20h@r-36.net>

9 years agoBrightening is again on the TODO list.
Christoph Lohmann [Sat, 24 May 2014 11:52:24 +0000 (13:52 +0200)]
Brightening is again on the TODO list.

9 years agoFixing trailing whitespaces.
Christoph Lohmann [Sat, 24 May 2014 11:48:44 +0000 (13:48 +0200)]
Fixing trailing whitespaces.

9 years agoFix colour-model and simplify xloadcols()
FRIGN [Thu, 22 May 2014 13:00:33 +0000 (15:00 +0200)]
Fix colour-model and simplify xloadcols()

Signed-off-by: Christoph Lohmann <20h@r-36.net>

9 years agoAllow mouse selection override using ShiftMask
Hiltjo Posthuma [Mon, 12 May 2014 12:39:37 +0000 (14:39 +0200)]
Allow mouse selection override using ShiftMask

Similar to xterm or urxvt holding shift before selecting text with the mouse
allows to override copying text. For example in tmux with "mode-mouse on" or
vim (compiled with --with-x), mc, htop, etc.

forceselmod in config.h sets the modifier to use this mode, by default
ShiftMask.

Signed-off-by: Hiltjo Posthuma <hiltjo@codemadness.org>

9 years agoRedraw needs all dirty lines to have flash etc. work.
Christoph Lohmann [Fri, 9 May 2014 15:12:58 +0000 (17:12 +0200)]
Redraw needs all dirty lines to have flash etc. work.

9 years agoEnd a sequence only on CAN, SUB, \a and C1s.
Colona [Fri, 9 May 2014 08:23:53 +0000 (10:23 +0200)]
End a sequence only on CAN, SUB, \a and C1s.

9 years agoAlso clears ESC_START on interrupt characters during sequences.
Colona [Wed, 7 May 2014 08:05:27 +0000 (10:05 +0200)]
Also clears ESC_START on interrupt characters during sequences.

Otherwise, the rest of the input is interpreted as a new escape
sequence.
For the ESC character, ESC_START is re-set in tcontrolcode.

9 years agoConsistent FALLTHROUGH comments.
noname [Fri, 25 Apr 2014 14:26:54 +0000 (18:26 +0400)]
Consistent FALLTHROUGH comments.

9 years agoFix displaying control code
Roberto E. Vargas Caballero [Tue, 29 Apr 2014 07:58:55 +0000 (09:58 +0200)]
Fix displaying control code

Control code are never displayed. It is not important if graphic
charset is displayed or not.

9 years agoRemove ATTR_GFX and tselcs
Roberto E. Vargas Caballero [Tue, 29 Apr 2014 06:58:10 +0000 (08:58 +0200)]
Remove ATTR_GFX and tselcs

ATTR_GFX was used long time ago to detect when terminal was in
graphic mode. Today graphic mode is implemented using a charset
pointer, so ATTR_GFX is not needed anymore because graphic
condition can be detected directly checking if current charset
is GRAPHICS C0.

9 years agoDo not save cursor in tresize.
noname [Mon, 28 Apr 2014 22:35:22 +0000 (02:35 +0400)]
Do not save cursor in tresize.

This patch fixes the bug introduced in
8f11e1cd034ff28ca47bb4955505db7fa8016ba8

To reproduce the bug:
1. Save cursor: printf '\e[s'
2. Load cursor: printf '\e[u'
3. Resize st window.
4. Load cursor again: printf '\e[u'

9 years agoSimplify selected().
noname [Sun, 27 Apr 2014 11:16:41 +0000 (15:16 +0400)]
Simplify selected().

9 years agoDo not export chscale and cwscale.
noname [Sun, 27 Apr 2014 11:39:53 +0000 (15:39 +0400)]
Do not export chscale and cwscale.

9 years agoSimplify xunloadfonts.
noname [Sun, 27 Apr 2014 22:41:17 +0000 (02:41 +0400)]
Simplify xunloadfonts.

9 years agoRemove one indentation level in getsel().
noname [Sun, 27 Apr 2014 11:40:23 +0000 (15:40 +0400)]
Remove one indentation level in getsel().

9 years agoRemove unused dump() function.
noname [Sun, 27 Apr 2014 11:40:01 +0000 (15:40 +0400)]
Remove unused dump() function.

9 years agoAdd missing function prototypes.
noname [Sun, 27 Apr 2014 11:39:47 +0000 (15:39 +0400)]
Add missing function prototypes.

9 years agoUse MODBIT in xseturgency.
noname [Sun, 27 Apr 2014 11:39:39 +0000 (15:39 +0400)]
Use MODBIT in xseturgency.

9 years agoCompute ena_sel as one expression.
noname [Sun, 27 Apr 2014 11:39:29 +0000 (15:39 +0400)]
Compute ena_sel as one expression.

9 years agoSimplify expressions in tputc()
Roberto E. Vargas Caballero [Sun, 27 Apr 2014 08:25:15 +0000 (10:25 +0200)]
Simplify expressions in tputc()

9 years agoFix misplaced break
Roberto E. Vargas Caballero [Sat, 26 Apr 2014 07:30:53 +0000 (09:30 +0200)]
Fix misplaced break

This misplaced break was causing an incorrect fall through
from DSR to DECSTBM.

9 years agoCreate a function for DEC test
Roberto E. Vargas Caballero [Sat, 26 Apr 2014 07:24:04 +0000 (09:24 +0200)]
Create a function for DEC test

Almost of the sequences execute their action in a separate function,
which is good because helps to read the full set of sequences
faster.

9 years agoFix tputc control code handling
Roberto E. Vargas Caballero [Mon, 28 Apr 2014 16:32:09 +0000 (18:32 +0200)]
Fix tputc control code handling

The patch 53105cf modified how control codes were detected, because
it tried to handle also C1 control codes (0x80-0x9f), that have
upper bit to 1, so they are multi byte character in utf8.
Code was checking the value of width in order to known that after
decoding the unicode point had a width of 1 byte, but it as incorrect
because this width is the columnb width.

9 years agoRemove repeated initialisation of term.esc
Roberto E. Vargas Caballero [Fri, 25 Apr 2014 23:45:10 +0000 (01:45 +0200)]
Remove repeated initialisation of term.esc

Once a sequence is completed term.esc must return to 0, so
instead of repeating this expression in all the cases is
better put it at the end of the block.

9 years agoCancel DCS with SUB, CAN, ESC or any CC1 code
Roberto E. Vargas Caballero [Fri, 25 Apr 2014 23:34:46 +0000 (01:34 +0200)]
Cancel DCS with SUB, CAN, ESC or any CC1 code

From http://www.vt100.net/docs/vt510-rm/chapter4:

*The VT510 ignores all following characters until it receives a
 SUB, ST, or any other C1 control character.

So OSC, PM and APC sequence ends with a SUB (it cancels the sequence
and show a question mark as error), ST or any another C1 (8 bits)
code, or their C0 (7 bits) equivalent sequences (at this moment we
do not handle C1 codes, but we should). But it is also said that:

Cancel  CAN
1/8     Immediately cancels an escape sequence, control sequence,
or device control string in progress. In this case, the
VT510 does not display any error character.

Escape  ESC
1/11    Introduces an escape sequence. ESC also cancels any escape
sequence, control sequence, or device control string in
progress.

9 years agoadd break;s for last cases in switch statements
Markus Teich [Sat, 26 Apr 2014 21:50:37 +0000 (23:50 +0200)]
add break;s for last cases in switch statements

9 years agoDo not eat ESC character if control string is not properly terminated.
noname [Fri, 25 Apr 2014 22:12:41 +0000 (00:12 +0200)]
Do not eat ESC character if control string is not properly terminated.

Currently tputc handles the case of too long control string waiting for
the end of control string.

Another case is when there is ESC character is encountered but is not
followed by '\\'.  In this case st stops processing control string,
but ESC character is ignored.

After this patch st processes ESC characters in control strings properly.

Test case:
printf '\e]0;abc\e[1mBOLD\e[0m'

Also ^[\ is actually processed in the code that handles ST.
According to ECMA-048 ST stands for STRING TERMINATOR and is used to
close control strings.

9 years agoFix for multibyte characters in techo.
noname [Fri, 25 Apr 2014 14:28:00 +0000 (18:28 +0400)]
Fix for multibyte characters in techo.

Works for both signed and unsigned char.

9 years agos/DSC/DCS/ DCS stands for DEVICE CONTROL STRING
noname [Fri, 25 Apr 2014 14:27:48 +0000 (18:27 +0400)]
s/DSC/DCS/ DCS stands for DEVICE CONTROL STRING

9 years agoUse xwrite instead of write.
noname [Fri, 25 Apr 2014 14:27:26 +0000 (18:27 +0400)]
Use xwrite instead of write.

9 years agoRemove unnecessary break
noname [Fri, 25 Apr 2014 14:27:15 +0000 (18:27 +0400)]
Remove unnecessary break

9 years agoComment fix.
noname [Fri, 25 Apr 2014 14:26:43 +0000 (18:26 +0400)]
Comment fix.

9 years agoOn terminal resize, clear the alt screen with its own cursor.
Colona [Thu, 24 Apr 2014 18:35:41 +0000 (20:35 +0200)]
On terminal resize, clear the alt screen with its own cursor.

Currently the alternate screen get messed up on resize if it has
different colors or mode.

9 years agoOptimize tputtab.
noname [Wed, 23 Apr 2014 19:12:45 +0000 (23:12 +0400)]
Optimize tputtab.

Before this patch executing
printf '\e[10000000000I'
or
printf '\e[10000000000Z'
resulted in long delay.

9 years agoUse != instead of ^ for logical values.
noname [Wed, 23 Apr 2014 19:12:36 +0000 (23:12 +0400)]
Use != instead of ^ for logical values.

sel.alt is only changed by
sel.alt = IS_SET(MODE_ALTSCREEN);

9 years agoUse BETWEEN macro in xsetcolorname and fix style.
noname [Wed, 23 Apr 2014 19:12:29 +0000 (23:12 +0400)]
Use BETWEEN macro in xsetcolorname and fix style.

9 years agoConformity in the -g geometry handling.
Christoph Lohmann [Fri, 25 Apr 2014 20:34:24 +0000 (22:34 +0200)]
Conformity in the -g geometry handling.

Thanks to Yuri Karaban for suggesting this!

These changes make -g correspond to <cols>x<rows> and honor it so non-tiling
window managers can work with the size hints afterwards. It also adds a -i
flag to force the window size. This is needed so -g keeps being useful in dwm.

9 years agoSimplify a bit more tdeletechar and tinsertblank
Roberto E. Vargas Caballero [Fri, 25 Apr 2014 15:24:12 +0000 (17:24 +0200)]
Simplify a bit more tdeletechar and tinsertblank

The large and repeated expression used in memmove to indirect
the line can be simplified using a pointer, that makes more
clear where begins and where ends the movement.

9 years agoSimplify tdeletechar and tinsertblank and fix memory corruption.
noname [Tue, 22 Apr 2014 22:08:13 +0000 (02:08 +0400)]
Simplify tdeletechar and tinsertblank and fix memory corruption.

Current CSI parsing code uses strtol to parse arguments and allows them
to be negative. Negative argument is not properly handled in tdeletechar
and tinsertblank and results in memory corruption in memmove.

Reproduce with printf '\e[-500@'

Patch also removes special handling for corner case and simplifies
the code.

Removed
term.dirty[term.c.y] = 1
because tclearregion sets dirty flag.

9 years agoStyle fixes in tscrollup.
noname [Sun, 20 Apr 2014 13:26:39 +0000 (17:26 +0400)]
Style fixes in tscrollup.

9 years agoDo not set dirty flag twice in tscrollup and tscrolldown.
noname [Sun, 20 Apr 2014 13:26:50 +0000 (17:26 +0400)]
Do not set dirty flag twice in tscrollup and tscrolldown.

9 years agoUse tsetdirt in tscrollup and tscrolldown.
noname [Sun, 20 Apr 2014 13:26:13 +0000 (17:26 +0400)]
Use tsetdirt in tscrollup and tscrolldown.

tscrollup and tscrolldown do not use tsetdirt, but their code is
equivalent to

        tsetdirt(orig, term.bot-n);
        tsetdirt(orig+n, term.bot);

tclearregion also marks cleared lines as dirty.
In tscrolldown it sets lines from term.bot-n+1 to term.bot dirty, and in
tscrollup it sets lines from orig to orig+n-1 dirty.

In both functions all lines from orig to term.bot are effectively set
dirty, but in tscrolldown lines from orig+n to term.bot are set dirty
twice, and in tscrollup lines from orig to term.bot-n are set dirty
twice.

These patches make it clear which lines are set dirty and sets them
dirty once in each funciton.

9 years agoFix techo handling of control and multibyte characters.
noname [Tue, 22 Apr 2014 20:26:07 +0000 (00:26 +0400)]
Fix techo handling of control and multibyte characters.

techo compares signed char to '\x20'. Any character with code less then
'\x20' is treated as control character.  This way characters with MSB
set to 1 are considered control characters too.

Also this patch makes techo display DEL character as ^?.

To reprocuce the bug, enable echo mode using printf '\e[12l',
then type DEL character or any non-ASCII character.

9 years agoMake xrealloc and xstrdup style consistent.
noname [Tue, 22 Apr 2014 17:59:48 +0000 (21:59 +0400)]
Make xrealloc and xstrdup style consistent.

9 years agoUse BETWEEN in tsetchar.
noname [Tue, 22 Apr 2014 17:59:39 +0000 (21:59 +0400)]
Use BETWEEN in tsetchar.

9 years agoUse uint and uchar instead of unsigned int and unsigned char.
noname [Tue, 22 Apr 2014 17:59:32 +0000 (21:59 +0400)]
Use uint and uchar instead of unsigned int and unsigned char.

9 years agoRemove argument names from function prototypes.
noname [Tue, 22 Apr 2014 17:59:21 +0000 (21:59 +0400)]
Remove argument names from function prototypes.

9 years agoStyle fix in tdumpsel.
noname [Tue, 22 Apr 2014 17:59:14 +0000 (21:59 +0400)]
Style fix in tdumpsel.

9 years agoUse BETWEEN in tinsertblankline and tdeleteline.
noname [Tue, 22 Apr 2014 17:59:01 +0000 (21:59 +0400)]
Use BETWEEN in tinsertblankline and tdeleteline.

9 years agoIt's 2014 now.
Christoph Lohmann [Tue, 22 Apr 2014 17:22:58 +0000 (19:22 +0200)]
It's 2014 now.

10 years agomove MODBIT to Macros section
noname@inventati.org [Sun, 20 Apr 2014 09:08:09 +0000 (13:08 +0400)]
move MODBIT to Macros section

Patch moves MODBIT to macros section and uses it in tselcs.

10 years agomisplaced parenthesis in LEN macro
noname@inventati.org [Sun, 20 Apr 2014 11:41:40 +0000 (15:41 +0400)]
misplaced parenthesis in LEN macro

10 years agoremove confusing SERRNO macro
Markus Teich [Sun, 20 Apr 2014 11:15:40 +0000 (13:15 +0200)]
remove confusing SERRNO macro

I found the SERRNO Macro slightly confusing, since you have to look
it up, if you don't know it already. A web search showed it does
not seem to be any kind of standard. Also there was no reason in
the commit log when it was introduced in 2009. As you can see it
also leads to new patches, which don't use this macro (probably the
author did not know about it).

10 years agoSimplify techo
noname@inventati.org [Sun, 13 Apr 2014 16:21:32 +0000 (20:21 +0400)]
Simplify techo

Remove special case for \e because it is handled well by "control code"
case.

10 years agotypedefs instead of #defines
noname@inventati.org [Sat, 5 Apr 2014 16:54:01 +0000 (20:54 +0400)]
typedefs instead of #defines

Replaced #defines with typedefs where possible, patch attached.

10 years agofix cursor handling when alt screen is disabled
q@c9x.me [Wed, 9 Apr 2014 18:37:23 +0000 (20:37 +0200)]
fix cursor handling when alt screen is disabled

I don't like this alt screen thing, but when
allowaltscreen == 0, the cursor is still saved
and restored after calling 'less' (or 'man').

This patch makes allowaltscreen == 0 usable.

10 years agoMerge remote-tracking branch 'origin/master' into omaster
Roberto E. Vargas Caballero [Tue, 15 Apr 2014 06:11:47 +0000 (08:11 +0200)]
Merge remote-tracking branch 'origin/master' into omaster

10 years agoRemoving two unnecessary gettimeofday().
Christoph Lohmann [Fri, 11 Apr 2014 16:11:25 +0000 (18:11 +0200)]
Removing two unnecessary gettimeofday().

10 years agoIncrement XSync consistency
Toaster Toaster [Mon, 7 Apr 2014 17:53:59 +0000 (10:53 -0700)]
Increment XSync consistency

This patch cleans up a style inconsistency.