| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
|
|
|
|
This makes it easier to figure out what's going wrong when
we forget to define _POSIX_SOURCE.
|
|
|
|
Remove false 3d, add borders for visual separation between
content area and navigation area.
|
|
|
|
|
|
mischief provided the following test that shows the issue:
ramfs -S crash
aux/9pcon /srv/crash <<EOF
Tversion 8192 9P2000
Tattach 0 -1 $user ''
Tcreate 0 dir 020000000777 0
Tattach 5 -1 $user ''
Twalk 5 6 dir
Tread 6 0 512
EOF
the problem is that lib9p wrongly allowed reads on closed fids,
due to the permission check only considering the lower 2 bits.
a closed fid has fid->omode == -1 and it would pass on read for:
(-1 & 3) == 3 == OEXEC
the following change explicitely checks for for the closed case
and also rejects writes on directories (they are rejected on
open/create, but a broken 9p client could still issue the request).
|
|
|
|
when reclaiming pages from an image, always reclaim all
the hash chains equally. that way, we avoid being biased
towards the chains at the start of the Image.pghash[] array.
images can be in two states: active or inactive. inactive
images are the ones which are not used by program while
active ones aare.
when reclaiming pages, we should try to reclaim pages
from inactive images first and only if that set becomes
exhausted attempt to release text pages and attempt to
reclaim pages from active images.
when we run out of Image structures, it makes only sense
to reclaim pages from inactive images, as reclaiming pages
from active ones will never free any Image structures.
change putimage() to require a image already locked and
make it unlock the image. this avoids many pointless
unlock()/lock() sequences as all callers of putimage()
already had the image locked.
|
|
- nil check pl_blue allocation, and don't do it every time pl_rtdraw is called
- fix re-adding previously removed flushimage calls
- correct format for pointer
- sysfatal in pl_drawinit on error
|
|
if we do 'CAT(foo bar, baz quux)', the resulting token row
should have 3 tokens: 'foo', 'barbaz', 'quux'.
tested by jonasa, rebuilding /sys/src, perl, netsurf, and nuklear.
|
|
|
|
Modify aux/kbdfs to work with neo2 layout.
- add new modifier, increase limit on layers.
- Add Kaltgr for mod3
- Add mod4
This change only implements the alphanumeric
block of the keyboard in the kbmap, as neo2
is mainly used for blind typing and is not
needed that much anywhere else, this leaves
the numpad like it is.
(Thanks, jstsmthrgk@jstsmthrgk.eu)
|
|
|
|
from my forked mess by Ori_B.
|
|
grow selection from point of click, not start of selection region.
starting at the beginning of the selection region causes the match
logic to kick in, which is confusing.
|
|
the caller of macexpand() needs one more byte in
the buffer to append peekc.
make macexpand() actually check for buffer overflow.
just use strdup() to duplicate include file name
instead of the hunk dance.
move GETC() macro in cc.h
|
|
the results of the time calculation were garbled -- and
apparently negative on my system when testing, so the
test passed when it shouldn't have.
|
|
not what I wanted to commit.
|
|
the results of the time calculation were garbled -- and
apparently negative on my system when testing, so the
test passed when it shouldn't have.
|
|
|
|
see the last 2 commits.
|
|
same as sam and vt, three clicks selects a
whitespace-delimited line.
|
|
three clicks selects a whitespace-delimited line.
|
|
|
|
it is probably almost always a good idea to use
double buffering with a unaccelerated framebuffer.
|
|
size detection (thanks kenji)
looks like linux changed the device tree names for
the memory node:
https://github.com/raspberrypi/linux/commit/4b17654f51fd87c42cda3217b0863e6aa1df49b9#diff-ac03c9402b807c11d42edc9e8d03dfc7
this fixes the memory size detection with latest firmware
on raspberry pi4-b (4GB) for kenji.
|
|
|
|
sanity-checked by Ori_B.
|
|
never looked right outside of drawterm on osx (mea culpa). this is easy to change for anyone who prefers something else, but it's difficult to imagine native users preferring our very fuzzy dejavusans to anything else.
|
|
|
|
with scroll-wheel scrolling, this case can happen with the
command window:
umbraticus → cinap: create fullscreen file buffer by right-clicking in cmd window
|
|
|
|
|
|
|
|
|
|
|
|
add a buffer size argument to macexpand() and check for
overflow.
check for overflow when parsing #include directives.
|
|
rune
|
|
slookup() copies to symb, so use the symb[NSYMB] buffer directly
to declare type conversion functions and get rid of the arbitrary
sized local buffer. replace sprint() with snprint().
|
|
we were implicitly depending on UMINUS being the last entry
in the operator table -- that's fragile.
|
|
We used to treat all operators as right associative,
which means that we would evaluate them incorrecty.
For example, '2 - 1 + 1' would evaluate as '2 - (1 + 2)',
instead of '(2 - 1) + 1'.
This adds an assoc parameter to struct pri, and then uses
it to decide how to evaluate operators.
|
|
|
|
when pushing expressions in cpp, particularly complex ones could
overflow the stack and silently corrupt our data structures. add
checks when we push, and bump the stack size up.
|
|
|
|
|
|
|
