That is: colorize text after a blank-plus-hash as a comment when there
is at least one non-hexadecimal character among the first three.
This relaxes the demand that a trailing comment should have a blank
after the hash too, as implemented by commit 90946c5e two weeks ago.
Make types a bluish green, for more contrast with strings, so that the
latter can be made greener, for more contrast with the yellow keywords.
Make the main keywords bluer and less bright, and unbold and soften the
yellow ones. Soften the comments too.
Reference for case-insensitivity:
https://www.postgresql.org/docs/14/sql-syntax-lexical.html#SQL-SYNTAX-IDENTIFIERS
Also, colorize comments more readably (on a dark background), and
schemas less yelling, and languages more distinct from keywords.
And add a few more keywords, like IS and NULL.
Also, combine some keywords that can only occur together.
And drop a few that are probably rare, like LARGE OBJECT.
Reference:
https://www.postgresql.org/docs/14/bookindex.html
And use a more readable color -- blue is too dark on a black base.
Regular expressions in SQL are not written between // or %r{},
strings are not written between <> or %Q{} or anything similar,
and comments do not start with # -- and officially not with //
either, but some dialects seem to allow it.
This fixes https://savannah.gnu.org/bugs/?61381.
Bug existed since version 2.4.2, since the SQL syntax was included.
The path after 'include' is not a regex and does not need to be quoted;
when it's unquoted, colorize it specially, instead of leaving it red.
The syntax name after 'extendsyntax' should be brightgreen, like after
'syntax', but it's better that the subsequent commands have their own
special color.
Saying "Unbound key" would imply that it could be bound in a nanorc.
This fixes https://savannah.gnu.org/bugs/?61574.
Bug existed in this form since version 2.6.0, commit e0c4f9c5.
Letting the dummy key KEY_FRESH do a full refresh in the tiny version
allows a 'fg' to properly resume after an external SIGSTOP.
(Reminder: the ungetch() is needed in the normal version to prevent
the eating of a keystroke: https://savannah.gnu.org/bugs/?48331.)
Achieve this by translating KEY_SUSPEND back to ^Z.
(Strictly speaking, this is incorrect, as the terminfo entry
might specify some other keystroke for 'kspd', but... meh.)
This fixes https://savannah.gnu.org/bugs/?61573.
Bug existed since before version 2.2.0, since keys became rebindable.
With --zero, any message that was there will get overwitten by the
text window anyway as result of the current operation. And with
--minibar, any message will get overwritten with the minibar.
This fixes https://savannah.gnu.org/bugs/?61539.
Bug existed since commit 03637030 from two weeks ago.
The unsetting should happen after the three possible early returns,
to avoid leaving the flag unset.
Also, don't bother to exclude the setting and unsetting from the
tiny version, to get rid of some cluttering conditionals.
Most terminal emulators support a palette of 216 indexed colors:
the 6x6x6 color cube pioneered by Xterm. But as an index gives
little indication about what kind of color it produces, allow the
user to specify the color with three hex digits instead, one for
the level of red, green, and blue each, in the format #rgb.
One hex digit allows for sixteen values, but as there are only
six levels available, only 0, 4, 8, a, c, and e are significant
-- other values get reduced to the nearest lower one.
This fulfills https://savannah.gnu.org/bugs/?56445.
Requested-by: Peter Liscovius, Erik Lundin, Owen Maresh, Dave Lemonby
Signed-off-by: Brad Town <brad@bradtown.com>
Only requested information (^C, M-D) and alerts stay onscreen until the
next keystroke, but less important messages are overwritten after 1.5
seconds (or 0.8 seconds with --quick). This makes the behavior more
consistent with dropping a message when there is a spotlighted match.
