| … | |
… | |
| 37 | TODO: c&c perl socket via daemon-ext mechanism |
37 | TODO: c&c perl socket via daemon-ext mechanism |
| 38 | TODO: simplify extension metainfo cache, cache on disk |
38 | TODO: simplify extension metainfo cache, cache on disk |
| 39 | TODO: URxvt::Ext::Name installs urxvt ext name and provides pod/manpage for URxvt::Ext::Name |
39 | TODO: URxvt::Ext::Name installs urxvt ext name and provides pod/manpage for URxvt::Ext::Name |
| 40 | TODO: üpixel droppins idenrasm,ll,scrollup |
40 | TODO: üpixel droppins idenrasm,ll,scrollup |
| 41 | TODO: cuu/cud and probably others default to 1 when arg is 0, not just missing, in xterm/vt102, but not in rxvt |
41 | TODO: cuu/cud and probably others default to 1 when arg is 0, not just missing, in xterm/vt102, but not in rxvt |
|
|
42 | TODO: clean up the resource vs. long option mess, especially in the manpage. |
| 42 | |
43 | |
| 43 | TODO: xterm-osc does not autoload background |
|
|
| 44 | TODO: think about daemon delaying urxvt client exit. |
44 | TODO: think about daemon delaying urxvt client exit. |
| 45 | |
45 | IDEA: use harfbuzz to render combining sequences, maybe also complex scripts |
| 46 | TODO: https://github.com/bbidulock/icewm/issues/492 urxvt wrongfully(?) needa configurenotify after mapwindow? also, black border? |
46 | TODO: https://github.com/bbidulock/icewm/issues/492 urxvt wrongfully(?) needa configurenotify after mapwindow? also, black border? |
| 47 | TODO: compose chars should be rendered "in one go" for xft. maybe. possibly. |
47 | TODO: match default regex does not match https://猫.移动/tmp/, but it does match if you special_decode the lines first. special_encode of the regex does not and would probably be the better solution -> investigate. |
| 48 | TODO: disbale bracketed paste mode option, too amny prorgams abuse it |
48 | |
|
|
49 | TODO: rclock date overwrite |
|
|
50 | - libptytty is now an external dependency. |
| 49 | - enable confirm-paste extzensions by default, can be disabled by e.g. |
51 | - enable confirm-paste extension by default, can be disabled by e.g. |
| 50 | URxvt.perl-ext: -confirm-paste |
52 | URxvt.perl-ext: -confirm-paste |
|
|
53 | - Key bindings specified with the -keysym option or with resources of |
|
|
54 | the form URxvt*keysym work again. Such resources, as opposed to |
|
|
55 | resources of the form URxvt.keysym, were not officially allowed but |
|
|
56 | accepted, so ignoring them is considered a regression. |
|
|
57 | - new option/resource -rm/rewrapMode, which sets the line rewrap |
|
|
58 | mode to auto, never or always. |
|
|
59 | - new option/resource -dpb/disablePasteBrackets, which can be used to |
|
|
60 | disable bracketed paste mode sequences, as more and more programs |
|
|
61 | abuse these sequences (typically in the name of security, but these |
|
|
62 | sequences are not improving security). Also, bracketed paste mode |
|
|
63 | is now under frills. |
|
|
64 | - add manpage for urclock. |
|
|
65 | - confirm-paste wrongly considered brackets as ctrlchars (patch by |
|
|
66 | Jakub Wilk, debian bug #995244). |
| 51 | - if urxvt cannot detect a good estimate for font width (because a font does |
67 | - if urxvt cannot detect a good estimate for font width (because a font does |
| 52 | not contain laitn characters at all for example), it will now fall |
68 | not contain latin characters at all for example), it will now fall |
| 53 | back to xft's max_advance_width, which works well with soem fonts. It |
69 | back to xft's max_advance_width, which works well with some fonts. It |
| 54 | currently will warn about this. |
70 | currently will warn about this. |
| 55 | - do not render variation selectors in the default font as boxes, treat |
71 | - do not render variation selectors in the default font as boxes, treat |
| 56 | them as zero-width characters. |
72 | them as zero-width characters. |
| 57 | - unicode characters outside of the BMP can now be rendered even |
73 | - change combining character architecture to allow fonts to handle |
| 58 | without --enable-unicode3, when using xft fonts. |
74 | combining sequences on their own. This could allow future renderers |
| 59 | - urxvt now passes full combining character sequences to xft, |
75 | to be more intelligent about these sequences. |
| 60 | increasing the chance of it doing something useful with it, such |
76 | - use the previous enhancement to allow unicode characters outside of the |
| 61 | as correctly applying variant selectors. |
77 | BMP to be rendered even without --enable-unicode3, as long as an |
|
|
78 | xft with the character is configured. |
|
|
79 | It's recommended to stay with disabled unicode3 even more now. |
|
|
80 | - freetype/xft have been verified to return nonsensical offsets for |
|
|
81 | combining characters. urxvt tries to compensate to some extent, |
|
|
82 | but diacritics will be slightly off in many cases, and far off in |
|
|
83 | some cases (e.g. input mono). |
|
|
84 | - when checking for overlap/careful handling, take non-basefont |
|
|
85 | character shifts into account. |
|
|
86 | - extensions can now request autoloading on OSC and Perl OSC sequences, |
|
|
87 | implement this for the background, clipboard-osc and overlay-osc |
|
|
88 | extensions. |
|
|
89 | - the background extension will now be autoloaded when the OSC 20 or 705 |
|
|
90 | sequence is received. |
|
|
91 | - The view_start perl method now scrolls to the bottom also when the |
|
|
92 | argument is greater than 0. |
|
|
93 | - eval scroll actions now scroll to the bottom when scrolling down and |
|
|
94 | there are less lines to scroll than the ones requested (reported by |
|
|
95 | Stephen Talley). |
| 62 | - allow perl-ext and perl-ext-common to override autoloaded extensions. |
96 | - allow perl-ext and perl-ext-common to override autoloaded extensions. |
| 63 | this allows tabbed to disable itself in subtabs, avoiding recursive |
97 | this allows tabbed to disable itself in subtabs, avoiding recursive |
| 64 | loading. |
98 | loading. |
| 65 | - update precompose table to unicode 13.0. |
99 | - update precompose table to unicode 13.0. |
| 66 | - update charset conversion tables to perl 5.32 (this only affected |
100 | - update charset conversion tables to perl 5.32 (this only affected |
| 67 | cs_koi8_u_from_unicode). |
101 | cs_koi8_u_from_unicode). |
| 68 | - unicode characters > 0x100000 were not handled correctly w.r.t. |
102 | - unicode characters > 0x100000 were not handled correctly w.r.t. |
| 69 | font rendering. |
103 | font rendering. |
|
|
104 | - replace bitstream vera by dejavu sans, as it has replaced the former |
|
|
105 | almost everywhere. |
|
|
106 | - urxvt will now try tic -x first and only fall back to tic without -x |
|
|
107 | when tic -x fails. |
|
|
108 | - shave off over 1K of code size from urclock by replacing the sine table |
|
|
109 | with a slightly less accurate fixed point approximation. |
|
|
110 | - do not specify indentation with =over in pod, this is no longer |
|
|
111 | needed with semi-modern pod tools and gives the formatter more |
|
|
112 | stylistic freedom. |
|
|
113 | - background extension did not set the "size-sensitive" flag |
|
|
114 | when repeat mode wasn't normal. |
|
|
115 | - extensions are now subjected to "use strict 'subs'". |
|
|
116 | - avoid crashing when xft successfully opens a font but then somehow |
|
|
117 | can't lock the face. |
| 70 | |
118 | |
| 71 | 9.26 Fri May 14 19:14:14 CEST 2021 |
119 | 9.26 Fri May 14 19:14:14 CEST 2021 |
| 72 | - ev_iouring.c was wrongly required during compilation, and wrongly |
120 | - ev_iouring.c was wrongly required during compilation, and wrongly |
| 73 | not packaged. |
121 | not packaged. |
| 74 | |
122 | |
