libvterm

Tested Software version 0.3 on Linux. libvterm tested is the one shipped with vIM when using command, :terminal. The homepage URL of this terminal is https://www.leonerd.org.uk/code/libvterm/. Full results available at ucs-detect repository path data/libvterm.yaml.

libvterm is a terminal multiplexer. These tests were executed in host terminal :ref:`ghostty. The homepage URL of this host terminal is https://ghostty.org/.

Score Breakdown

Detailed breakdown of how scores are calculated for libvterm:

#	Score Type	Raw Score	Final Scaled Score
1	WIDE	99.98%	100.0%
2	NARROW	99.60%	99.6%
3	ZWJ	0.00%	0.0%
4	LANG	96.22%	96.2%
5	VS16	50.00%	50.0%
6	VS15	0.00%	(excluded)
7	SRI	100.00%	100.0%
8	SFZ	100.00%	100.0%
9	RI	0.00%	0.0%
10	FEAT	43.75%	46.7%
11	Graphics	0%	0.0%
12	Resources	27.0%	27.0%

Score Comparison Plot:

The following plot shows how this terminal’s scores compare to all other terminals tested.

Scaled scores comparison across all metrics (normalized 0-100%)

Final Scaled Score Calculation:

• Raw Final Score: 48.59% (weighted average: WIDE + NARROW + ZWJ + LANG + VS16 + 0.33 * SRI + 0.33 * SFZ + RI + CAP + 0.5 * GFX + 0.5 * RSC) the categorized ‘average’ absolute support level of this terminal.

  Note

  RSC (Resources) is a composite CPU, memory, and runtime score. RSC is weighted at 0.5 (half as powerful as other metrics). FEAT (Features) is the fraction of notable features supported. GFX (Graphics) scores 100% for modern protocols (iTerm2, Kitty), 50% for legacy only (Sixel, ReGIS), 0% for none.

• Final Scaled Score: 24.0% (normalized across all terminals tested). Final Scaled scores are normalized (0-100%) relative to all terminals tested

WIDE Score Details:

Wide character support calculation:

• Total successful codepoints: 43585

• Total codepoints tested: 43592

• Formula: 43585 / 43592

• Result: 99.98%

NARROW Score Details:

Narrow character support calculation:

• Total successful codepoints: 36110

• Total codepoints tested: 36254

• Formula: 36110 / 36254

• Result: 99.60%

ZWJ Score Details:

Emoji ZWJ (Zero-Width Joiner) support calculation:

• Total successful sequences: 0

• Total sequences tested: 1445

• Formula: 0 / 1445

• Result: 0.00%

VS16 Score Details:

Variation Selector-16 support calculation:

• Errors: 213 of 426 codepoints tested

• Success rate: 50.0%

• Formula: 50.0 / 100

• Result: 50.00%

VS15 Score Details (excluded from final score):

Variation Selector-15 support calculation:

• Errors: 158 of 158 codepoints tested

• Success rate: 0.0%

• Formula: 0.0 / 100

• Result: 0.00%

SRI Score Details:

Standalone Regional Indicator support calculation:

• Total successful codepoints: 26

• Total codepoints tested: 26

• Formula: 26 / 26

• Result: 100.00%

SFZ Score Details:

Standalone Fitzpatrick skin tone modifier support calculation:

• Total successful codepoints: 5

• Total codepoints tested: 5

• Formula: 5 / 5

• Result: 100.00%

RI Score Details:

Regional Indicator flag sequence support calculation:

• Total successful sequences: 0

• Total sequences tested: 262

• Formula: 0 / 262

• Result: 0.00%

Features Score Details:

Notable terminal features (7.0 / 16):

• Set bracketed paste mode (2004): yes

• Synchronized Output (2026): no

• Send FocusIn/FocusOut events (1004): yes

• Enable SGR Mouse Mode (1006): yes

• Grapheme Clustering (2027): no

• Bracketed Paste MIME (5522): no

• Kitty Keyboard: yes

• XTGETTCAP: no

• Text Sizing (OSC 66): no

• Kitty Clipboard Protocol: no

• OSC 52 Clipboard: no

• Kitty Pointer Shapes (OSC 22): no

• Kitty Notifications (OSC 99): no

• Color Report (OSC 10/11): yes

• Terminal Identification (XTVERSION): yes

• Truecolor Detection: yes

Raw score: 43.75%

Graphics Score Details:

Graphics protocol support (0%):

• Sixel: no

• ReGIS: no

• iTerm2: no

• Kitty: no

Scoring: 100% for modern (iTerm2/Kitty), 50% for legacy only (Sixel/ReGIS), 0% for none

Resource Score Details:

• Duration: 50.3s

• Mean CPU: 119.1%

• Mean RSS: 279.3 MB

• Resources Score: 27/100

• Note: log-scale composite cost = log(CPU+1) + log(RSS+1) + log(time+1)

• Scaled result: 27.0%

LANG Score Details (Geometric Mean):

Geometric mean calculation:

• Formula: (p₁ × p₂ × … × pₙ)^(1/n) where n = 118 languages

• About geometric mean

• Result: 96.22%

Wide character support

Wide character support of libvterm is 100.0% (7 errors of 43592 codepoints tested).

Sequence of a WIDE character, from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0001FAC8	‘\U0001fac8’	Cn	2	na

Total codepoints: 1

• Shell test using printf(1), '|' should align in output:

  $ printf "\xf0\x9f\xab\x88|\\n12|\\n"
  🫈|
  12|
  

Screenshot:

• python wcwidth.wcswidth() measures width 2, while libvterm measures width 1.

Narrow character support

Narrow character support of libvterm is 99.6% (144 errors of 36254 codepoints tested).

Sequence of a NARROW character, from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0001F3CD	‘\U0001f3cd’	So	1	RACING MOTORCYCLE

Total codepoints: 1

• Shell test using printf(1), '|' should align in output:

  $ printf "\xf0\x9f\x8f\x8d|\\n1|\\n"
  🏍|
  1|
  

• python wcwidth.wcswidth() measures width 1, while libvterm measures width 2.

Emoji ZWJ support

Compatibility of libvterm with the Unicode Emoji ZWJ sequence table is 0.0% (1445 errors of 1445 sequences tested).

Sequence of an Emoji ZWJ Sequence, from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0001F3C3	‘\U0001f3c3’	So	2	RUNNER
2	U+0001F3FF	‘\U0001f3ff’	Sk	2	EMOJI MODIFIER FITZPATRICK TYPE-6
3	U+200D	‘\u200d’	Cf	0	ZERO WIDTH JOINER
4	U+2640	‘\u2640’	So	1	FEMALE SIGN
5	U+FE0F	‘\ufe0f’	Mn	0	VARIATION SELECTOR-16

Total codepoints: 5

• Shell test using printf(1), '|' should align in output:

  $ printf "\xf0\x9f\x8f\x83\xf0\x9f\x8f\xbf\xe2\x80\x8d\xe2\x99\x80\xef\xb8\x8f|\\n12|\\n"
  🏃🏿‍♀️|
  12|
  

Screenshot:

• python wcwidth.wcswidth() measures width 2, while libvterm measures width 11.

Variation Selector-16 support

Emoji VS-16 results for libvterm is 213 errors out of 426 total codepoints tested, 50.0% success. Sequence of a NARROW Emoji made WIDE by Variation Selector-16, from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+26D1	‘\u26d1’	So	1	HELMET WITH WHITE CROSS

Total codepoints: 1

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe2\x9b\x91|\\n1|\\n"
  ⛑|
  1|
  

Screenshot:

• python wcwidth.wcswidth() measures width 1, while libvterm measures width 2.

Variation Selector-15 support

Emoji VS-15 results for libvterm is 158 errors out of 158 total codepoints tested, 0.0% success. Sequence of a WIDE Emoji made NARROW by Variation Selector-15, from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0001F3AE	‘\U0001f3ae’	So	2	VIDEO GAME
2	U+FE0E	‘\ufe0e’	Mn	0	VARIATION SELECTOR-15

Total codepoints: 2

• Shell test using printf(1), '|' should align in output:

  $ printf "\xf0\x9f\x8e\xae\xef\xb8\x8e|\\n1|\\n"
  🎮︎|
  1|
  

Screenshot:

• python wcwidth.wcswidth() measures width 1, while libvterm measures width 2.

Standalone Regional Indicator support

Standalone Regional Indicator support of libvterm is 100.0% (0 errors of 26 codepoints tested).

Standalone Fitzpatrick modifier support

Standalone Fitzpatrick skin tone modifier support of libvterm is 100.0% (0 errors of 5 codepoints tested).

Regional Indicator flag sequence support

Regional Indicator flag sequence support of libvterm is 0.0% (262 errors of 262 sequences tested).

Sequence of a Regional Indicator flag sequence, from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0001F1F0	‘\U0001f1f0’	So	2	REGIONAL INDICATOR SYMBOL LETTER K
2	U+0001F1FF	‘\U0001f1ff’	So	2	REGIONAL INDICATOR SYMBOL LETTER Z

Total codepoints: 2

• Shell test using printf(1), '|' should align in output:

  $ printf "\xf0\x9f\x87\xb0\xf0\x9f\x87\xbf|\\n12|\\n"
  🇰🇿|
  12|
  

Screenshot:

• python wcwidth.wcswidth() measures width 2, while libvterm measures width 4.

Graphics Protocol Support

libvterm does not report support for any graphics protocols.

Detection Methods:

• Sixel and ReGIS: Detected via the Device Attributes (DA1) query CSI c (\x1b[c). Extension code 4 indicates Sixel support, 3 ReGIS.

• Kitty graphics: Detected by sending a Kitty graphics query and checking for an OK response.

• iTerm2 inline images: Detected via the iTerm2 capabilities query OSC 1337 ; Capabilities.

Device Attributes Response:

• Extensions reported: 2

• Sixel indicator (4): not present

• ReGIS indicator (3): not present

Language Support

The following 94 languages were tested with 100% success:

(Jinan), (Yeonbyeon), Aja, Amarakaeri, Arabic, Standard, Assyrian Neo-Aramaic, Baatonum, Bamun, Belanda Viri, Bora, Catalan (2), Chickasaw, Chinantec, Chiltepec, Chinese, Gan, Chinese, Hakka, Chinese, Jinyu, Chinese, Mandarin (Beijing), Chinese, Mandarin (Guiyang), Chinese, Mandarin (Harbin), Chinese, Mandarin (Nanjing), Chinese, Mandarin (Simplified), Chinese, Mandarin (Tianjin), Chinese, Mandarin (Traditional), Chinese, Min Nan, Chinese, Wu, Chinese, Xiang, Chinese, Yue, Colorado, Dagaare, Southern, Dangme, Dendi, Dinka, Northeastern, Ditammari, Dzongkha, Evenki, Fon, French (Welche), Fur, Ga, Gen, Gilyak, Gumuz, Japanese, Japanese (Osaka), Japanese (Tokyo), Kabyle, Korean, Lamnso’, Lao, Lingala (tones), Maldivian, Maori (2), Mazahua Central, Mirandese, Mixtec, Metlatónoc, Mongolian, Halh (Mongolian), Mòoré, Nanai, Navajo, Nuosu, Orok, Otomi, Mezquital, Panjabi, Eastern, Pashto, Northern, Picard, Pular (Adlam), Saint Lucian Creole French, Secoya, Seraiki, Shan, Shipibo-Conibo, Siona, South Azerbaijani, Tagalog (Tagalog), Tai Dam, Tamazight, Central Atlas, Tamil, Tamil (Sri Lanka), Tem, Thai, Thai (2), Tibetan, Central, Ticuna, Uduk, Urdu, Urdu (2), Veps, Vietnamese, Vietnamese (Han nom), Waama, Yaneshaʼ, Yiddish, Eastern, Yoruba, Éwé.

The following 24 languages are not fully supported:

lang	n_errors	n_total	pct_success
Javanese (Javanese)	321	530	39.4%
Malayalam	271	845	67.9%
Bengali	117	385	69.6%
Sanskrit	145	493	70.6%
Marathi	88	391	77.5%
Hindi	82	390	79.0%
Sanskrit (Grantha)	61	293	79.2%
Khmer, Central	90	443	79.7%
Maithili	71	357	80.1%
Nepali	70	352	80.1%
Tamang, Eastern	11	70	84.3%
Gujarati	50	343	85.4%
Magahi	43	314	86.3%
Bhojpuri	41	313	86.9%
Telugu	42	384	89.1%
Farsi, Western	4	49	91.8%
Dari	4	54	92.6%
Kannada	15	287	94.8%
Chakma	13	267	95.1%
Sinhala	11	258	95.7%
Burmese	10	268	96.3%
Mon	7	332	97.9%
Panjabi, Western	1	62	98.4%
Khün	4	396	99.0%

Javanese (Javanese)

Sequence of language Javanese (Javanese) from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+A98F	‘\ua98f’	Lo	1	JAVANESE LETTER KA
2	U+A9BA	‘\ua9ba’	Mc	0	JAVANESE VOWEL SIGN TALING
3	U+A9B4	‘\ua9b4’	Mc	0	JAVANESE VOWEL SIGN TARUNG

Total codepoints: 3

• Shell test using printf(1), '|' should align in output:

  $ printf "\xea\xa6\x8f\xea\xa6\xba\xea\xa6\xb4|\\n12|\\n"
  ꦏꦺꦴ|
  12|
  

Screenshot:

• python wcwidth.wcswidth() measures width 2, while libvterm measures width 3.

Malayalam

Sequence of language Malayalam from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0D15	‘\u0d15’	Lo	1	MALAYALAM LETTER KA
2	U+0D4D	‘\u0d4d’	Mn	0	MALAYALAM SIGN VIRAMA
3	U+0D15	‘\u0d15’	Lo	1	MALAYALAM LETTER KA
4	U+0D3E	‘\u0d3e’	Mc	0	MALAYALAM VOWEL SIGN AA

Total codepoints: 4

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe0\xb4\x95\xe0\xb5\x8d\xe0\xb4\x95\xe0\xb4\xbe|\\n12|\\n"
  ക്കാ|
  12|
  

Screenshot:

• python wcwidth.wcswidth() measures width 2, while libvterm measures width 3.

Bengali

Sequence of language Bengali from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0995	‘\u0995’	Lo	1	BENGALI LETTER KA
2	U+09BE	‘\u09be’	Mc	0	BENGALI VOWEL SIGN AA
3	U+200C	‘\u200c’	Cf	0	ZERO WIDTH NON-JOINER

Total codepoints: 3

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe0\xa6\x95\xe0\xa6\xbe\xe2\x80\x8c|\\n12|\\n"
  কা‌|
  12|
  

Screenshot:

• python wcwidth.wcswidth() measures width 2, while libvterm measures width 8.

Sanskrit

Sequence of language Sanskrit from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0915	‘\u0915’	Lo	1	DEVANAGARI LETTER KA
2	U+094D	‘\u094d’	Mn	0	DEVANAGARI SIGN VIRAMA
3	U+0924	‘\u0924’	Lo	1	DEVANAGARI LETTER TA
4	U+093F	‘\u093f’	Mc	0	DEVANAGARI VOWEL SIGN I

Total codepoints: 4

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe0\xa4\x95\xe0\xa5\x8d\xe0\xa4\xa4\xe0\xa4\xbf|\\n12|\\n"
  क्ति|
  12|
  

Screenshot:

Marathi

Sequence of language Marathi from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0915	‘\u0915’	Lo	1	DEVANAGARI LETTER KA
2	U+094D	‘\u094d’	Mn	0	DEVANAGARI SIGN VIRAMA
3	U+0924	‘\u0924’	Lo	1	DEVANAGARI LETTER TA
4	U+093F	‘\u093f’	Mc	0	DEVANAGARI VOWEL SIGN I

Total codepoints: 4

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe0\xa4\x95\xe0\xa5\x8d\xe0\xa4\xa4\xe0\xa4\xbf|\\n12|\\n"
  क्ति|
  12|
  

Screenshot:

Hindi

Sequence of language Hindi from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0915	‘\u0915’	Lo	1	DEVANAGARI LETTER KA
2	U+094D	‘\u094d’	Mn	0	DEVANAGARI SIGN VIRAMA
3	U+0924	‘\u0924’	Lo	1	DEVANAGARI LETTER TA
4	U+093F	‘\u093f’	Mc	0	DEVANAGARI VOWEL SIGN I

Total codepoints: 4

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe0\xa4\x95\xe0\xa5\x8d\xe0\xa4\xa4\xe0\xa4\xbf|\\n12|\\n"
  क्ति|
  12|
  

Screenshot:

Sanskrit (Grantha)

Sequence of language Sanskrit (Grantha) from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+00011315	‘\U00011315’	Lo	1	GRANTHA LETTER KA
2	U+0001133E	‘\U0001133e’	Mc	0	GRANTHA VOWEL SIGN AA
3	U+00011302	‘\U00011302’	Mc	0	GRANTHA SIGN ANUSVARA

Total codepoints: 3

• Shell test using printf(1), '|' should align in output:

  $ printf "\xf0\x91\x8c\x95\xf0\x91\x8c\xbe\xf0\x91\x8c\x82|\\n12|\\n"
  𑌕𑌾𑌂|
  12|
  

Screenshot:

• python wcwidth.wcswidth() measures width 2, while libvterm measures width 3.

Khmer, Central

Sequence of language Khmer, Central from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+1780	‘\u1780’	Lo	1	KHMER LETTER KA
2	U+17D2	‘\u17d2’	Mn	0	KHMER SIGN COENG
3	U+178A	‘\u178a’	Lo	1	KHMER LETTER DA
4	U+17C5	‘\u17c5’	Mc	0	KHMER VOWEL SIGN AU

Total codepoints: 4

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe1\x9e\x80\xe1\x9f\x92\xe1\x9e\x8a\xe1\x9f\x85|\\n12|\\n"
  ក្ដៅ|
  12|
  

Screenshot:

• python wcwidth.wcswidth() measures width 2, while libvterm measures width 3.

Maithili

Sequence of language Maithili from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0915	‘\u0915’	Lo	1	DEVANAGARI LETTER KA
2	U+094D	‘\u094d’	Mn	0	DEVANAGARI SIGN VIRAMA
3	U+0924	‘\u0924’	Lo	1	DEVANAGARI LETTER TA
4	U+093F	‘\u093f’	Mc	0	DEVANAGARI VOWEL SIGN I

Total codepoints: 4

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe0\xa4\x95\xe0\xa5\x8d\xe0\xa4\xa4\xe0\xa4\xbf|\\n12|\\n"
  क्ति|
  12|
  

Screenshot:

Nepali

Sequence of language Nepali from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0915	‘\u0915’	Lo	1	DEVANAGARI LETTER KA
2	U+094D	‘\u094d’	Mn	0	DEVANAGARI SIGN VIRAMA
3	U+0924	‘\u0924’	Lo	1	DEVANAGARI LETTER TA
4	U+093F	‘\u093f’	Mc	0	DEVANAGARI VOWEL SIGN I

Total codepoints: 4

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe0\xa4\x95\xe0\xa5\x8d\xe0\xa4\xa4\xe0\xa4\xbf|\\n12|\\n"
  क्ति|
  12|
  

Screenshot:

Tamang, Eastern

Sequence of language Tamang, Eastern from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0915	‘\u0915’	Lo	1	DEVANAGARI LETTER KA
2	U+094D	‘\u094d’	Mn	0	DEVANAGARI SIGN VIRAMA
3	U+0924	‘\u0924’	Lo	1	DEVANAGARI LETTER TA
4	U+093F	‘\u093f’	Mc	0	DEVANAGARI VOWEL SIGN I

Total codepoints: 4

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe0\xa4\x95\xe0\xa5\x8d\xe0\xa4\xa4\xe0\xa4\xbf|\\n12|\\n"
  क्ति|
  12|
  

Screenshot:

Gujarati

Sequence of language Gujarati from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0A95	‘\u0a95’	Lo	1	GUJARATI LETTER KA
2	U+0ACD	‘\u0acd’	Mn	0	GUJARATI SIGN VIRAMA
3	U+0A95	‘\u0a95’	Lo	1	GUJARATI LETTER KA
4	U+0ABE	‘\u0abe’	Mc	0	GUJARATI VOWEL SIGN AA

Total codepoints: 4

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe0\xaa\x95\xe0\xab\x8d\xe0\xaa\x95\xe0\xaa\xbe|\\n12|\\n"
  ક્કા|
  12|
  

Screenshot:

• python wcwidth.wcswidth() measures width 2, while libvterm measures width 3.

Magahi

Sequence of language Magahi from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0915	‘\u0915’	Lo	1	DEVANAGARI LETTER KA
2	U+094D	‘\u094d’	Mn	0	DEVANAGARI SIGN VIRAMA
3	U+0924	‘\u0924’	Lo	1	DEVANAGARI LETTER TA
4	U+093F	‘\u093f’	Mc	0	DEVANAGARI VOWEL SIGN I

Total codepoints: 4

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe0\xa4\x95\xe0\xa5\x8d\xe0\xa4\xa4\xe0\xa4\xbf|\\n12|\\n"
  क्ति|
  12|
  

Screenshot:

Bhojpuri

Sequence of language Bhojpuri from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0915	‘\u0915’	Lo	1	DEVANAGARI LETTER KA
2	U+094D	‘\u094d’	Mn	0	DEVANAGARI SIGN VIRAMA
3	U+0918	‘\u0918’	Lo	1	DEVANAGARI LETTER GHA
4	U+094D	‘\u094d’	Mn	0	DEVANAGARI SIGN VIRAMA
5	U+0918	‘\u0918’	Lo	1	DEVANAGARI LETTER GHA
6	U+093F	‘\u093f’	Mc	0	DEVANAGARI VOWEL SIGN I
7	U+094D	‘\u094d’	Mn	0	DEVANAGARI SIGN VIRAMA

Total codepoints: 7

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe0\xa4\x95\xe0\xa5\x8d\xe0\xa4\x98\xe0\xa5\x8d\xe0\xa4\x98\xe0\xa4\xbf\xe0\xa5\x8d|\\n12|\\n"
  क्घ्घि्|
  12|
  

Screenshot:

• python wcwidth.wcswidth() measures width 2, while libvterm measures width 4.

Telugu

Sequence of language Telugu from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0C15	‘\u0c15’	Lo	1	TELUGU LETTER KA
2	U+0C41	‘\u0c41’	Mc	0	TELUGU VOWEL SIGN U
3	U+0C02	‘\u0c02’	Mc	0	TELUGU SIGN ANUSVARA

Total codepoints: 3

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe0\xb0\x95\xe0\xb1\x81\xe0\xb0\x82|\\n12|\\n"
  కుం|
  12|
  

Screenshot:

• python wcwidth.wcswidth() measures width 2, while libvterm measures width 3.

Farsi, Western

Sequence of language Farsi, Western from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+062A	‘\u062a’	Lo	1	ARABIC LETTER TEH
2	U+200C	‘\u200c’	Cf	0	ZERO WIDTH NON-JOINER

Total codepoints: 2

• Shell test using printf(1), '|' should align in output:

  $ printf "\xd8\xaa\xe2\x80\x8c|\\n1|\\n"
  ت‌|
  1|
  

Screenshot:

Dari

Sequence of language Dari from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+062A	‘\u062a’	Lo	1	ARABIC LETTER TEH
2	U+200C	‘\u200c’	Cf	0	ZERO WIDTH NON-JOINER

Total codepoints: 2

• Shell test using printf(1), '|' should align in output:

  $ printf "\xd8\xaa\xe2\x80\x8c|\\n1|\\n"
  ت‌|
  1|
  

Screenshot:

• python wcwidth.wcswidth() measures width 1, while libvterm measures width 7.

Kannada

Sequence of language Kannada from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0C95	‘\u0c95’	Lo	1	KANNADA LETTER KA
2	U+0CBE	‘\u0cbe’	Mc	0	KANNADA VOWEL SIGN AA
3	U+0C82	‘\u0c82’	Mc	0	KANNADA SIGN ANUSVARA

Total codepoints: 3

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe0\xb2\x95\xe0\xb2\xbe\xe0\xb2\x82|\\n12|\\n"
  ಕಾಂ|
  12|
  

Screenshot:

• python wcwidth.wcswidth() measures width 2, while libvterm measures width 3.

Chakma

Sequence of language Chakma from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+00011107	‘\U00011107’	Lo	1	CHAKMA LETTER KAA
2	U+00011133	‘\U00011133’	Mn	0	CHAKMA VIRAMA
3	U+00011120	‘\U00011120’	Lo	1	CHAKMA LETTER YYAA
4	U+0001112C	‘\U0001112c’	Mc	0	CHAKMA VOWEL SIGN E

Total codepoints: 4

• Shell test using printf(1), '|' should align in output:

  $ printf "\xf0\x91\x84\x87\xf0\x91\x84\xb3\xf0\x91\x84\xa0\xf0\x91\x84\xac|\\n12|\\n"
  𑄇𑄳𑄠𑄬|
  12|
  

Screenshot:

• python wcwidth.wcswidth() measures width 2, while libvterm measures width 3.

Sinhala

Sequence of language Sinhala from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+0D9A	‘\u0d9a’	Lo	1	SINHALA LETTER ALPAPRAANA KAYANNA
2	U+0DCA	‘\u0dca’	Mn	0	SINHALA SIGN AL-LAKUNA
3	U+200D	‘\u200d’	Cf	0	ZERO WIDTH JOINER

Total codepoints: 3

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe0\xb6\x9a\xe0\xb7\x8a\xe2\x80\x8d|\\n1|\\n"
  ක්‍|
  1|
  

Screenshot:

• python wcwidth.wcswidth() measures width 1, while libvterm measures width 7.

Burmese

Sequence of language Burmese from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+1000	‘\u1000’	Lo	1	MYANMAR LETTER KA
2	U+103B	‘\u103b’	Mc	0	MYANMAR CONSONANT SIGN MEDIAL YA
3	U+1031	‘\u1031’	Mc	0	MYANMAR VOWEL SIGN E

Total codepoints: 3

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe1\x80\x80\xe1\x80\xbb\xe1\x80\xb1|\\n12|\\n"
  ကျေ|
  12|
  

Screenshot:

• python wcwidth.wcswidth() measures width 2, while libvterm measures width 3.

Mon

Sequence of language Mon from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+1007	‘\u1007’	Lo	1	MYANMAR LETTER JA
2	U+103C	‘\u103c’	Mc	0	MYANMAR CONSONANT SIGN MEDIAL RA
3	U+1031	‘\u1031’	Mc	0	MYANMAR VOWEL SIGN E

Total codepoints: 3

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe1\x80\x87\xe1\x80\xbc\xe1\x80\xb1|\\n12|\\n"
  ဇြေ|
  12|
  

Screenshot:

• python wcwidth.wcswidth() measures width 2, while libvterm measures width 3.

Panjabi, Western

Sequence of language Panjabi, Western from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+06D2	‘\u06d2’	Lo	1	ARABIC LETTER YEH BARREE
2	U+200C	‘\u200c’	Cf	0	ZERO WIDTH NON-JOINER

Total codepoints: 2

• Shell test using printf(1), '|' should align in output:

  $ printf "\xdb\x92\xe2\x80\x8c|\\n1|\\n"
  ے‌|
  1|
  

Screenshot:

• python wcwidth.wcswidth() measures width 1, while libvterm measures width 7.

Khün

Sequence of language Khün from midpoint of alignment failure records:

#	Codepoint	Python	Category	wcwidth	Name
1	U+1A23	‘\u1a23’	Lo	1	TAI THAM LETTER LOW KA
2	U+1A55	‘\u1a55’	Mc	0	TAI THAM CONSONANT SIGN MEDIAL RA
3	U+1A6E	‘\u1a6e’	Mc	0	TAI THAM VOWEL SIGN E
4	U+1A60	‘\u1a60’	Mn	0	TAI THAM SIGN SAKOT

Total codepoints: 4

• Shell test using printf(1), '|' should align in output:

  $ printf "\xe1\xa8\xa3\xe1\xa9\x95\xe1\xa9\xae\xe1\xa9\xa0|\\n12|\\n"
  ᨣᩕᩮ᩠|
  12|
  

Screenshot:

• python wcwidth.wcswidth() measures width 2, while libvterm measures width 3.

DEC Private Modes Support

DEC private modes results for libvterm: 3 changeable modes of 3 supported out of 8 total modes tested (37.5% support, 37.5% changeable).

Complete list of DEC private modes tested:

Mode	Name	Description	Supported	Changeable	Enabled
1004	FOCUS_IN_OUT_EVENTS	Send FocusIn/FocusOut events	Yes	Yes	No
1006	MOUSE_EXTENDED_SGR	Enable SGR Mouse Mode	Yes	Yes	No
2004	BRACKETED_PASTE	Set bracketed paste mode	Yes	Yes	No
2026	SYNCHRONIZED_OUTPUT	Synchronized Output	No	No	No
2027	GRAPHEME_CLUSTERING	Grapheme Clustering	No	No	No
2031	COLOR_PALETTE_UPDATES	Color palette updates	No	No	No
2048	IN_BAND_WINDOW_RESIZE	In-Band Window Resize Notifications	No	No	No
5522	BRACKETED_PASTE_MIME	Bracketed Paste MIME	No	No	No

Summary: 3 changeable, 5 not changeable.

Kitty Keyboard Protocol

libvterm supports the Kitty keyboard protocol.

#	Flag	Key	State
1	Disambiguate escape codes	disambiguate	No
2	Report event types	report_events	No
3	Report alternate keys	report_alternates	No
4	Report all keys as escape codes	report_all_keys	No
5	Report associated text	report_text	No

Detection is performed by sending CSI ? u to query the current progressive enhancement flags. A terminal that supports this protocol responds with the active flags value.

XTGETTCAP (Terminfo Capabilities)

libvterm does not support the XTGETTCAP sequence.

Text Sizing Protocol (OSC 66)

libvterm does not support the Text Sizing protocol.

Truecolor Support

libvterm supports 24-bit truecolor, detectable via:

• XTGETTCAP (RGB capability): no

• DECRQSS (truecolor probe): yes (probe)

• COLORTERM: yes (truecolor)

OSC 52 Clipboard Support

libvterm does not advertise OSC 52 clipboard support via DA1 extension 52 or XTGETTCAP Ms.

• DA1 extension 52: no

• XTGETTCAP Ms: no

Terminal Identification

libvterm is identified as libvterm version 0.3 (detected via XTVERSION).

• XTVERSION (raw): libvterm(0.3)

• XTVERSION: yes

• XTGETTCAP TN: no

• ENQ: no

• TERM_PROGRAM: no

• TERM: yes (xterm-ghostty) (inherited from host terminal)

Reproduction

To reproduce these results for libvterm, install and run ucs-detect with the following commands:

uvx ucs-detect --rerun data/libvterm.yaml

Test Performance

The test suite completed in 50.33 seconds (50s).

• Mean CPU: 119.1%

• Mean RSS: 279.3 MB

• Total time: 50.3s

CPU usage during test execution for libvterm.

RSS memory usage during test execution for libvterm.

Test duration for libvterm compared to all other terminals.

CPU % vs duration trade-off for libvterm.