Update requested for country.sys on Freedos.devel

[andrewbird]

Just pasting this here so it doesn't get lost.

Sorry for disturbing you, it is not very important for 2044, but maybe you will have a look at country.sys.
 
I just made a quick look and several "new" countries are not listed there:
 
a) brazilian / brasilianisch  (pb)  Sim/Não           / Sair      S N Ss  858  pb  ?     BR  (strictly no language code!) - not sure if there will be changes 
b) esperanto / esperanto      (eo)  Jes/Ne            / Fini      J N Ff  858  eo  ok    NONE! (NO country exists!) - not sure if there will be changes
c) estonian / estnisch        (et)  Jah/Ei            / Välju     J E Vv  775  et  ?     EE   (ET=NEW, EE old + country differs)  (NOT IN COUNTRY.SYS)
d) icelandic / isländisch     (is)  Já/Nei            / Hætta     J N Hh  861  is  ok    IS   (NOT IN COUNTRY.SYS, a state of its own)
e) latvian / lettisch         (lv)  Jā/Nē             / Iziet     J N Ii  775  lv  ok    LV   (NOT IN COUNTRY.SYS, former part of RU)
f) lithuanian / litauisch     (lt)  Taip/Ne           / Išeiti    T N Ii  775  lt  ok    LT   (NOT IN COUNTRY.SYS, former part of RU)
g) slovak / slowakisch        (sk)  Áno/Nie           / Ukončiť   Á N Uu  852  sk  ok    SK   (NOT IN COUNTRY.SYS, former part of Czechia)
h) slovenian / slowenisch     (sl)  Da/Ne             / Izhod     D N Ii  852  sl  ok    SL   (SI - country differs)
i) swedish / schwedisch       (sv)  Ja/Nej            / Avsluta   J N Aa  858  sv  ok    SE   (SE country differs)
j) ukrainian / ukrainisch     (uk)  Так (Tak)/Ні (Ni) / Вихід     Т Н Вв  855  uk  ok    UA   (UA country differs)
 
"country differs" Maybe this has to do with differing between "Language abbreviation" and "Country abbreviation)
 
It would be great if you could check this before publishing 2045 country.sys.
 
Thank you very much!
 
Fritz

[PerditionC]

Quick list of possible issues to resolve:

Esperanto what country code to use? if none available then can't include it.
Slovenian country code SI or SL? SI seems proper one based on ISO standards, but is SL more correct?
Ukrainian is CP855 correct, quick search suggests CP866?
The rest seem fine to update/add.

[andrewbird]

I don't really have much experience with non English DOS, but I would think the country abbreviations should match those in the general translation project https://github.com/shidel/fd-nls .

[andrewbird]

Now back at the computer, I was hoping to write a fuller reply about seeing only sl for Slovenian over at https://shidel.github.io/fd-nls/report.html and that si only seems to be part of internal labels in country itself https://github.com/FDOS/country/bob/13587a32d6fda6cf1a1ce9bd9c3da7d3ae5a8dd5/country.asm#L297-L302 , however when reading the definitions for Ukraine I wonder if these country codes are not expected to be the same as those in the LANG variable. In particular I see that there's a country code for United Kingdom of uk when we'd use en_GB for LANG. I realise that my DOS knowledge is lacking here, it's probably better to ask the requester Fritz on Freedos devel list.

[fritzmueller]

Just to clarify a little: Do you remember the discussion with Brazil and "br" and "pt" "ptb" and "pb" now?

There are two different things: The LANGUAGE ending code used for the FreeDOS NLS files, see: https://shidel.github.io/fd-nls/report.html and the country ending (sounds crazy, but there are really ISO definitions for both!) And they may differ!
https://www.loc.gov/standards/iso639-2/php/code_list.php ISO 639-1 is the language code (according to Jerome it is overtaken for NLS with the exceptions of dk (denmark - should be da) and cz (czechia - should be cs) and brazil (pb - there was none). Estland also has two different endings!
see also: https://de.wikipedia.org/wiki/Liste_der_ISO-639-2-Codes

The COUNTRY codes CAN but MUST not differ, see: https://de.wikipedia.org/wiki/ISO-3166-1-Kodierliste

So SE(Sveden), UA (Ukraine) and Slovenian (SI) may be correct

in country.sys are mentioned:

se (sveden) (language: sv)
si (slovenian) (language: sl)
ua_848 cnf 380,848,DMY,0A3h,0E0h,0ADh,".",0," ",",",".",":",3,2,_24;Ukraine Oleg
ua_1125 cnf 380,1125,DMY,0A3h,0E0h,0ADh,".",0," ",",",".",":",3,2,_24; Ukraine uk for language (is cp855 wrong? FreeDOS cpidos offers 3 different codpages for Ukraine and 7 for Lithuanian )

In the hope to have confused you completely (haha)

My problem is that I do not know which things are really stored in country.sys, an alternative for country.sys is "localcfg" from mateusz viste, available via fdnpkg, but I think not everything there is inside of country.sys
The best thing to get this out is to ask the mailinglist or the bttr forum https://www.bttr-software.de/forum/board.php

Why am I interested in this? The codepages of NLS will have to be coordinated with country.sys - and it makes no fun to rewrite all NLS files again.

[PerditionC]

I will revisit later today when at a computer, but I think part of the confusion is nls files (language, codepage, and translations) vs country configuration (kernel api and codepages).

nls files need to use ISO language codes where available and specify codepage used (ideally all translations are in utf8 and converted to users active codepage with a preferred one for accurate display, accents, codepage specific characters, etc).

country.sys has no notion of language, but provides information that may be language/locale specific. this is the problem.
Country.sys specifies only country, keyed by numerical value (generally matches international phone prefix) and 2 letter country [not language] code. then a bunch of language/local specific values like yes or no, how date shown, and separaters (. vs ,).

So some language and country codes that users may think would be same aren't and this could cause misconfigurations or incorrect handling by programs not expecting this difference.

For country.sys, ISO country codes need to be used, not ISO language codes.

I think our nls should use directories and not extensions so pt-br and en-ca or even 3 alpha codes if no 2 alpha exists can be used. but this is not directly relevant and for further discussion not this bug.

Anyway just trying to jot down some thoughts while trying to wrap my head around best course of action. I personally want to keep to standards and compatible values, but open to what is best since reality is I'm not the target user of country nor the translations.

[andrewbird]

@fritzmueller

In the hope to have confused you completely (haha)

No kidding!

So looking at the country.asm file, the 'ua_848' seems only to be an internal label, which if true means it can be anything unique within the file. I suspect the lookup is done by reading each entry and comparing the '380' Telephone country code and '848' Codepage fields.

I wrote a quick python validator to check some of the labels (there are others like '__ca_ES_850' which seem to be made up for catalan/galicia etc) and all but UK and YU seem to exist as https://en.wikipedia.org/wiki/ISO_3166-1_alpha-2. I have yet to check if the country code matches the phone number code, but I'm working on it.

@PerditionC

nls files need to use ISO language codes where available and specify codepage used (ideally all translations are in utf8 and converted to users active codepage with a preferred one for accurate display, accents, codepage specific characters, etc).

I love that idea. I wrote a validator for fd-nls that flags when a codepage specific file doesn't agree with the iconv converted utf-8 if it exists, so at least you know straight away there's something awry.

country.sys has no notion of language, but provides information that may be language/locale specific. this is the problem.
Country.sys specifies only country, keyed by numerical value (generally matches international phone prefix) and 2 letter country [not language] code. then a bunch of language/local specific values like yes or no, how date shown, and separaters (. vs ,).

That aligns with what I'm learning by looking at the asm. Some of my supposition above is confirmed by what you write. This post has been gestating for a while.

[andrewbird]

I ran my little validation script on the country.asm file and the master table gave these results (after adding a couple of fixups for French-Canada and Latin-America)

Country ISO3166-1-A2 (yu) invalid at line 122 '__yu_852 dw 12, 38,852,0,0'
Country ISO3166-1-A2 (yu) invalid at line 124 '__yu_855 dw 12, 38,855,0,0'
Country ISO3166-1-A2 (yu) invalid at line 126 '__yu_872 dw 12, 38,872,0,0'
Country ISO3166-1-A2 (yu) invalid at line 128 '__yu_858 dw 12, 38,858,0,0'
Country ISO3166-1-A2 (yu) invalid at line 130 '__yu_850 dw 12, 38,850,0,0'
Country ISO3166-1-A2 (cz) mismatch with International Phone Prefix at line 150 '__cz_852 dw 12, 42,852,0,0'
Country ISO3166-1-A2 (cz) mismatch with International Phone Prefix at line 152 '__cz_850 dw 12, 42,850,0,0'
Country ISO3166-1-A2 (cz) mismatch with International Phone Prefix at line 154 '__cz_858 dw 12, 42,858,0,0'
Country ISO3166-1-A2 (uk) invalid at line 162 '__uk_858 dw 12, 44,858,0,0'
Country ISO3166-1-A2 (uk) invalid at line 164 '__uk_850 dw 12, 44,850,0,0'
Country ISO3166-1-A2 (uk) invalid at line 166 '__uk_437 dw 12, 44,437,0,0'
Country ISO3166-1-A2 (sr) mismatch with International Phone Prefix at line 280 '__sr_855 dw 12, 381,855,0,0 ; Serbia and Montenegro, Serbian, Cyrillic'
Country ISO3166-1-A2 (sr) mismatch with International Phone Prefix at line 282 '__sr_872 dw 12, 381,872,0,0'
Country ISO3166-1-A2 (sr) mismatch with International Phone Prefix at line 284 '__sr_852 dw 12, 381,852,0,0 ; Serbia and Montenegro, Serbian, Latin'
Country ISO3166-1-A2 (sr) mismatch with International Phone Prefix at line 286 '__sr_850 dw 12, 381,850,0,0'
Country ISO3166-1-A2 (sr) mismatch with International Phone Prefix at line 288 '__sr_858 dw 12, 381,858,0,0'
Country ISO3166-1-A2 (hr) mismatch with International Phone Prefix at line 290 '__hr_852 dw 12, 384,852,0,0 ; Croatia, Croatian'
Country ISO3166-1-A2 (hr) mismatch with International Phone Prefix at line 292 '__hr_850 dw 12, 384,850,0,0'
Country ISO3166-1-A2 (hr) mismatch with International Phone Prefix at line 294 '__hr_858 dw 12, 384,858,0,0'
Country ISO3166-1-A2 (me) mismatch with International Phone Prefix at line 320 '__me_858 dw 12,785,858,0,0'
Country ISO3166-1-A2 (me) mismatch with International Phone Prefix at line 322 '__me_850 dw 12,785,850,0,0'
Country ISO3166-1-A2 (me) mismatch with International Phone Prefix at line 324 '__me_864 dw 12,785,864,0,0'
Errors = 22
Number of entries (171)

So the problem (if there is any - they are only internal identifiers) countries are:
YU - Yugoslavia (no longer exists), now several countries
CZ - Czechoslovakia (by phone identifier) no longer exists, now two countries Czech Republic and Slovakia
UK - United Kingdom - but should be 'GB'
SR - Serbia
HR - Croatia
ME - (I looked up the phone prefix 785 which appears to be former Soviet Union)

Also the made up identifiers seem fine '__nl_BE_850' means Dutch Language, Country Belgium, Codepage 850. Its identifier '40032' is made up of '4' means multilingual country, '0' first language within, and '032' original country code.

I'm happy to update the UK -> GB identifiers.

[andrewbird]

Actually found a comment later in the country.asm file that indicates the intention of 'me' was middle east. No sure how that can work with international phone number prefix, unless there's an exception like
'2' == FrenchCanada
'3' == LatinAmerica

[fritzmueller]

Hi, you are great! Just only for safety - and as you already work on it - and it would complete the whole region for the next years. I hope I overlooked nothing.   What about Austria (same settings as germany) +43 Switzerland +41 (speak french, italian, german and some very small slangs) Liechtenstein + 423 (very very small, between Austria, Switzerland and Germany, should speak German only) Luxemburg + 352 Iceland +354 Estonia +372 Latvia + 371 Lithuania +370 Ireland +353 Belarus +375 Greece +30 Albania +355 Romania +40 Bulgaria + 359 Moldova +373 (small, close to Ukrainia and Russia)   this should hopefully cover whole Europe. - they are not all on Jeromes NLS list, but maybe some day...   You are right, Yugoslavia has gone and is splitted into a lot of new countries. Czechoslovakia has become Czechia +420  and Slovakia +421 ME should be Montenegro, see below.   former yugoslavia now has 7 separate states: (according to wiki): Slowenien (slovenia) + 386 Kroatien   (croatia) + 385 Bosnien und Herzegowina Bosnia and Herzegovina +387 Serbien  (serbia) +381 Montenegro Montenegro +382 Nordmazedonien northern Mazedonia, former Mazedonia +389 Kosovo +383   Fritz     Sent: Tuesday, January 27, 2026 at 7:32 PM From: "Andrew Bird" ***@***.***> To: "FDOS/kernel" ***@***.***> Cc: "fritzmueller" ***@***.***>, "Mention" ***@***.***> Subject: Re: [FDOS/kernel] Update requested for country.sys on Freedos.devel (Issue #235) andrewbird left a comment (FDOS/kernel#235) I ran my little validation script on the country.asm file and the master table gave these results (after adding a couple of fixups for French-Canada and Latin-America) Country ISO3166-1-A2 (yu) invalid at line 122 '__yu_852 dw 12, 38,852,0,0' Country ISO3166-1-A2 (yu) invalid at line 124 '__yu_855 dw 12, 38,855,0,0' Country ISO3166-1-A2 (yu) invalid at line 126 '__yu_872 dw 12, 38,872,0,0' Country ISO3166-1-A2 (yu) invalid at line 128 '__yu_858 dw 12, 38,858,0,0' Country ISO3166-1-A2 (yu) invalid at line 130 '__yu_850 dw 12, 38,850,0,0' Country ISO3166-1-A2 (cz) mismatch with International Phone Prefix at line 150 '__cz_852 dw 12, 42,852,0,0' Country ISO3166-1-A2 (cz) mismatch with International Phone Prefix at line 152 '__cz_850 dw 12, 42,850,0,0' Country ISO3166-1-A2 (cz) mismatch with International Phone Prefix at line 154 '__cz_858 dw 12, 42,858,0,0' Country ISO3166-1-A2 (uk) invalid at line 162 '__uk_858 dw 12, 44,858,0,0' Country ISO3166-1-A2 (uk) invalid at line 164 '__uk_850 dw 12, 44,850,0,0' Country ISO3166-1-A2 (uk) invalid at line 166 '__uk_437 dw 12, 44,437,0,0' Country ISO3166-1-A2 (sr) mismatch with International Phone Prefix at line 280 '__sr_855 dw 12, 381,855,0,0 ; Serbia and Montenegro, Serbian, Cyrillic' Country ISO3166-1-A2 (sr) mismatch with International Phone Prefix at line 282 '__sr_872 dw 12, 381,872,0,0' Country ISO3166-1-A2 (sr) mismatch with International Phone Prefix at line 284 '__sr_852 dw 12, 381,852,0,0 ; Serbia and Montenegro, Serbian, Latin' Country ISO3166-1-A2 (sr) mismatch with International Phone Prefix at line 286 '__sr_850 dw 12, 381,850,0,0' Country ISO3166-1-A2 (sr) mismatch with International Phone Prefix at line 288 '__sr_858 dw 12, 381,858,0,0' Country ISO3166-1-A2 (hr) mismatch with International Phone Prefix at line 290 '__hr_852 dw 12, 384,852,0,0 ; Croatia, Croatian' Country ISO3166-1-A2 (hr) mismatch with International Phone Prefix at line 292 '__hr_850 dw 12, 384,850,0,0' Country ISO3166-1-A2 (hr) mismatch with International Phone Prefix at line 294 '__hr_858 dw 12, 384,858,0,0' Country ISO3166-1-A2 (me) mismatch with International Phone Prefix at line 320 '__me_858 dw 12,785,858,0,0' Country ISO3166-1-A2 (me) mismatch with International Phone Prefix at line 322 '__me_850 dw 12,785,850,0,0' Country ISO3166-1-A2 (me) mismatch with International Phone Prefix at line 324 '__me_864 dw 12,785,864,0,0' Errors = 22 Number of entries (171) So the problem (if there is any - they are only internal identifiers) countries are: YU - Yugoslavia (no longer exists), now several countries CZ - Czechoslovakia (by phone identifier) no longer exists, now two countries Czech Republic and Slovakia UK - United Kingdom - but should be 'GB' SR - Serbia HR - Croatia ME - (I looked up the phone prefix 785 which appears to be former Soviet Union) Also the made up identifiers seem fine '__nl_BE_850' means Dutch Language, Country Belgium, Codepage 850. Its identifier '40032' is made up of '4' means multilingual country, '0' first language within, and '032' original country code. I'm happy to update the UK -> GB identifiers. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[fritzmueller]

uff, middle east is at least the following (from german AI):   (Es gibt keine einzelne Telefonvorwahl für den gesamten Mittleren Osten; stattdessen hat jedes Land eine eigene internationale Ländervorwahl. Die meisten Länder in dieser Region nutzen Vorwahlen, die mit +9 (West- und Zentralasien) oder +2 (Nordostafrika) beginnen.  Hier sind die wichtigsten Ländervorwahlen für 2026:)   there is no prefix for the whole middle east, each country has an own international country code. Most countries in this region use prefixxes starting with +9(west and central asia) and +2(north east africa). I am not sure if these countries were ever in country.sys:   Arabische Halbinsel & Golfstaaten Saudi-Arabien: +966 Vereinigte Arabische Emirate: +971 Katar: +974 Oman: +968 Kuwait: +965 Bahrain: +973 Jemen: +967    Levante & Naher Osten Israel: +972 Palästina: +970 Jordanien: +962 Libanon: +961 Syrien: +963 Irak: +964    Weitere Länder der Region Türkei: +90 Iran: +98 Ägypten: +20 Zypern: +357      Sent: Tuesday, January 27, 2026 at 8:03 PM From: "Andrew Bird" ***@***.***> To: "FDOS/kernel" ***@***.***> Cc: "fritzmueller" ***@***.***>, "Mention" ***@***.***> Subject: Re: [FDOS/kernel] Update requested for country.sys on Freedos.devel (Issue #235) andrewbird left a comment (FDOS/kernel#235) Actually found a comment later in the country.asm file that indicates the intention of 'me' was middle east. No sure how that can work with international phone number prefix, unless there's an exception like '2' == FrenchCanada '3' == LatinAmerica — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[fritzmueller]

No idea, have a look at the following (sorry, german but should be self explainable) site:   https://www.placetel.de/blog/laendervorwahlen   Scroll to the top: Greenland belongs to: tadaa! Africa (green colour!) Fritz     Sent: Tuesday, January 27, 2026 at 8:03 PM From: "Andrew Bird" ***@***.***> To: "FDOS/kernel" ***@***.***> Cc: "fritzmueller" ***@***.***>, "Mention" ***@***.***> Subject: Re: [FDOS/kernel] Update requested for country.sys on Freedos.devel (Issue #235) andrewbird left a comment (FDOS/kernel#235) Actually found a comment later in the country.asm file that indicates the intention of 'me' was middle east. No sure how that can work with international phone number prefix, unless there's an exception like '2' == FrenchCanada '3' == LatinAmerica — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>