Hi Hans, 1. changing the English sorting rules as you suggested had no effect, neither did the “add_uppercase_mappings('pl',1)”. 2. I think my original question was stated imprecisly, so let me emphasize what I'm after: Suppose you've got three string aaa, Aaa and aab. They are tested _as if_ they had the same case, i.e. “aaa == Aaa” (the sorter returns 0). Then (only if the case-indifferent test returned equal) another check is done for the _first_ character only. If both strings differ in case of the first char, then the string with the lowercase one gets precedence. The correct order will be: [1] = "aaa", [2] = "Aaa", [3] = "aab" whereas with uppercase after lowercase (as I understand it) you'd get: [1] = "aaa", [2] = "aab", [3] = "Aaa" And that is why I extended the splitter (a) to keep the state of the first character as a boolean as well as (b) to return lowercase sort strings, and the comparer to do an extra check for this whenever basicsort returns 0. I really don't expect you to change the sorter, far from it. Perhaps you can keep an extra comparer around to do the job -- after all the table is called “comparers” but for now contains only a single one. Same for splitters. And as this rule seems to be quite popular around the world it might probably become useful someday. If you decide against it I'll just put it on the wiki which will be fine enough, I guess. Philipp On 2010-08-19 <00:48:14>, Hans Hagen wrote:

On 18-8-2010 6:08, Philipp Gesang wrote:

...

Hi,

I'm creating some sorting tables. While researching this topic I stumbled on the Polish dictionary sorting rules: if two strings are equal except for case then the one gets precedence that begins lowercase.[1] (This seems to apply to the Swedish order as well but I have no means to verify that. Apparently, my German dictionary (from 1991) follows the same rule without explicitly stating so.)

Context seems to prefer it the other way round, so I modified two functions from sort-ini.lua to handle that; but I'm not happy with this solution.

So my question: is there already, or could we have some mechanism to influence the details of sorting in context?

i wonder if this works out ok (needs a test index):

sorters.replacements["pl"] = { -- no replacements }

sorters.entries["pl"] = { ["a"] = "a", ["ą"] = "ą", ["b"] = "b", ["c"] = "c", ["ć"] = "ć", ["d"] = "d", ["e"] = "e", ["ę"] = "ę", ["f"] = "f", ["g"] = "g", ["h"] = "h", ["i"] = "i", ["j"] = "j", ["k"] = "k", ["l"] = "l", ["ł"] = "ł", ["m"] = "m", ["n"] = "n", ["ń"] = "ń", ["o"] = "o", ["ó"] = "ó", ["p"] = "p", ["q"] = "q", ["r"] = "r", ["s"] = "s", ["ś"] = "ś", ["t"] = "t", ["u"] = "u", ["v"] = "v", ["w"] = "w", ["x"] = "x", ["y"] = "y", ["z"] = "z", ["ź"] = "ź", ["ż"] = "ż", }

sorters.mappings["pl"] = { ["a"] = 1, ["ą"] = 2, ["b"] = 3, ["c"] = 4, ["ć"] = 5, ["d"] = 6, ["e"] = 7, ["ę"] = 8, ["f"] = 9, ["g"] = 10, ["h"] = 11, ["i"] = 12, ["j"] = 13, ["k"] = 14, ["l"] = 15, ["ł"] = 16, ["m"] = 17, ["n"] = 18, ["ń"] = 19, ["o"] = 20, ["ó"] = 21, ["p"] = 22, ["q"] = 23, ["r"] = 24, ["s"] = 25, ["ś"] = 26, ["t"] = 27, ["u"] = 28, ["v"] = 29, ["w"] = 30, ["x"] = 31, ["y"] = 32, ["z"] = 33, ["ź"] = 34, ["ż"] = 35, }

add_uppercase_entries ('pl') add_uppercase_mappings('pl',1)

----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl ----------------------------------------------------------------- ___________________________________________________________________________________ If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___________________________________________________________________________________

-- () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments

Hi,

a aaa 1 Aaa 1 aab 1 Aab 1

aaa aab Aaa Aab

so, what order do you expect here? Exactly this! Could you please post an example how to achieve this in lua using comparers.basic (the registers implementation gives me

On 2010-08-19 <11:41:43>, Hans Hagen wrote: headaches)? And how can I invert this behaviour so that [1] = "Aaa", [2] = "aaa", [3] = "Aab", &c.? Language settings don't seem to have an effect on the order. Thanks, Philipp PS: Did you just include the table from my example into the beta? Wow, that's fast! There's more to come …

----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl ----------------------------------------------------------------- ___________________________________________________________________________________ If your question is of interest to others as well, please add an entry to the Wiki!

maillist : ntg-context@ntg.nl / http://www.ntg.nl/mailman/listinfo/ntg-context webpage : http://www.pragma-ade.nl / http://tex.aanhet.net archive : http://foundry.supelec.fr/projects/contextrev/ wiki : http://contextgarden.net ___________________________________________________________________________________

-- () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments