[PATCH] Test sorting order for tr_TR

[PATCH] Test sorting order for tr_TR

[Petter Reinholdtsen]

I got a sorting order test from the tr_TR author.  Please include in
CVS.  There is the changelog entry for localedata/ChangeLog: I hope
the UTF-8 encoded file survived this email.

2004-07-17  Petter Reinholdtsen  <pere@hungry.com>

	* Makefile (test-input): Add tr_TR.UTF-8 sorting test.  Test case
	for BZ #19.
	* tr_TR.in: New file

Index: localedata/Makefile
===================================================================
RCS file: /cvs/glibc/libc/localedata/Makefile,v
retrieving revision 1.100
diff -u -3 -p -u -r1.100 localedata/Makefile
--- localedata/Makefile	20 Nov 2003 23:31:38 -0000	1.100
+++ localedata/Makefile	17 Jul 2004 10:52:32 -0000
@@ -41,7 +41,7 @@ test-srcs := collate-test xfrm-test tst-
 	     tst-mbswcs1 tst-mbswcs2 tst-mbswcs3 tst-mbswcs4 tst-mbswcs5 \
 	     tst-ctype tst-wctype tst-langinfo tst-numeric
 test-input := de_DE.ISO-8859-1 en_US.ISO-8859-1 da_DK.ISO-8859-1 \
-	      hr_HR.ISO-8859-2 sv_SE.ISO-8859-1
+	      hr_HR.ISO-8859-2 sv_SE.ISO-8859-1 tr_TR.UTF-8
 test-input-data = $(addsuffix .in, $(basename $(test-input)))
 test-output := $(foreach s, .out .xout, \
 			 $(addsuffix $s, $(basename $(test-input))))
Index: localedata/Makefile
===================================================================
diff -u -3 -p -u -r1.100 localedata/tr_TR.in
--- localedata/tr_TR.in	Thu Jan  1 01:00:00 1970
+++ localedata/tr_TR.in	Sat Jul 17 12:45:29 2004
@@ -0,0 +1,126 @@
+A
+a
+Â
+â
+alem
+âlem
+Amasya
+ana
+anı
+bakkal
+bardak
+Bolu
+Ceyhan
+cuma
+cumhur
+Ç
+ç
+Çankırı
+çok
+çorba
+deniz
+derin
+Diyarbakır
+Edirne
+ezel
+eziyet
+Fatsa
+fincan
+firavun
+ganî
+gaz
+GümüÅŸhane
+Äž
+ÄŸ
+hala
+hâlâ
+havlu
+Havza
+I
+ı
+Iğdır
+ıslak
+ıspanak
+Ä°
+i
+ÃŽ
+î
+imlâ
+imza
+Ä°stanbul
+jale
+jandarma
+Japonya
+kâğıt
+Kırıkkale
+kimlik
+kiraz
+lâle
+limon
+lira
+Lüleburgaz
+Mardin
+mesih
+mesîh
+meskup
+meskûp
+mühür
+müstesna
+neden
+nezle
+NiÄŸde
+okçu
+OktaÅŸ
+okul
+Ordu
+Ö
+ö
+Ödemiş
+ördek
+örnek
+pırasa
+pırlanta
+Plevne
+quantum
+Quebec
+Recai
+Rize
+rüya
+rüzgar
+Samsun
+selâm
+süpürge
+sürahi
+Åž
+ÅŸ
+ÅŸemsiye
+ÅŸerbet
+Şereflikoçhisar
+tımar
+tırnak
+Trabzon
+U
+u
+Û
+û
+Urfa
+uygar
+uyluk
+Ü
+ü
+Üsküdar
+üzgün
+üzüm
+Van
+vilâyet
+virane
+wall
+Washington
+xanadu
+Xman
+yaÄŸmur
+yar
+Yozgat
+zehir
+zeki
+Zonguldak

Re: [PATCH] Test sorting order for tr_TR

[Petter Reinholdtsen]

[Petter Reinholdtsen, 2004-07-18]
> I got a sorting order test from the tr_TR author.  Please include in
> CVS.  There is the changelog entry for localedata/ChangeLog: I hope
> the UTF-8 encoded file survived this email.
>
> 2004-07-17  Petter Reinholdtsen  <pere@hungry.com>
>
> 	* Makefile (test-input): Add tr_TR.UTF-8 sorting test.  Test case
> 	for BZ #19.
> 	* tr_TR.in: New file

Is there anything controversial in this patch?  Will it help to
mention in the changelog that I got the test data from Recai Oktas
<roktas@omu.edu.tr> to get it accepted?  Or are you just all busy and
I just to impatient?

Should such test cases go through Bugzilla, or is it OK for me to just
pass them on directly?

Re: [PATCH] Test sorting order for tr_TR

[Ulrich Drepper]

Petter Reinholdtsen wrote:

> Is there anything controversial in this patch?

Yes.  The problem is that it requires yet another locale to be created
at test time.  And there is not even an error which is tested for.

To test the actual locale definition, and not the locale code in libc,
we need a separate target which the interested parties can run if
necessary.  I don't have the time to make all these changes to the
Makefile in the moment.

-- 
➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖

Re: [PATCH] Test sorting order for tr_TR

[Petter Reinholdtsen]

[Ulrich Drepper]
> Yes.  The problem is that it requires yet another locale to be created
> at test time.  And there is not even an error which is tested for.

It does?  As far as I can se the tr_TR.UTF-8 is already created at
test time, and all my patch is doing is adding a new sorting test
file to check that the sorting order of the locale is correct.

> To test the actual locale definition, and not the locale code in
> libc, we need a separate target which the interested parties can run
> if necessary.  I don't have the time to make all these changes to
> the Makefile in the moment.

Yes, this is a good idea.  We would be nice to be able to check both
the implementation and the content of the locales.  I'll see what I
can do.  :)

I guess the hard part will be to come up with a target name for
this. :)

Re: [PATCH] Test sorting order for tr_TR

[Ulrich Drepper]

Petter Reinholdtsen wrote:
> I got a sorting order test from the tr_TR author.  Please include in
> CVS. 

Not until the problem is figured out.  The test fails for me, several
lines are in a different order after the test.

-- 
➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖

Re: [PATCH] Test sorting order for tr_TR

[Petter Reinholdtsen]

[Ulrich Drepper]
> Not until the problem is figured out.  The test fails for me, several
> lines are in a different order after the test.

Oh.  It worked for me when I sent the patch a long time ago.  I'll
look into why it fail.  Do you have any more details?

Perhaps I forgot to remove my patch to get the sorting order test to
check _all_ characters in the file, and not ignore "\s+[:(;].*$" when
testing the sorting order of locales.  That is the only related change
I can think of, even though I'm pretty sure I tested the patch on a
clean CVS tree.

Re: [PATCH] Test sorting order for tr_TR

[Petter Reinholdtsen]

[-- Attachment #1: Type: text/plain, Size: 993 bytes --]


[Petter Reinholdtsen]
> Oh.  It worked for me when I sent the patch a long time ago.  I'll
> look into why it fail.  Do you have any more details?

I got the output of the failing test from Ulrich.

> Perhaps I forgot to remove my patch to get the sorting order test to
> check _all_ characters in the file, and not ignore "\s+[:(;].*$"
> when testing the sorting order of locales.  That is the only related
> change I can think of, even though I'm pretty sure I tested the
> patch on a clean CVS tree.

Nope, that is not it.  I tested again on a "clean" CVS tree, and the
test still work for me.

I'm starting to suspect some transfer encoding issue with the UTF-8
file.  Here is the patch again, this time gzipped and included as an
attachment to avoid any mail transfer encoding problems.  I updated
the date of the changelog entry.

2004-08-07  Petter Reinholdtsen  <pere@hungry.com>

	* Makefile (test-input): Add tr_TR.UTF-8 sorting test.  Test case
	for BZ #19.
	* tr_TR.in: New file


[-- Attachment #2: tr_TR UTF-8 sorting test data --]
[-- Type: application/octet-stream, Size: 1060 bytes --]

Re: [PATCH] Test sorting order for tr_TR

[Ulrich Drepper]

Petter Reinholdtsen wrote:

> I'm starting to suspect some transfer encoding issue with the UTF-8
> file.

Might be, it works for me now.

-- 
➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖