[tarantool-patches] Re: [PATCH v1 1/1] sql: UPPER and LOWER support COLLATE

[Vladislav Shpilevoy]

Hi! Thanks for the patch!

Please, write me in 'Copy' when sending a patch on review.

See 3 comments below.

On 27/07/2018 18:40, imeevma at tarantool.org wrote:
> SQL functions UPPER and LOWER now works
> with COLLATE as they should according to
> ANSI Standart.
> 
> Closes #3052.
> ---
> Branch: https://github.com/tarantool/tarantool/tree/imeevma/gh-3052-collate-for-upper-lower
> Issue: https://github.com/tarantool/tarantool/issues/3052
> 
>   src/box/sql/func.c          | 18 ++++++++++-----
>   test/sql/collation.result   | 53 +++++++++++++++++++++++++++++++++++++++++++++
>   test/sql/collation.test.lua | 20 +++++++++++++++++
>   3 files changed, 86 insertions(+), 5 deletions(-)
> 
> diff --git a/src/box/sql/func.c b/src/box/sql/func.c
> index e211de1..637121c 100644
> --- a/src/box/sql/func.c
> +++ b/src/box/sql/func.c
> @@ -503,6 +504,14 @@ case_type##ICUFunc(sqlite3_context *context, int argc, sqlite3_value **argv)   \
>   		return;                                                        \
>   	}                                                                      \
>   	UErrorCode status = U_ZERO_ERROR;                                      \
> +	struct coll *pColl = sqlite3GetFuncCollSeq(context);                   \

1. Please, use Tarantool style for new code.

> +	const char *locale = NULL;                                             \
> +	if (pColl != NULL) {                                                   \
> +		locale = ucol_getLocaleByType(pColl->collator,                 \
> +					      ULOC_VALID_LOCALE, &status);     \
> +	}                                                                      \
> +	UCaseMap *pUCaseMap = ucasemap_open(locale, 0, &status);               \
> +	assert(pUCaseMap);                                                     \
>   	int len = ucasemap_utf8To##case_type(pUCaseMap, z1, n, z2, n, &status);\
>   	if (len > n) {                                                         \
>   		status = U_ZERO_ERROR;                                         \
> diff --git a/test/sql/collation.result b/test/sql/collation.result
> index 7fec96d..e057bb6 100644
> --- a/test/sql/collation.result
> +++ b/test/sql/collation.result
> @@ -32,6 +32,59 @@ box.sql.execute("SELECT 1 LIMIT 1 COLLATE BINARY, 1;")
>   ---
>   - error: 'near "COLLATE": syntax error'
>   ...
> +-- gh-3052: sql: upper/lower support only default locale

2. I do not think that in /test/sql directory it is necessary to
specify 'sql:' subsystem.

> +-- For tr-TR result depends on collation
> +box.internal.collation.create('TURKISH', 'ICU', 'tr-TR', {strength='primary'});
> +---
> +...
> +box.sql.execute([[CREATE TABLE tu (descriptor CHAR(50) PRIMARY KEY, letter CHAR)]]);
> +---
> +...
> +box.sql.execute([[INSERT INTO tu VALUES ('Latin Capital Letter I U+0049','I');]])
> +---
> +...
> +box.sql.execute([[INSERT INTO tu VALUES ('Latin Small Letter I U+0069','i');]])
> +---
> +...
> +box.sql.execute([[INSERT INTO tu VALUES ('Latin Capital Letter I With Dot Above U+0130','İ');]])
> +---
> +...
> +box.sql.execute([[INSERT INTO tu VALUES ('Latin Small Letter Dotless I U+0131','ı');]])
> +---
> +...
> +-- Without collation
> +box.sql.execute([[SELECT descriptor, upper(letter) AS upper,lower(letter) AS lower FROM tu;]])
> +---
> +- - ['Latin Capital Letter I U+0049', 'I', 'i']
> +  - ['Latin Capital Letter I With Dot Above U+0130', 'İ', 'i̇']
> +  - ['Latin Small Letter Dotless I U+0131', 'I', 'ı']
> +  - ['Latin Small Letter I U+0069', 'I', 'i']
> +...
> +-- With collation
> +box.sql.execute([[SELECT descriptor, upper(letter COLLATE "TURKISH") AS upper,lower(letter COLLATE "TURKISH") AS lower FROM tu;]])
> +---
> +- - ['Latin Capital Letter I U+0049', 'I', 'ı']
> +  - ['Latin Capital Letter I With Dot Above U+0130', 'İ', 'i']
> +  - ['Latin Small Letter Dotless I U+0131', 'I', 'ı']
> +  - ['Latin Small Letter I U+0069', 'İ', 'i']
> +...
> +-- For de-DE result is actually the same
> +box.internal.collation.create('GERMAN', 'ICU', 'de-DE', {strength='primary'});
> +---
> +...
> +box.sql.execute([[INSERT INTO tu VALUES ('German Small Letter Sharp S U+00DF','ß');]])
> +---
> +...
> +-- Without collation
> +box.sql.execute([[SELECT descriptor, upper(letter), letter FROM tu where UPPER(letter) = 'SS';]])
> +---
> +- - ['German Small Letter Sharp S U+00DF', 'SS', 'ß']
> +...
> +-- With collation
> +box.sql.execute([[SELECT descriptor, upper(letter COLLATE "GERMAN"), letter FROM tu where UPPER(letter COLLATE "GERMAN") = 'SS';]])
> +---
> +- - ['German Small Letter Sharp S U+00DF', 'SS', 'ß']
> +...
>   box.schema.user.grant('guest','read,write,execute', 'universe')
>   ---
>   ...

3. Cleanup the test: drop tables, collations.