Hi again
The diff is quite huge (- = in our list and not on web, + = listed on web and not it our list):
-ACTION -AFTER -AGAINST -AGGREGATE -ANALYSE -AUTOCOMMIT
i agree to adding any reserved words that are on the mysql-website but not in the list, but i wouldn't delete those that are in the list and not on the web - that would mean that you can have a table named 'autocommit'?? even if mysql does allow it, this can only mean trouble. Robin would your parser mind that?
regards
Mike Beck mikebeck@users.sourceforge.net
On 15.04.2003 12:11, Beck, Mike wrote:
i agree to adding any reserved words that are on the mysql-website but not in the list, but i wouldn't delete those that are in the list and not on the web - that would mean that you can have a table named 'autocommit'?? even if mysql does allow it, this can only mean trouble. Robin would your parser mind that?
I don't see why we should have something different than MySQL does. With current version it has some bad side effects:
On MySQL commandline: "create table autocommit (i int);" goes okay, in phpMyAdmin it is converted to "CREATE TABLE AUTOCOMMIT (i int)" so it creates uppercased table. This is IMHO something we don't want...
If we wouldn't change case of reserved words, it would be okay, but we do change this so we should have same reserved words as MySQL does.
On Tue, Apr 15, 2003 at 12:15:48PM +0200, Michal ?iha? wrote:
On 15.04.2003 12:11, Beck, Mike wrote:
i agree to adding any reserved words that are on the mysql-website but not in the list, but i wouldn't delete those that are in the list and not on the web - that would mean that you can have a table named 'autocommit'?? even if mysql does allow it, this can only mean trouble. Robin would your parser mind that?
I don't see why we should have something different than MySQL does. With current version it has some bad side effects: On MySQL commandline: "create table autocommit (i int);" goes okay, in phpMyAdmin it is converted to "CREATE TABLE AUTOCOMMIT (i int)" so it creates uppercased table. This is IMHO something we don't want... If we wouldn't change case of reserved words, it would be okay, but we do change this so we should have same reserved words as MySQL does.
Previously I have found that the website list is NOT up to date. Instead I have based the list on the actual parser inside the MySQL server (look in the yacc and flex files).
On 15.04.2003 03:25, Robin H. Johnson wrote:
On Tue, Apr 15, 2003 at 12:15:48PM +0200, Michal ?iha? wrote:
On 15.04.2003 12:11, Beck, Mike wrote:
i agree to adding any reserved words that are on the mysql-website but not in the list, but i wouldn't delete those that are in the list and not on the web - that would mean that you can have a table named 'autocommit'?? even if mysql does allow it, this can only mean trouble. Robin would your parser mind that?
I don't see why we should have something different than MySQL does. With current version it has some bad side effects: On MySQL commandline: "create table autocommit (i int);" goes okay, in phpMyAdmin it is converted to "CREATE TABLE AUTOCOMMIT (i int)" so it creates uppercased table. This is IMHO something we don't want... If we wouldn't change case of reserved words, it would be okay, but we do change this so we should have same reserved words as MySQL does.
Previously I have found that the website list is NOT up to date. Instead I have based the list on the actual parser inside the MySQL server (look in the yacc and flex files).
Okay, I looked into source, but I didn't see it in yacc files clearly, so I used sql-bench/crash-me.sh, which tries which word is really treated as reserved and which not (results are also in sources - sql-bench/limits/mysql.cfg). As little surprise this list has only 166 entries :-). If somebody interested in diff to our list, it is attached.
The question is: Which should we use?
On Tue, Apr 15, 2003 at 02:09:05PM +0200, Michal ?iha? wrote:
Previously I have found that the website list is NOT up to date. Instead I have based the list on the actual parser inside the MySQL server (look in the yacc and flex files).
Okay, I looked into source, but I didn't see it in yacc files clearly, so I used sql-bench/crash-me.sh, which tries which word is really treated as reserved and which not (results are also in sources - sql-bench/limits/mysql.cfg). As little surprise this list has only 166 entries :-). If somebody interested in diff to our list, it is attached. The question is: Which should we use?
look at sql/lex.h MANY more symbols there.
-
Beck, Mike -
Michal Čihař -
Michal Čihař -
Robin H. Johnson -
Robin H.Johnson