Обсуждение: increase message string buffer size of watch command of psql
Hello. In po.ko (korean message) at psql #: command.c:2971 #, c-format msgid "Watch every %lds\t%s" msgstr "%ld초 간격으로 지켜보기\t%s" this message string is a cut string, because buffer size is small. At line 2946 in src/bin/psql/command.c char title[50]; size of message string for korean is over 50 bytes. (at least 80 columns terminal for common) Increase size of this title, please. 50 bytes is so small for multi language. And. I suggest that date string might be local language, or current_timestamp string. Regards, Ioseph.
On Sun, Jun 12, 2016 at 10:55 AM, Ioseph Kim <pgsql-kr@postgresql.kr> wrote: > Hello. > In po.ko (korean message) at psql > #: command.c:2971 > #, c-format > msgid "Watch every %lds\t%s" > msgstr "%ld초 간격으로 지켜보기\t%s" > > this message string is a cut string, because buffer size is small. > At line 2946 in src/bin/psql/command.c > char title[50]; > > size of message string for korean is over 50 bytes. > (at least 80 columns terminal for common) > > Increase size of this title, please. > 50 bytes is so small for multi language. > > And. I suggest that date string might be local language, > or current_timestamp string. This was already changed in commit dea2b5960. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company
Robert Haas <robertmhaas@gmail.com> writes: > On Sun, Jun 12, 2016 at 10:55 AM, Ioseph Kim <pgsql-kr@postgresql.kr> wrote: >> Increase size of this title, please. >> 50 bytes is so small for multi language. >> >> And. I suggest that date string might be local language, >> or current_timestamp string. > This was already changed in commit dea2b5960. Well, we did part of that, but it's still using asctime(). Should we change that to strftime(..."%c"...) to be less English-centric? (The result seems to be the same in C locale. pg_controldata has done it that way for a long time, with few complaints.) If we want to do so, now would be the time, since 9.6 already whacked around the format of \watch output. regards, tom lane
I wrote: > Robert Haas <robertmhaas@gmail.com> writes: >> On Sun, Jun 12, 2016 at 10:55 AM, Ioseph Kim <pgsql-kr@postgresql.kr> wrote: >>> Increase size of this title, please. >>> 50 bytes is so small for multi language. >>> And. I suggest that date string might be local language, >>> or current_timestamp string. >> This was already changed in commit dea2b5960. > Well, we did part of that, but it's still using asctime(). Should we > change that to strftime(..."%c"...) to be less English-centric? > (The result seems to be the same in C locale. pg_controldata has done > it that way for a long time, with few complaints.) If we want to do so, > now would be the time, since 9.6 already whacked around the format > of \watch output. I take it from the vast silence that nobody particularly cares one way or the other. On reflection I think that this would be a good change to make, so I'll go do so unless I hear complaints soon. regards, tom lane
2016년 06월 15일 01:56에 Tom Lane 이(가) 쓴 글: > I take it from the vast silence that nobody particularly cares one way > or the other. On reflection I think that this would be a good change > to make, so I'll go do so unless I hear complaints soon. regards, tom > lane I propose to change from asctime() to sql current_timestamp value, then users will change date format with set command DateStyle. Regards, Ioseph.
Ioseph Kim <pgsql-kr@postgresql.kr> writes: > 2016년 06월 15일 01:56에 Tom Lane 이(가) 쓴 글: >> I take it from the vast silence that nobody particularly cares one way >> or the other. On reflection I think that this would be a good change >> to make, so I'll go do so unless I hear complaints soon. regards, tom >> lane > I propose to change from asctime() to sql current_timestamp value, > then users will change date format with set command DateStyle. That would require an additional SQL query each time through the loop, which seems like undue expense. It's also not terribly friendly to the goal of localization, I should think, given the limited number of datestyle options and the fact that none of them actually change day or month names to non-English choices. And it would imply changing the timestamps from psql's timezone to the backend's. While that might have been a good way to do it in a green field, it's not the way \watch has worked in the past, and given the lack of complaints I'm disinclined to change that. regards, tom lane
Tom Lane wrote: > I wrote: > > Well, we did part of that, but it's still using asctime(). Should we > > change that to strftime(..."%c"...) to be less English-centric? > > (The result seems to be the same in C locale. pg_controldata has done > > it that way for a long time, with few complaints.) If we want to do so, > > now would be the time, since 9.6 already whacked around the format > > of \watch output. > > I take it from the vast silence that nobody particularly cares one way > or the other. On reflection I think that this would be a good change > to make, so I'll go do so unless I hear complaints soon. +1 to strftime("%c"). If we wanted to provide additional flexibility we could have a \pset option to change the strftime format string to something other than %c, but I don't think there's enough demand to justify it. -- Álvaro Herrera http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
Alvaro Herrera <alvherre@2ndquadrant.com> writes: > +1 to strftime("%c"). If we wanted to provide additional flexibility we > could have a \pset option to change the strftime format string to > something other than %c, but I don't think there's enough demand to > justify it. Agreed, that seems like something for later (or never). Pushed that way. I also widened the buffer a bit in the back branches, to address the original complaint. regards, tom lane
Thanks, I agree that strftime() is better then asctime(). regards, Ioseph 2016년 06월 16일 08:37에 Tom Lane 이(가) 쓴 글: > Alvaro Herrera <alvherre@2ndquadrant.com> writes: >> +1 to strftime("%c"). If we wanted to provide additional flexibility we >> could have a \pset option to change the strftime format string to >> something other than %c, but I don't think there's enough demand to >> justify it. > Agreed, that seems like something for later (or never). Pushed that way. > > I also widened the buffer a bit in the back branches, to address the > original complaint. > > regards, tom lane