Обсуждение: pgsql: pg_test_fsync: update output to show usecs/op clearer
pg_test_fsync: update output to show usecs/op clearer Branch ------ master Details ------- http://git.postgresql.org/pg/commitdiff/095018bc32b70eff95bdf3c2959e6eef09dcfa6a Modified Files -------------- contrib/pg_test_fsync/pg_test_fsync.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
On Thu, May 2, 2013 at 7:27 AM, Bruce Momjian <bruce@momjian.us> wrote: > pg_test_fsync: update output to show usecs/op clearer I don't think that this is an improvement. Not everyone knows that usec is shorthand for microsecond. If I Google usec, I see something about the United States Enrichment Corporation, but that's about it. -- Peter Geoghegan
Peter Geoghegan <pg@heroku.com> writes: > On Thu, May 2, 2013 at 7:27 AM, Bruce Momjian <bruce@momjian.us> wrote: >> pg_test_fsync: update output to show usecs/op clearer > > I don't think that this is an improvement. Not everyone knows that > usec is shorthand for microsecond. Use unicode super power and ouput µsec, as it's meant to be? Then google is useful again, for the smaller number of people who would need it. Regards, -- Dimitri Fontaine http://2ndQuadrant.fr PostgreSQL : Expertise, Formation et Support
On Thu, May 2, 2013 at 08:38:39AM -0700, Peter Geoghegan wrote:
> On Thu, May 2, 2013 at 7:27 AM, Bruce Momjian <bruce@momjian.us> wrote:
> > pg_test_fsync: update output to show usecs/op clearer
>
> I don't think that this is an improvement. Not everyone knows that
> usec is shorthand for microsecond. If I Google usec, I see something
> about the United States Enrichment Corporation, but that's about it.
Well, we use 'usecs' all over our code, and in pg_test_timing:
Histogram of timing durations:
< usec % of total count
1 95.86499 69510498
2 4.13395 2997478
4 0.00062 453
8 0.00042 301
16 0.00001 8
32 0.00000 1
Microseconds just isn't going to fit there. Postgres understands usec:
test=> select '1 usec'::interval;
interval
-----------------
00:00:00.000001
(1 row)
I think the people running these tools are going to know usec.
--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ It's impossible for everything to be true. +
On Thu, May 2, 2013 at 05:44:07PM +0200, Dimitri Fontaine wrote: > Peter Geoghegan <pg@heroku.com> writes: > > On Thu, May 2, 2013 at 7:27 AM, Bruce Momjian <bruce@momjian.us> wrote: > >> pg_test_fsync: update output to show usecs/op clearer > > > > I don't think that this is an improvement. Not everyone knows that > > usec is shorthand for microsecond. > > Use unicode super power and ouput µsec, as it's meant to be? Then google > is useful again, for the smaller number of people who would need it. That would be cool, but I don't think we can assume everyone running these tools is on a UTF8-enabled terminal. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com + It's impossible for everything to be true. +
On 2013-05-02 12:09:22 -0400, Bruce Momjian wrote: > On Thu, May 2, 2013 at 05:44:07PM +0200, Dimitri Fontaine wrote: > > Peter Geoghegan <pg@heroku.com> writes: > > > On Thu, May 2, 2013 at 7:27 AM, Bruce Momjian <bruce@momjian.us> wrote: > > >> pg_test_fsync: update output to show usecs/op clearer > > > > > > I don't think that this is an improvement. Not everyone knows that > > > usec is shorthand for microsecond. > > > > Use unicode super power and ouput µsec, as it's meant to be? Then google > > is useful again, for the smaller number of people who would need it. > > That would be cool, but I don't think we can assume everyone running > these tools is on a UTF8-enabled terminal. I don't care about the symbol here, but I think psql has code to determine that for its line drawing stuff... Greetings, Andres Freund -- Andres Freund http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Training & Services
On Thu, May 2, 2013 at 06:13:11PM +0200, Andres Freund wrote: > On 2013-05-02 12:09:22 -0400, Bruce Momjian wrote: > > On Thu, May 2, 2013 at 05:44:07PM +0200, Dimitri Fontaine wrote: > > > Peter Geoghegan <pg@heroku.com> writes: > > > > On Thu, May 2, 2013 at 7:27 AM, Bruce Momjian <bruce@momjian.us> wrote: > > > >> pg_test_fsync: update output to show usecs/op clearer > > > > > > > > I don't think that this is an improvement. Not everyone knows that > > > > usec is shorthand for microsecond. > > > > > > Use unicode super power and ouput µsec, as it's meant to be? Then google > > > is useful again, for the smaller number of people who would need it. > > > > That would be cool, but I don't think we can assume everyone running > > these tools is on a UTF8-enabled terminal. > > I don't care about the symbol here, but I think psql has code to > determine that for its line drawing stuff... Uh, yeah, I think it does. I question whether a command-line diagnostic tool really should use it though. It might also make grep of the result difficult. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com + It's impossible for everything to be true. +
Bruce Momjian <bruce@momjian.us> writes: > Uh, yeah, I think it does. I question whether a command-line diagnostic > tool really should use it though. It might also make grep of the result > difficult. Hey, it's 2013 outside. Just saying: $ echo 'µsec' | grep µsec µsec $ echo 'µsec' | awk '/µsec/' µsec Maybe there's no practical problem here, -- Dimitri Fontaine http://2ndQuadrant.fr PostgreSQL : Expertise, Formation et Support
On Thu, May 2, 2013 at 06:29:34PM +0200, Dimitri Fontaine wrote:
> Bruce Momjian <bruce@momjian.us> writes:
> > Uh, yeah, I think it does. I question whether a command-line diagnostic
> > tool really should use it though. It might also make grep of the result
> > difficult.
>
> Hey, it's 2013 outside. Just saying:
>
> $ echo 'µsec' | grep µsec
> µsec
> $ echo 'µsec' | awk '/µsec/'
> µsec
>
> Maybe there's no practical problem here,
My point is that tools would need to adjust to output not knowing if
they are getting UTF8 output, e.g.:
grep '(µ|u)sec'
Do any of the OS tools output UTF8 bytes as headers? I have never seen
it.
--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ It's impossible for everything to be true. +
Bruce Momjian <bruce@momjian.us> writes: > My point is that tools would need to adjust to output not knowing if > they are getting UTF8 output, e.g.: Ah, yes. Or use "LANG=C pg_test_fsync" or the documented stanza that always works in scripts as it forces the output to something predictable? -- Dimitri Fontaine http://2ndQuadrant.fr PostgreSQL : Expertise, Formation et Support
On Thu, May 2, 2013 at 06:41:53PM +0200, Dimitri Fontaine wrote: > Bruce Momjian <bruce@momjian.us> writes: > > My point is that tools would need to adjust to output not knowing if > > they are getting UTF8 output, e.g.: > > Ah, yes. Or use "LANG=C pg_test_fsync" or the documented stanza that > always works in scripts as it forces the output to something > predictable? Actually, pg_upgrade does that so it gets the output always in English so it can do its comparisons. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com + It's impossible for everything to be true. +
On Thu, May 2, 2013 at 4:44 PM, Dimitri Fontaine <dimitri@2ndquadrant.fr> wrote: > Use unicode super power and ouput µsec, as it's meant to be? Then google > is useful again, for the smaller number of people who would need it. The SI abbreviation would be µs actually. I don't think it's worth spending an U+03B9 of time on this however. -- greg