bug#31644: tests/ls/abmon-align does not work today (May 29th)

2018-05-30 Thread Kamil Dudka
On Tuesday, May 29, 2018 7:16:43 PM CEST Pádraig Brady wrote:
> On 29/05/18 09:16, Kamil Dudka wrote:
> > Re-posting with a fresh subject.  debbugs.gnu.org rejected my original
> > post
> > because the original bug has been closed and has received no comments for
> > more than 28 days:
> > 
> > https://debbugs.gnu.org/cgi/bugreport.cgi?bug=30814
> > 
> > On Wednesday, March 14, 2018 7:40:31 PM CEST Pádraig Brady wrote:
> >> Given the increase in supported size should only impact relatively few
> >> languages it probably makes sense to increase to 12. The attached does
> >> that
> >> and also augments the test to find ambiguous cases.
> > 
> > The ls/abmon-align test does not work today (May 29th) because of the
> > added
> > 
> > check.  The problem is that:
> > touch '-d+N month' for N={01..12}
> > 
> > ... in general does not create 12 files with timestamps in different
> > months.
> > 
> > If you run it today (May 29th), touch '-d+09 month' results in March 1st
> > while touch '-d+10 month' results in March 29th.  Consequently, both
> > 09.ts and 10.ts have the same month (March) to begin with and the check
> > for duplicates has to
> > fail:
> That was a latent issue with the original test, which I've
> now pushed a fix for at:
> https://git.sv.gnu.org/gitweb/?p=coreutils.git;a=commitdiff;h=c8eb21c

This works for me.  Thanks!

Kamil

> I've a local prototype to address the more general gotcha here
> wrt relative date adjustments, which I hope to push soon.
> 
> thanks!
> Pádraig







bug#31644: tests/ls/abmon-align does not work today (May 29th)

2018-05-29 Thread Pádraig Brady
On 29/05/18 09:16, Kamil Dudka wrote:
> Re-posting with a fresh subject.  debbugs.gnu.org rejected my original post 
> because the original bug has been closed and has received no comments for 
> more 
> than 28 days:
> 
> https://debbugs.gnu.org/cgi/bugreport.cgi?bug=30814
> 
> 
> On Wednesday, March 14, 2018 7:40:31 PM CEST Pádraig Brady wrote:
>> Given the increase in supported size should only impact relatively few
>> languages it probably makes sense to increase to 12. The attached does that
>> and also augments the test to find ambiguous cases.
> 
> The ls/abmon-align test does not work today (May 29th) because of the added 
> check.  The problem is that:
> 
> touch '-d+N month' for N={01..12}
> 
> ... in general does not create 12 files with timestamps in different months. 
> 
> If you run it today (May 29th), touch '-d+09 month' results in March 1st 
> while 
> touch '-d+10 month' results in March 29th.  Consequently, both 09.ts and 
> 10.ts 
> have the same month (March) to begin with and the check for duplicates has to 
> fail:

That was a latent issue with the original test, which I've
now pushed a fix for at:
https://git.sv.gnu.org/gitweb/?p=coreutils.git;a=commitdiff;h=c8eb21c

I've a local prototype to address the more general gotcha here
wrt relative date adjustments, which I hope to push soon.

thanks!
Pádraig





bug#31644: tests/ls/abmon-align does not work today (May 29th)

2018-05-29 Thread Kamil Dudka
Re-posting with a fresh subject.  debbugs.gnu.org rejected my original post 
because the original bug has been closed and has received no comments for more 
than 28 days:

https://debbugs.gnu.org/cgi/bugreport.cgi?bug=30814


On Wednesday, March 14, 2018 7:40:31 PM CEST Pádraig Brady wrote:
> Given the increase in supported size should only impact relatively few
> languages it probably makes sense to increase to 12. The attached does that
> and also augments the test to find ambiguous cases.

The ls/abmon-align test does not work today (May 29th) because of the added 
check.  The problem is that:

touch '-d+N month' for N={01..12}

... in general does not create 12 files with timestamps in different months. 

If you run it today (May 29th), touch '-d+09 month' results in March 1st while 
touch '-d+10 month' results in March 29th.  Consequently, both 09.ts and 10.ts 
have the same month (March) to begin with and the check for duplicates has to 
fail:

++ echo 'Jun
Jul
Aug
Sep
Oct
Nov
Dec
Jan
Mar
Mar
Apr
May'
++ sort
++ uniq -d
++ wc -l
+ n_dupes=1
+ test 1 = 1
+ test 1 = 0
+ fail=1
+ break 2
+ test 1 = 1
+ echo 'misalignment or ambiguous output in C locale:'
misalignment or ambiguous output in C locale:
+ LC_ALL=C
+ TIME_STYLE=+%b
+ ls -lgG 01.ts 02.ts 03.ts 04.ts 05.ts 06.ts 07.ts 08.ts 09.ts 10.ts 11.ts 
12.ts
-rw-rw-r-- 1 0 Jun 01.ts
-rw-rw-r-- 1 0 Jul 02.ts
-rw-rw-r-- 1 0 Aug 03.ts
-rw-rw-r-- 1 0 Sep 04.ts
-rw-rw-r-- 1 0 Oct 05.ts
-rw-rw-r-- 1 0 Nov 06.ts
-rw-rw-r-- 1 0 Dec 07.ts
-rw-rw-r-- 1 0 Jan 08.ts
-rw-rw-r-- 1 0 Mar 09.ts
-rw-rw-r-- 1 0 Mar 10.ts
-rw-rw-r-- 1 0 Apr 11.ts
-rw-rw-r-- 1 0 May 12.ts