Re: [Toybox] Stat %Z - What are valid values?

2017-01-21 Thread Rich Felker 
On Sat, Jan 21, 2017 at 03:18:28PM -0600, Rob Landley wrote:
> > Or if it's signed, that's -1346458162 which would be... sometime in the
> > 1930's? hmmm... "./date -D %s -d -1346458162" is failing under glibc,
> > and failing _differently_ under musl. (Wheee.)
> > 
> > /me goes down tangent rathole debugging why.
> 
> > 
> > (Answer: musl doesn't implement %s at all, and glibc doesn't allow the
> > %s value it converts to be negative.)
> 
> Query: does bionic strptime() handle %s, and if so does it handle
> negative input values? (If not I suppose I can try to special case this
> in toybox, but ew.)
> 
> Also, Rich: any interest in adding this to musl?

strptime with %s? I suspect there are some nasty underspecified issues
with how it interacts with timezones. All the standard specifiers work
with broken-down (struct tm) time so timezone is irrelevant to how
they operate.

So the answer isn't no, but "it's complicated", and needs more
research on how other implementations work, if they're consistent,
pros and cons of different possible behaviors, etc.

Rich
___
Toybox mailing list
Toybox@lists.landley.net
http://lists.landley.net/listinfo.cgi/toybox-landley.net

Re: [Toybox] Stat %Z - What are valid values?

2017-01-21 Thread Rob Landley 
On 12/30/2016 04:01 PM, enh wrote:
> On Fri, Dec 30, 2016 at 1:51 PM, Rob Landley  wrote:
>> On 12/30/2016 01:39 PM, enh wrote:
>>> 1928-04-26 17:31:55.746667836 18446744072394174731
>>
>> Really we can partially blame posix here for not specifying whether
>> time_t is signed or unsigned. (If it's unsigned it can't represent times
>> before January 1, 1970, as the code on the left is doing. And there
>> _are_ times before then. So I think we have to treat it as signed and go
>> "32 bit timestamps gotta go away before 2038".)
> 
> if you want to be amused/horrified, you should look at the tzcode
> stuff. iirc they only recently agreed that time_t is an integral type.

Note to self: it's easier to review patches after a while if you include
the test case of the thing it fixes in the patch.

Right, this thread...

> Or if it's signed, that's -1346458162 which would be... sometime in the
> 1930's? hmmm... "./date -D %s -d -1346458162" is failing under glibc,
> and failing _differently_ under musl. (Wheee.)
> 
> /me goes down tangent rathole debugging why.
...
> 
> (Answer: musl doesn't implement %s at all, and glibc doesn't allow the
> %s value it converts to be negative.)

Query: does bionic strptime() handle %s, and if so does it handle
negative input values? (If not I suppose I can try to special case this
in toybox, but ew.)

Also, Rich: any interest in adding this to musl?

> with 64-bit toybox:
> 
> angler:/ # date 06022064
> Mon Jun  2 00:00:00 GMT 2064

Under Ubuntu 14.04 with 64-bit glibc, that becomes

  $ ./date 06022064
  date: bad date '06022064'; Fri Jun  2 00:00:00 CST 2064 != Mon
  Jun  2 01:00:00 CDT 2064

So it's adjusting it for daylight savings time, and breaking.

Sigh. _Why_ are we doing this date normalization checking thing again?
It's kinda brittle. I'm starting to remember why this got pushed onto
the todo stack. (Which is really a todo heap with a leaking problem.)

Rob
___
Toybox mailing list
Toybox@lists.landley.net
http://lists.landley.net/listinfo.cgi/toybox-landley.net