[#7476] Net::HTTP Bug in Ruby 1.8.4? — James Edward Gray II <james@...>
Can a Net::HTTP guru comment on this message:
[#7485] Bugzilla for ruby? — Hadmut Danisch <hadmut@...>
Hi,
[#7493] how to introduce reference objects into ruby — "Geert Fannes" <Geert.Fannes@...>
Hello,
[#7497] Re: how to introduce reference objects into ruby — "Geert Fannes" <Geert.Fannes@...>
Hello,
[#7500] Re: how to introduce reference objects into ruby — "Geert Fannes" <Geert.Fannes@...>
The problem with the code you sent is that you have to go through ALL
The columns store the actual values (doubles), and the rows store pointers to the corresponding doubles. This way, I can update a double directly via the columns, via the rows after dereferencing the pointers.
[#7518] Proposal: String#notempty? — Bertram Scharpf <lists@...>
Hi,
[#7524] Sefe level: bug or feature? — "Kirill A. Shutemov" <k.shutemov@...>
Why cannot do eval with $SAFE=3 and can with $SAFE=4? Is it bug or
Hi,
On Mon, 13 Mar 2006, Yukihiro Matsumoto wrote:
[#7529] Re: Proposal: String#notempty? — "Berger, Daniel" <Daniel.Berger@...>
> -----Original Message-----
[#7546] Re: how to introduce reference objects into ruby — "Geert Fannes" <Geert.Fannes@...>
In Ruby, there's the []= and [] operators which you can define together.
[#7553] "not" operator used in expression that is a method parameter can generate syntax error — noreply@...
Bugs item #3843, was opened at 2006-03-15 22:09
Hi,
Nobu, you are not answering to the question.... You have to unveil why
Hi,
Hello,
Zev Blut wrote:
On 3/16/06, Joel VanderWerf <vjoel@path.berkeley.edu> wrote:
On 3/16/06, Zev Blut <rubyzbibd@ubit.com> wrote:
Hello,
Yukihiro Matsumoto wrote:
On 3/16/06, mathew <meta@pobox.com> wrote:
Brian Mitchell wrote:
On 3/16/06, mathew <meta@pobox.com> wrote:
Dear all
What you've described is the basic predence difference between
Evan Phoenix wrote:
[#7600] ruby_script ? — "Nicolas Despr鑚" <nicolas.despres@...>
Hi list,
>>>>> "N" == Nicolas Despr=E8s?= <ISO-8859-1> writes:
On 3/25/06, ts <decoux@moulon.inra.fr> wrote:
>>>>> "N" == Nicolas Despr=E8s?= <ISO-8859-1> writes:
[#7601] to_str, to_s and StringValue — "Gerardo Santana Gez Garrido" <gerardo.santana@...>
If I understand correctly, StringValue is a way for writing duck-type
[#7614] PATCH: A subclassable Pathname — "Evan Phoenix" <evanwebb@...>
A simply change (changing all references of "Pathname.new" to
In article <92f5f81d0603262350k796fe48fp2224b9f2108ac507@mail.gmail.com>,
Quite right on the .glob and .getwd. I guess the tests don't test hit
In article <92f5f81d0603270903g2fb02244i6a395be708dfffa3@mail.gmail.com>,
In article <87fyl3x0wd.fsf@m17n.org>,
Hm, well, thats because of the shortcut behavior in Pathname#+ which
In article <92f5f81d0603271717r1ce51d30p6c28e363dc32a09b@mail.gmail.com>,
Re: GC Question
ts wrote:
>>>>>>"z" == zdennis <zdennis@mktec.com> writes:
>
>
> z> Thank you Guy for posting this with your comments. So this is a bug I am taking it, is this
> z> something that can be fixed rather easily or is this problem more difficult then that?
>
> Well this is not really a bug, this is how work a conservative GC which
> can mark object, because it find a reference on the stack. even if these
> objects are potentially dead (i.e. never referenced by other objects).
From some tests, it appears that ruby will taper off memory usage and reuse unused slots in the
heaps. My big worry was that I need to do processing on millions of records in a database. I didn't
want to process the data in chunks of 50,000, if each time I did a new query for a new set of 50,000
ruby wasn't going to GC well-enough to where I ran out of memory. It appears this will taper off at
a certain point.
Here are my results from querying 1,000 new records in iteration:
iteration: 0
Calling GC.start
String
count = 63,455
Array
count = 516
Mem usage: 10Mb
-----------
iteration: 1
Calling GC.start
String
count = 63,295
Array
count = 516
Mem usage: 11Mb
-----------
iteration: 2
Calling GC.start
String
count = 63,332
Array
count = 516
Mem usage: 11Mb
-----------
iteration: 3
Calling GC.start
String
count = 63,121
Array
count = 516
Mem usage: 11Mb
It tapers off at 11Mb. Now moving up to querying a new set of 50,000 records for each iteration:
iteration: 0
Calling GC.start
String
count = 2,708,846
Array
count = 531
Mem usage: 177Mb
-----------
iteration: 1
Calling GC.start
String
count = 2,657,524
Array
count = 531
Mem usage: 189Mb
-----------
iteration: 2
Calling GC.start
String
count = 2,631,215
Array
count = 531
Mem usage: 193Mb
-----------
iteration: 3
Calling GC.start
String
count = 2,656,035
Array
count = 531
Mem usage: 189Mb
-----------
Calling GC.start
String
count = 9,794
Array
count = 514
Mem usage: 189Mb
It tapers off at 189Mb, which I can live with. I was afraid it would jump 189Mb to 300Mb to 500Mb,
and so on.
>
> Like I've said I think that the Boehm GC has some way to "clean" a stack
> when it's unused, just for this reason (to don't retrieve old reference to
> an object) but I don't think that :
>
> 1) it can solve *all* these problems
>
> 2) it's easy to do it
>
>
> Some time you must just try to adapt you to the ruby GC, which generally do a
> really good job.
>
I am trying to adapt. I think I need to become more familiar with Garbage Collection. It would be
very useful if ruby GC's was compacting.
Thank you for your reply,
Zach