Pike-devel

pike-devel@lists.lysator.liu.se

4 participants
3044 discussions

Optimizations
by Martin Nilsson (Fake Build Master) ＠ Pike (-) developers forum 04 Nov '02

04 Nov '02

I've run some of the shootout tests with disabled optimizations and found some surprising results: Pike: pike PikeNP: pike -dp (without peephole optimization) PikeNT: pike -dt (without tail recursion) PikeNTP: pike -dp -dt Pike PikeNP PikeNT PikeNTP 1: 42.81 32.83 37.49 34.75 (ackermann) 2: 34.92 37.05 37.64 37.80 (ary3) 3: 148.92 304.90 118.21 128.94 (fibo) Is the test flawed or is there something fishy in the optimization. The compilation overhead between Pike with and without optimizations is at most 0.13s.

4 5

lyskom export problems?
by Martin Baehr 04 Nov '02

04 Nov '02

the export mailing list is getting multiple copies of each message with different dates (2-3 minutes different) since today. what's happening? is the exporter trying to compensate for the lack of messages over the last few days? greetings, martin.

2 1

Backend stuff in 7.3
by Marek Habersack (Grendel) ＠ Pike (-) developers forum 03 Nov '02

03 Nov '02

Would it be possible to get a quick overview about how to use it? Looking through backend.cmod doesn't give much information. If I understand that well, Backend_struct is a transparent structure used in a generic way by backend.cmod - if it is so, where can I find specs about what does backend.cmod expect to find in the structure? I'd like to implement my own I/O backend and it seems that Backend_struct somehow makes it possible. How? :) TIA

1 0

Re: Proposal for new ADT interface
by Martin Stjernholm, Roxen IS ＠ Pike developers forum 03 Nov '02

03 Nov '02

> (when `== on the data elements are true?). Rather when a < b and a > b are both false. But that's right; even in ordered collections it would be wise to keep an integer position in the iterators as often as possible. / Martin Stjernholm, Roxen IS Previous text: >2002-11-03 15:55: >Subject: Re: Proposal for new ADT interface >-------------------------------------------------------------------- >> Well, that could require an element walk through in some cases, e.g. >> in the new multiset implementation, which would be more expensive. >> Anyway, if there is an order then "before" should always be the same >> as "less" so it doesn't matter much. > >Yes. Since the multiset order is known, it can very well cheat and >only do walking when needed (when `== on the data elements are true?). > >Giving an error when the iterators aren't comparable seems like a good >behaviour (iterators from different or edited multisets or mappings, >for instance), since doing < and > comparison on those are bound to be >a mistake. > >Any plans on a Heap, btw? > > > / Brevbäraren >

1 0

Re: Proposal for new ADT interface
by Martin Nilsson (Fake Build Master) ＠ Pike (-) developers forum 03 Nov '02

03 Nov '02

Plans to do what? Add an iterator to ADT.Heap? / Martin Nilsson (Fake Build Master) Previous text: >2002-11-03 15:55: >Subject: Re: Proposal for new ADT interface >-------------------------------------------------------------------- >> Well, that could require an element walk through in some cases, e.g. >> in the new multiset implementation, which would be more expensive. >> Anyway, if there is an order then "before" should always be the same >> as "less" so it doesn't matter much. > >Yes. Since the multiset order is known, it can very well cheat and >only do walking when needed (when `== on the data elements are true?). > >Giving an error when the iterators aren't comparable seems like a good >behaviour (iterators from different or edited multisets or mappings, >for instance), since doing < and > comparison on those are bound to be >a mistake. > >Any plans on a Heap, btw? > > > / Brevbäraren >

1 0

Re: Proposal for new ADT interface
by Martin Stjernholm, Roxen IS ＠ Pike developers forum 03 Nov '02

03 Nov '02

> Anyway, the iterator's `< shouldn't compare the data values, it should > compare the iterators - "before", not "less". Well, that could require an element walk through in some cases, e.g. in the new multiset implementation, which would be more expensive. Anyway, if there is an order then "before" should always be the same as "less" so it doesn't matter much. / Martin Stjernholm, Roxen IS Previous text: >2002-11-03 08:00: >Subject: Re: Proposal for new ADT interface >-------------------------------------------------------------------- >> Not really. If you have an unordered set you might get two iterators >> that iterates over all members in different order. > >Of course. But with for instance mappings, the iterator uses the same >order until the mapping is changed or indexed. (Then the operator can >give an error, just like when the operator is used on an iterator over >some other data.) > >Anyway, the iterator's `< shouldn't compare the data values, it should >compare the iterators - "before", not "less". > > > / Brevbäraren >

1 0

Re: Proposal for new ADT interface
by Martin Nilsson (Fake Build Master) ＠ Pike (-) developers forum 03 Nov '02

03 Nov '02

Not really. If you have an unordered set you might get two iterators that iterates over all members in different order. / Martin Nilsson (Fake Build Master) Previous text: >2002-11-02 22:56: >Subject: Re: Proposal for new ADT interface >-------------------------------------------------------------------- >> Ok. Good point. Come to think about it, iterators in ordered >> collections ought to have `< and `> too. > >Shouldn't all iterators? >An iterator is by definition doing something ordered, isn't it...? > > > / Brevbäraren >

4 6

Re: unix_time and UTC
by Johan Sundstr�m (�rkehertig av Lysators rootgrupp) ＠ Pike (-) developers forum 03 Nov '02

03 Nov '02

Noted. Thanks for the clarifications. / Johan Sundström (ärkehertig av Lysators rootgrupp) Previous text: >2002-11-03 08:03: >Subject: Re: unix_time and UTC >-------------------------------------------------------------------- >> Calendar->set_timezone() does not seem to apply to unix_time() at >> present. Is this intentional? The two operations below should IMO >> return the same number: > >unix_time() and time() returns the number of seconds since 1 Jan 1970 >0:00 UTC (when UTC was introduced and GMT was dropped, BTW). >That number doesn't change with timezones. > > > / Brevbäraren >

1 0

Re: Proposal for new ADT interface
by Martin Nilsson (Fake Build Master) ＠ Pike (-) developers forum 02 Nov '02

02 Nov '02

1 0

#X
by Johan Sundstr�m (�rkehertig av Lysators rootgrupp) ＠ Pike (-) developers forum 02 Nov '02

02 Nov '02

The extra space generated by this is a bug, I presume? #define FOO(X) #X #define BAR(Y) FOO(Y) int main() { write("%O\n", BAR(baz)); } (The output from the above is "baz " in 7.3, at least.)

3 8

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

Pike-devel