REBOL 3.0

REBOL 3.0 Front Line

Carl Carl Sassenrath, CTO
REBOL Technologies
18-Sep-2014 21:42 GMT

2000000 visits since 15-March-2007
RSS Feed URL
Send me feedback

Search:

Purpose:
This blog provides advanced technical notes and discussions regarding the design and implementation of the new REBOL 3.0 system. It is intended for experts.

Also Visit:
Carl's REBOL Blog
REBOL 3 Home Page
R3 CureCode

Recent Comments::
0184 0351 0352 0345 0344 0340 0339 0350
Review all comments for week or month

Recent Articles:

20-Feb-2011 - Relative speeds from compiler optimizations [0352] 13 Cmts
12-Feb-2011 - Is Git working ok for R3? [0351] 11 Cmts
12-Nov-2010 - The SPEC-OF an Object [0350] 5 Cmts
28-Oct-2010 - Host-kit Source Control [0349] 10 Cmts
28-Oct-2010 - Idioms Wanted [0348] 0 Cmts
23-Oct-2010 - Remember help lib, sys, and self [0347] 10 Cmts
23-Oct-2010 - A109 Host Kit Status [0346] 1 Cmts
23-Oct-2010 - Amiga A109 Uploaded [0345] 12 Cmts
22-Oct-2010 - Example of a minimal module [0344] 2 Cmts
20-Oct-2010 - A109 Released [0343] 5 Cmts
18-Oct-2010 - RFC: A108 LOAD/SAVE support changes [0342] 16 Cmts
13-Oct-2010 - About /local words [0341] 20 Cmts
10-Oct-2010 - So, do you want RESET? [0340] 18 Cmts
5-Oct-2010 - Battling the Module Monster, Again [0339] 13 Cmts
29-Sep-2010 - Rename ISSUE datatype? [0338] 16 Cmts
28-Sep-2010 - Boot Levels [0337] 28 Cmts
Contents- Index of all articles.

20-Feb-2011 - Relative speeds from compiler optimizations [0352]

As an experiment for A111, I compared various compiler optimizations using the simple technique shown in this line:

r3 --do "print [size? system/options/boot speed?]"

Here are the results for different optimizations:

optexe-sizeinterpnativememorydisk
none5195125855245993137
-Os36340455444132116164
-O141678471504098104169
-O245774483484112114171
-O356014482154208117168

You can see that the default -O2 used for releases produces a good blend of results.

Measurements taken on a mid-priced CPU: Intel i5-2500 CPU @ 3.30GHz (quad core.)

13 Comments


12-Feb-2011 - Is Git working ok for R3? [0351]

A few months back we set-up Git for the R3 Host Kit source code control. This was a suggestion from a REBOL user for tracking releases and changes. We picked Git for the reasons listed in the above link.

It has come time to release R3 A111, and I'm looking at Git for source submissions. On github, I see changes from Oldes. That's great, thanks Oldes. So, are there any other change submissions? If so, where might they be posted?

The purpose of github was to use a well-supported central source management service. But, if Oldes and I are the only github users, I'm not sure it's worth the effort. It's easier for me to use R3 devbase to fetch and post source code changes, and it works on every target platform, and I've got the full power of R3 scripting behind it. Perhaps that's what I should use, and someone else can become a "gateway" back and forth to github.

Also, I will admit that these days I'm very busy with special projects that I don't have time to follow various discussions and message threads. But, I check in with BrianH, Cyphre, Ladislav, RobertM, Henrik, Gregg, and DocKimbel on a regular basis. If you need to reach me, contacting one of them might be the best path.

Anyway... R3 A111 is ready for release. If I can find any other changes this weekend, I will merge them. Otherwise, A111 will be released as is, and we can merge other changes into A112.

11 Comments


12-Nov-2010 - The SPEC-OF an Object [0350]

In R3, functions have "specs". These are interface and operational specifications. For example, arguments passed to a function, help strings, return values, exception handling, trace/profile controls, etc.

It is intended for objects (hence modules) to also accept a spec. This is one way to provide help strings for an object, enforce typing of the object's fields, along with other features.

The problem is: what should be done with specs when a sub-object is made. If we keep the spec, it would be nice to make it a reference, not a copy.

Then, there's the issue of mold: would all sub-objects mold the spec back or not?

So, it's not well defined. It may need to wait for post 3.0.0.

5 Comments

View index of all articles...

REBOL 3.0
Updated 18-Sep-2014 - Edit - Copyright REBOL Technologies - REBOL.net