• Home
  • Script library
  • AltME Archive
  • Mailing list
  • Articles Index
  • Site search
 

World: r4wp

[Ann-Reply] Reply to Announce group

Ladislav
30-Mar-2013
[1732]
, i.e. if you really get identical builds when getting identical 
system/version
Kaj
30-Mar-2013
[1733]
What I'm doing now is just a Linux build. A build on Syllable Desktop 
yields a different host executable: that is not compatible. The library 
would also be internatlly different when compiled on Syllable Desktop, 
but Desktop can still load a library compiled on Linux
Ladislav
30-Mar-2013
[1734x2]
I am especially curious whether any of your 0.4.4 is really what 
I get when building r3 on Linux?
(which is what is intended to be the true 0.4.4)
Kaj
30-Mar-2013
[1736]
Builds are not byte for byte compatible across Linux platforms, because 
they depend on compilers, headers and startup fragments from the 
compiler and possibly the assembler
Ladislav
30-Mar-2013
[1737]
I supposed that was why there were other platforms for Linux than 
just 0.4.4?
Kaj
30-Mar-2013
[1738]
Those are for different versions of the GNU GLibC C library. In practice, 
it's impossible to compile binaries with the GNU toolchain that are 
compatible across all C libraries
Ladislav
30-Mar-2013
[1739x2]
Do you really want Syllable to be ignored as a platform?
I suppose it may have some unwanted side effects.
Andreas
30-Mar-2013
[1741]
AFAIU, 0.4.x for Syllable Server sounds appropriate (it seems to 
be "just" a Linux distribution).
Kaj
30-Mar-2013
[1742x2]
I've tested it quite well. We've been running R3 Linux binaries for 
years on Syllable Desktop. Having the source it's properly built 
on Syllable, but it hasn't fixed any problems
It's fine if the Syllable Desktop configuration becomes explicit, 
but I currently don't have a reason to spend time on it
Ladislav
30-Mar-2013
[1744]
Yes, but there already are some Rebol scripts examining system/version 
to "know how to call system libraries" or some such. I guess it may 
be a good reason to define a platform?
Kaj
30-Mar-2013
[1745x2]
Yes, that would be a good reason. In my patches to port Python, for 
example, I solve it by looking if the /system/index/ directory exists, 
which would identify not only Syllable Desktop, but also Syllable 
Server
As long as programs use the Linux configuration as default in a decision, 
it almost always also works on Syllable Desktop
Ladislav
30-Mar-2013
[1747]
well, then it may be just a matter of preference (if you want to 
make Syllable "visible" at least for Rebol)
Kaj
30-Mar-2013
[1748x3]
Again, it would be nice if just a copy of the Linux configuration 
were made under the name "Syllable", but currently I wouldn't want 
to ask anyone to spend time on it
I see I penciled in a TO_SYLLABLE parameter here in the Syllable 
overlay of the build recipe:
http://syllable.cvs.sourceforge.net/viewvc/syllable/syllable/system/apps/utils/Builder/packages/Syllable/REBOL-Core--current/
Arnold
31-Mar-2013
[1751]
Carl has given a sign of life on rebol.com rebolution blog.
Bo
1-Apr-2013
[1752]
I saw Carl for two minutes on Thursday.  He said he was going to 
poke his head in for a second, although he is still really busy.
Henrik
1-Apr-2013
[1753]
He has an account here, but has never logged in.
Bo
3-Apr-2013
[1754]
Kaj: Sorry, I should have posted that reply here.
james_nak
5-Apr-2013
[1755]
Kaj, thanks for all your hard work. When I click on http://red.esperconsultancy.nl/index.red
I see the actual code. Is it suppose to behave that way?
Kaj
5-Apr-2013
[1756x2]
If you are using a web browser, yes. If you want to run it, you should 
run GTK-browser
Same way the R2 Viewtop works
james_nak
5-Apr-2013
[1758]
I see. Thanks. I also tried to get the android version to run on 
my Nexus 7.No luck there so I went back to trying the test that was 
originally posted and I can't even get that to run. Sorry, you'll 
need a more experienced test. :-(
Kaj
5-Apr-2013
[1759x3]
There have been failure reports before, so something low level may 
be wrong with current Red on Android
Did you get any response?
What do you mean by the originally posted test? The http://sidl.fr/red/hello
download from Doc's blog is gone
james_nak
5-Apr-2013
[1762]
Yes, that one.
Kaj
6-Apr-2013
[1763x2]
It's gone, so you wouldn't be able to download it
Did you get any response?
Gregg
6-Apr-2013
[1765]
Go Kaj Go!
Kaj
6-Apr-2013
[1766]
Sleep first
Gregg
8-Apr-2013
[1767]
INCLUDE: So, the output %.r file, without option, will now be molded 
flat? If so, what is the reasoning?
Ladislav
8-Apr-2013
[1768x2]
Just a request. If you do not want the default, you can use INCLUDE/ONLY 
and save it as you see fit.
Do you think it is a complication for you?
Gregg
8-Apr-2013
[1770x4]
OK, good. Next question. Why not make the new behavior the option?


I only use the linked files directly on rare occasions. I'm just 
wondering what the benefit is. And, at first glance, INCLUDE/FLAT 
seems clearer.
I don't think it will complicate things much for me.
It won't be like CALL/SHOW breaking all kinds of stuff. :-)
So, what is the benefit of molding flat in this case?
Ladislav
8-Apr-2013
[1774x2]
The original demand was to have a /FLAT version which I offered to 
add, but Robert asked me to modify /LINK behaviour instead. I do 
not mind, in fact.
...but the fact is that adding another refinement would complicate 
the interface a bit without being absolutely necessary, perhaps
Gregg
8-Apr-2013
[1776]
I'm sure it will be fine. /ONLY has a different meaning everywhere 
else, and that meaning could be assumed here, so people just need 
to be aware.
Ladislav
8-Apr-2013
[1777]
I guess it was Carl who coined /ONLY for INCLUDE.
Gregg
8-Apr-2013
[1778]
Well, as I said, this is my first impression. :-)
GrahamC
8-Apr-2013
[1779]
@Kaj, that will be a test of rebolbot's error handling :)
Endo
10-Apr-2013
[1780]
AdrianS: Thank you for sharing RebolBot, it is a good example script 
of R3 as well.
AdrianS
10-Apr-2013
[1781]
Well, thank Graham - he did all the work. All I did was rip it apart! 
:-)