AltME groups: search

DocKimbel, alright. I take it, the somewhat equivalent in REBOL would 
be to encap a REBOL/Core? That would of course be much bigger.

It depends how you define "unneeded parts". It's more complicated 
to define that it seems at first look. For example, you might want 
to drop date! type from runtime library if your code is not using 
it...but if you LOAD some external data that might contain some date! 
values, you'll have a problem.


Anyway, we'll provide options for stripping from end binaries everything 
that can be stripped without altering program behavior.

ok, I was just thinking in terms of, say, not needing networking 
or FTP protocols for a very specific app. I don't think it would 
be a good idea to eliminate datatypes, although perhaps it makes 
sense in some cases.

Henrik: networking protocols at mezz level shouldn't take much space. 
In compressed souce form, each should weight only a few KB.

as I think about it, it would be best to have some kind of interpreter/library, 
which would download just once, and then only the app package would 
be downloaded. But that's just most probably prohibited by mobile 
platforms. Having just each "hello world" app to take 60KB etc. is 
a bit too much ....

I thought, that once compiled = you know what is resolving to what

 Certainly not possible with a REBOL-like language, as it needs only 
 one "unresolvable" expression to force you to attach whole runtime 
 library.


Don't expect any Red app to fit in 5KB or anywhere close, that would 
be magic, not computer science. Higher-level abstractions have a 
cost, if you don't want to pay for them, there's Red/System.


There's also no "magic" in compilation, some expressions can be reduced 
to very close to CPU counterparts, most can't. Don't expect the first 
Red alpha bootstrap compiler to do extensive optimizations (it does 
none currently). Also, a sentence you will probably hear from me 
very often: "Don't expect a v0.3.0 to be a v1.0.0!".

Having just each 

hello world" app to take 60KB etc. is a bit too much ...." You're 
not having the right metrics in mind. What matters is not the size 
of a "Hello World" app, but size of a real app. A typical real app, 
like with REBOL + /Encap, will be just a few dozens KB bigger than 
the runtime library. So, most Red apps will be less than 500KB and 
still less than 1MB for really big ones (not counting additional 
resources data). A "Hello World" using Appcelerator (one of the leader 
in Android dev tools market) is around 600KB. Also have a look at 
the size of apps installed on your Android devices, most of them 
are bigger than the typical apps size Red will produce.


If you want to compare with app size that Java can produce, take 
Red/System in order to compare apples to apples. Anyway, Java apps 
are cheated, because the JVM is built in Android. I don't remember 
seeing any way to install a shared library across different apps 
on Android, but maybe there's a way?

Any app that can convert string/binary data into internal datatypes 
like blocks and such will pretty much need the whole datatype collection 
and their associated actions. That's most of a Rebol-like language 
right there. Given how rare Rebol-like apps are without those features, 
we are better off making the runtime efficient rather than trying 
to figure out how to break it up.

For more open platforms, you'll be able to compile your Red apps 
to a binary with externalized runtime library (in form of a shared 
library pre-installed or to install with the app).

But thinking about it, those mobile platform policies make it really 
innefficient. I know that most app you download, are just megabytes, 
but still - why not to allow a runtimes to find their shared place 
on the device? Then, all your apps would need just kilobytes of the 
traffic, and mobile operators would be happier. But maybe it is not 
easily technically solvable on such OSes, who knows ....

Graham - do you regard having REBOL/Core interpreter of certain version 
available being a DLL hell? We are talking about just ONE shared 
lib, a runtime for certain language. Anyway - the discussion is most 
probably meaningless, for Red architecture itself, and even more 
so for the mobile platforms, well, just in my opinion ...

Pekr, runtimes interpreters are allowed on iOS now, as long as they 
are approved by Apple. What is prohibited is running external scripts 
or scripts downloaded from the internet. All interpreted scripts 
need to be bundled with the app, so the app can be evaluated as a 
whole by the app store gatekeepers. I've given this some thought, 
and it may be possible to cripple an R3-like interpreter sufficiently 
to meet these criteria; all you need to do is put some limits into 
the LOAD and DO mezzanine code. Red could be similarly crippled if 
you want.

As I said before, I'll announce it officially. I still have a couple 
of things to do before: decide about one bugtracker ticket and write 
the blog entry to announce it. :-)

I've cleaned up the open tickets today and created a wish list page 
in the project's wiki on github.

Freddy reports a number of strange problems with the MSDOS and Darwin 
versions of the test binaries. I'll relay more when I get more info 
from him

HI Kaj, I download Your Gtk Bindings and and GLib too, put it in 
red-system directory (it's a write place ?) and I have too problems:

You need the Red-common repository, too, installed in a subdirectory 
called common/. Same pattern as GTK and GLib. See the Announce channel 
a few announcements back

You're confusing it with a Red/System file called common. They have 
nothing in common ;-)

Alvira thinks there's a Trojan in Red/Hello.exe and Red/Fibonacci.exe 
( TR/Crypt.XPACK.Gen )

Antivirus alert: I've scanned Hello.exe with virustotal.com, no virus 
signature detected at all:

https://www.virustotal.com/file/ecde1916047993067be692235fb87352860b4596f5a2d54ba7efbf7482bcbe3a/analysis/1351329447/

Fibonacci.exe does raise some alerts from a few AV:

https://www.virustotal.com/file/6a86ef360727e907b6fc8cd7f9c0607a3ccb11f8f953747167b349dc43d6bf8a/analysis/1351329779/


I will look into these false positives signatures to see if I can 
improve that. Not sure it will be possible if they are caught by 
undocumented heuristics...

I've sent a few info requests to various AV vendors, I hope to get 
some clues about what's triggering their heuristics.

I will release v0.3.0 in a couple of hours, if you want to do some 
tests to track regressions, that would help.

I've sent an email to Freddy already asking a few more questions 
about his issues.

Thanks! I have added a mention about missing Chinese font on some 
of the screenshots, just in case someone thinks it's a bug.

Is there currently a difference between MSDOS and Windows targets?

So far, it doesn't make a big difference in practice.

But Windows seems to be quite permissive in recent versions, so a 
GUI app can also be started from DOS console.

Calling it something like Windows-Console rather than MSDOS might 
be better, since the real MSDOS (or rather clones of it) is still 
a valid target for embedded development. The Windows console environment 
hasn't been MSDOS since the '90s.

Brian: I'm aware of that. The probabilty of someone porting Red to 
old MSDOS (no Unicode, no multitasking, no native TCP/IP) is very 
close to zero. If someone does it anyway, we'll adjust our targets 
ID accordingly. In the meantime, I prefer typing "-t MSDOS" rather 
than "-t Windows-Console" on command-line. Also, it's easier to remember 
for everyone, after all it's just an ID, nothing else.


If you are thinking about FreeDOS, which is probably a more likely 
target than real old MSDOS, I guess we won't have any name collision 
then. ;-)

Call it DOS. But the MSDOS target name still annoys me. I don't want 
a legitimate, common target build to bring back such bad memories 
:(

Actually, not all of the memories were bad. In recent years I was 
thinking of making sure R3 would run on a set of libraries that enables 
Windows apps on DOS, but at the moment I'm having trouble finding 
the link/name of those libraries.

what about calling it CMD (as is the command, cmd.exe, which you 
have to issue to open that black old console), or CON, as a "console"?

Pekr: Red is not a Windows-only product. ;-)

Brian: as I've explained to Kaj earlier, Windows does make a difference 
between apps built with GUI or built for console- only. If it wasn't 
the case, we would only have the "Windows" target.

It's a cyber-party and it's happening here. ;-)

Pekr: I think the Twitter widget I'm using on my Blogger's site is 
dead, not a single other Twitter plugin for Blogger is working...

They changed their API a few weeks back

Doc, are you thinking of doing anything in particular to make the 
lexer be easily accessed from a hosting environment? It would be 
nice to see Red supported in a deeper way by something like Sublime 
Text.

DanielN: thanks for testing, I will have a look at the GTK issue 
later today.

Kaj: I got a reply from AVIRA telling me that the binary I've submitted 
was clean:

https://analysis.avira.com/en/status?uniqueid=KwPWqW429CmT1fNpbHWQgDxQ8ryDHO4T&incidentid=1301128

WebKit has been changing their API a lot, so the binding is only 
for the older version. Segfaults and such would be expected when 
fooling it. Also, the GTK binding is developed for GTK 2 at the moment. 
I'll update them when I start running newer versions myself

Only if Red does stuff that will trip the heuristics, or be run on 
a system with AVG (since they require you to pay them money to revert 
a false positive diagnosis). Which we can't know ahead of time because 
those heruistics are unpublished and thus untrustable.

It was a while ago, so pardon any mixups with the version.

MS has a really great false positive dispute process. I've never 
had a problem with them. Even in cases where the diagnosis isn't 
false as much as it is iffy, such as diagnostic tools that admins 
need to use but which you don't want non-admins to use, it just labels 
tham as "Medium Risk" and lets you choose to allow them to run if 
you like on a case-by-case basis.

The new Windows Defenfer in Win8 is pretty much a relabeled Microsoft 
Security Essentials with a few more features. Haven't had a problem 
with MSE yet that wasn't resolvable.

I've just got an answer from F-Prot, they just whitelisted the sample 
binary I've sent to them but didn't send me any info about their 
heuristics... So I think that in order to avoid loosing my time trying 
to get any info from those AV vendors, I'll just add a specific signature 
to Red generated binaries, so that they can be whitelisted by all 
AV vendors (when possible). I can't see what else I could do, except 
warn users about some crappy AV software.

So, I could just add a static c-string!, what good unique signature 
sequence could we use, and how long should it be?

Be sure to not include such a signature in Red's output of user binaries. 
We want to keep Red itself whitelisted even if Red's users decide 
to make malware with it.

OTOH, that could be a good selling point for Red. ;-)

Afaik, authenticode is supposed to handle that process, at least 
to a certain extent. You cryptographically sign your binaries so 
when they do bad things there's someone to blame. If your binaries 
aren't signed, they're assumed to be bad by some people.

That's signing the complete binary, that's much stronger than adding 
a signature

Right, it's definitely better than a whitelist constant.

All these anti-virus program makers tell you something you already 
knew: your software does not qualify as a virus or did not contain 
any known viruses.

Or you are 1 of the worst script kiddies. Or you managed to disguise 
the virus relly well this time ;)

I see Ruud prepared a new batch for you :-)

I'm sitting next to a Windows 7 Ultimate machine to test the bindings, 
but it's unusable because it has been updating itself unasked for 
hours, on shutdown and on startup, and now won't start up anymore

An Internet connection for a Windows machine is close to a lethal 
injection

well, that's the first thing I do when I set up a Windows machine

I never ever had issue caused by Windows auto-updating. You guys 
don't like Windows so much, that you are seeing a ghosts :-)

Petr, this machine just became a ghost

Kaj: I get a lot of type casting warnings since the last commit on 
GTK-widget.reds, could you please check if they are legitimate or 
not?

Wich doesn't mean the problems are solved, but it's a start ;-\

Kaj: the changes I did today are generating a lot of warnings with 
GTK binding, I will have a look at it later to see if it's a bug 
or if the binding code needs some minor changes.

Natasja: seems that you'll have a brand new system soon. :-)

Probably the easiiest way would be to inculde the Red runtime and 
use red/unicode/load-utf8 to create a Red string and Red/Platform/print-ucs4 
to print it.

The other way is to convert the utf-8 c-string! to  UTF-16E integers 
on the fly  and feed them into llibc putwchar yourself. More work 
upfront but may be easier in the long term.


The code in red/runitme/platform/win32.reds is a pretty clear exmpale 
of how to do it but you wuld still need to write the UTF-8 to UTF16-LE 
on the fly conversion yourself. (That one is UCS-4 to UTF16LE).

Kaj: you can switch the Windows console to an UTF-8 compatible mode 
using _setmode():
http://msdn.microsoft.com/en-us/library/tw4k6df8.aspx


I haven't test it but it should work. Windows uses natively UTF-16LE, 
so you would probably have a speed penalty using that mode.

Question for everyone: I was thinking since a while to publish on 
red-lang.org the donations I get, including donator names (unless 
they want to remain anonymous). Is that ok? Are they best practices 
for doing that?

The other option is to do categories like kickstarter does and just 
indicate a number of donations in a certain range

The new casting warnings seem to assume that all structs are equal. 
Especially in GTK, many objects are both generic widgets and a specific 
class of widget. They need to be passed in different identities to 
different functions, even if those structs are only defined with 
the same dummy content

I don't see Kaj calling his binaries "MSDOS apps". It is just a target 
ID meant to be easy to remember, nothing else.

IMO, the only way to get a new good replacement ID for "MSDOS" is 
to change all targets ID and adopt a different naming convention 
(one that is easy to remember).

Gabriele: that was my first intent, but I gave up to avoid having 
to code and maintain a dynamic web page...Also, I would need to set 
up a workflow, so that donators would need to fill a form first before 
accessing the donation options. I wanted to keep it simple, but maybe 
we need that.

Click click click click, nonsense timeline, nonsense hashes, then 
you finally get to the point, where you can download a ZIP file, 
and it is named like a crap, subdir's name is weird too. I am really 
upset anytime I have to go to such an interface. It is not imo normal, 
in order to get a clue how to get a file, to twiddle around for a 
minute!

SQLite is a masterpiece. Fossil might be the masterpice under the 
hood too. But really - Git's user interface is much easier to grasp 
imo for an occassional user ...

Git doesn't have a web UI

I don't necessarily disagree, but the Fossil UI is a generic interface 
to the generic Fossil versioning database. It's just the first visible 
step in my project to create custom apps with underlying version 
control

Windows (whatever)/ARM support: I would love to see that, I just 
need to get access to such platform first....Anyone know if there's 
a virtualized option for that OS?

Kaj - e.g. do-sql errored out on not finding a library, ditto GTK 
demo ...

Pekr: maybe you would like to write a handy REBOL script that would 
download all the latest bindings from Fossil repos? That would be 
a useful tool.

Yeah, the Zip names seem to be weird on non-Unix systems. I'll have 
a look at that later

Glut triangle exe missing a dependency

champlain map browser - missing a library

mandelbrot - just crashes. In console I can see a message "Access 
denied"

I have a win8 release preview. Would this work?

hello-Unicode is a Red/System program, so the Unicode printing is 
straight from the source code, without awareness

OK, I just installed win 8 Pro 32 on one of my machines and I downloaded 
the v0.3.0-0 zip and ran the run-all.r file. Perhaps that's not the 
right thing to do since it printed "running lexer-test" and then 
proceeded to lock up my machine. I can't even ctrl-alt-del. Let me 
know how you would recommend me testing Red since I may be the first 
to have win 8 among us but the last to really use Red passed the 
original "let's see if I can see 'hello world' on my Droid" a few 
months back :-)

It's probably a CALL issue, we had similar issues a year ago, and 
it was CALL-related IIRC. I and Peter need to install a Win8 system 
to be able to debug that.

Doc: I wonder if there's a web service out there that lets you take 
donations and shows a page like the humble bundle one. if not, perhaps 
we should make one. :-)

I remember searching for such service that would allow outsourcing 
the donations handling a few months ago, but couldn't find anything. 
That would certainly be a good idea! Also if it could handle the 
money transfer aswell, that could be a good ground for a new startup. 
;-)

Pekr: well, if I had billions, I probably would spend them into science 
research rather than that. One of the main science project would 
be to build me a light-saber. ;-)

I'm very glad to announce that I've received the biggest amount of 
monthly donations in October: 997.65€ received (~$1277) from 8 donators. 
That not only easily covers my monthly bills, but also allows me 
to save some money for future trips to devcons. So thanks a lot to 
all donators!

A special huge thank to Bas de Lange for making a big donation!

I'll have a look at it tonight.

In R3, a symbol is the internal unique string that a word refers 
to. A word is a pointer to a symbol and another to its binding. Same 
in Red?

That can deal with the series! type, unless you have an internal 
reason to have it be a real type.

Also, the ipv6! type might not be derivated from binary! if it can 
fit in a value slot.

I don't see how it could fit in a 128bit value slot, but maybe in 
a larger one.

Can't fit a value slot, so it needs it's own buffer.

Side effect of compilation. The compiled code retrieves a value from 
a value slot or other indirect type, manipulates the values in typed 
stack or register variables, then puts the results back in the value 
slot or wherever.

The issue! type was changed from a string-like type in R2 to a word-like 
type in R3, but the R3 behavior isn't completely final. It will continue 
to be a word-like type, but the syntax might get some tweaking and 
some string-like operations might be added back where possible, perhaps 
in a similar way to how tuples are series/like at times but actually 
immutable.

Jerry: I haven't decided yet for issue! datatype. By default, I stick 
to R2 model. For scalar!, Red definition might differ a bit from 
R3 one, we'll adjust that if required when the work on date! will 
begin.