r3wp [groups: 83 posts: 189283]
  • Home
  • Script library
  • AltME Archive
  • Mailing list
  • Articles Index
  • Site search
 

World: r3wp

[!REBOL3-OLD1]

RobertS
19-Nov-2008
[8201x2]
TortoiseSVN on Windows XP has been a nightmare: the more you checkou, 
the slower XP Explorer runs until you uninstall.
RapidSVN has worked flawlessly for anything that was not convenient 
on a command line
Henrik
19-Nov-2008
[8203]
TortoiseSVN: Good name then. :-)
BrianH
19-Nov-2008
[8204x5]
There are many advantages to using a distributed vcs, all variants 
on a theme. For example:

- You can track changes within your own code without submitting it 
until it is ready.

- It is easier to collaborate with other people without having to 
put your changes into the central repository.

This makes it easier to make large changes with the help of other 
people, and is why I suggested Mercurial over CVS.


The problem with all of this is that the main purpose of DevBase 
isn't source control, it is managing development workflow. No other 
vcs handles the workflow management function of DevBase. Without 
the workflow management Carl can't manage contributions.
Yesterday I came up with a way to change the DevBase model so that 
it would keep (and enhance) the workflow management functions and 
still get most of the benefits of a distributed vcs. Yes, it would 
depend on making the DevBase client a reblet run by the browser above, 
but that would be a bonus for R3 anyways.
Pekr, the navigational functions of web browsers are what makes the 
web work so well. You shouldn't misjudge them by the way that some 
badly designed web apps don't take them into account. We will still 
have the ability to create regular aplications in R3 when the browser 
model doesn't work for you.
Steeve, the reasons for minimizing the parse proposals are given 
in the proposal doc, but they boil down to these:

- Clean, simple and powerful is not only the REBOL way, but more 
efficient at runtime. It's the greatest strength of PARSE.

- PARSE is only one dialect of many in REBOL, and they are designed 
to be used together.
If anyone wants to discuss the source control and workflow management 
of the R3 project, mention it in the !DevBase group here.
Pekr
19-Nov-2008
[8209]
BrianH: so which of parse proposals will make it into?
BrianH
19-Nov-2008
[8210]
Don't know yet. I know my preferences - should I put an Editor's 
Note in the Priorities section?
Pekr
19-Nov-2008
[8211]
Don't know - it makes sense, only if Carl takes the list of priorities 
seriously. Maybe he has his own opinion and will choose to implement 
according to his own priorities ...
BrianH
19-Nov-2008
[8212]
I haven't edited the priorities section at all - I figured that was 
Carl's job. I should take a look at it. The trouble is, I don't know 
whether Carl has already started work or whether he has put it on 
hold in order to get a developers' release out ASAP.
Pekr
19-Nov-2008
[8213x3]
Henrik claimed Carl is working on parser for one week or so :-) But 
maybe he just guessed?
I would prefer public relese, and continuous, incremental releases. 
But - can it happen without CVS?
Is there any conclusion to it? Will it be Mercurial? Hg? DevBase?
BrianH
19-Nov-2008
[8216]
No conclusion yet, discussion continues in the DevBase groups.
Graham
19-Nov-2008
[8217]
Mercurial = Hg
Edgar
19-Nov-2008
[8218]
If you don't want TortoiseXXX to slow down your explorer, turn off 
overlay icons. That is what slows them down. If you decide to turn 
them on, set the settings so it is only on for specific folders where 
you need the overlay icons.
Graham
19-Nov-2008
[8219]
New bug tracker is now online http://curecode.org/rebol3/
Henrik
20-Nov-2008
[8220x2]
No screenshots for a bit. My server is going down for a few days.
(office is being repainted)
Graham
20-Nov-2008
[8222x2]
Carl wants to stick to devbase instead of using Hg
and he can't decide on whether to use a BBS or not.
BrianH
20-Nov-2008
[8224x4]
We are using a BBS, just not a web-based one.
I mean now. We don't know yet whether we will be switching from the 
AltME BBS to another one.
The problem with AltME is not that it is not a BBS, but that it isn't 
a very good one for our purposes.
BBS's predate the web, so you can't just assume that a BBS would 
be web-based.
Graham
20-Nov-2008
[8228x3]
I recall zmodem protocols :)
acoustic coupled modems
good old fido net
BrianH
20-Nov-2008
[8231]
Those were the days :)
Pekr
21-Nov-2008
[8232]
Nice - screenshots there - http://www.rebol.net/wiki/GUI_Panels- 
they are nicely self-explanatory addition to source codes ...
[unknown: 5]
21-Nov-2008
[8233]
I have deleted the REBOL Developer's Networking Group on Linkedin 
due to lack of participation.
AdrianS
21-Nov-2008
[8234x2]
Has anyone involved with VID3.4 development looked at MigLayout, 
the Java layout manager (http://www.miglaout.com), for inspiration? 
I've recently started using MigLayout with Groovy and this manager 
is great for intuitive programmatic control placement. I'm guessing 
that whomever is doing all the hard work in R3 has really thought 
about the problem and will provide smilar functionality and ease 
of use, but just in case, it might be worth a look.
The Java WebStart Swing demo on the page allows you to see the source 
as well as the individual constraints by right-clicking on a control.
Steeve
21-Nov-2008
[8236]
are u sure for the link ?
Graham
21-Nov-2008
[8237]
add a 'y"
Steeve
21-Nov-2008
[8238]
looking for screenshots...
AdrianS
21-Nov-2008
[8239]
I don't think there are screenshots - the live swing demo app is 
probably the best way to see/experience how the manager works - you 
don't need to install anything (you do need a Java VM, though)
Henrik
26-Nov-2008
[8240]
Short VID3.4 status:

- Some extra formatting options in the DOC style.

- I added a numeric field today similar to the one on my old blog 
site for VID, except that this one took half a day to do instead 
of a week and was way easier to implement.

- Skin is changing appearance a bit, both to simplify and to extend 
its possibilities. This made it possible to build pixel-perfect text 
areas, lists and other combined faces, finally.

- Some bugs remain in the layout engine which make certain layouts 
look a little messy. Still not sure that MAX-SIZE is the right approach.

- Added CONTAINER to help produce aforementioned pixel perfect styles. 
As a bonus, such styles will have a common border, which we then 
can change in a simple way.

- Looked a bit at a CALENDAR style for a date picker, but decided 
that it is better to wait a bit with that to finish some other base 
styles first.
Rod
26-Nov-2008
[8241]
Any word on the updated alpha release timing yet?
Graham
26-Nov-2008
[8242]
I think we have moved into a time dilation field
PeterWood
26-Nov-2008
[8243]
I believe the American holiday season starts today, so I'd guess 
we're looking at next year now.
Pekr
27-Nov-2008
[8244x5]
next year? That would definitely suck. I want something for Christmass 
to play during the long holiday nights ....
Henrik - were there any draw fixes done? Or Cyphre still not being 
involved?
Henrik - what was the problem of former skin, that it did not allowed 
us to have pixel-perfect text areas, etc.?
btw - thanks for update, I was currious, what is happening in GUI 
field, as r3-alpha is mainly discussing supporting infrastructure 
nowadays ....
Henrik - re MAX-SIZE - what was the technique Gabriele used in his 
VID3? Is the model different?
Henrik
27-Nov-2008
[8249]
- No time table given, so I can't say when there will be a release. 
I can only say that we're moving forward all the time, fixing bugs, 
adding new styles.

- No DRAW fixes yet. I haven't seen Cyphre since 5th November. We 
can work around most DRAW bugs for now.

- Skin lacked a CONTAINER style to put things in. Lining up scrollers 
with text fields, lists, etc. proved to be problematic. When one 
looked OK, the other would look terrible. We had TIGHT, GROUP and 
PANEL, but none of those are useful as decorative containers. CONTAINER 
works llike the frame of a painting, and now all styles that use 
CONTAINER look almost identical and pixel perfect. It also allows 
me to build complex field styles, like a search field with icons 
or numeric field with arrows.

- MAX-SIZE: It's a combination of weighting and maximum size of a 
face. When you set MAX-SIZE, it will affect the layout of the face 
in code, even if you are nowhere near maximum size, something I would 
never expect such an attribute to do. Setting MAX-SIZE correctly 
is an annoying process of trial and error, can't be predicted and 
depends on the MAX-SIZE of other faces in the panel, which requires 
debugging and studying other styles or setting them manually with 
a lot of typing. There are also magic numbers, so sometimes you use 
a size of 4000 and other times 100000 to achieve similar results. 
This is really the only part of VID3.4 that I don't like. However: 
There are still bugs, so intended behavior may turn out to be much 
better than I think. I can't remember what VID3 uses, but I would 
guess it's a more traditional weighting model.
Pekr
27-Nov-2008
[8250]
Why there is another style needed, to properly align elements? Why 
isn't panel and group enough?