| # | User | Message | Date |
| 386 | Oldes | it must be from old times before union was available for charsets.. union is a little bit faster, but mainly it does not require function in memory for each xml-parser object. (there is also diffset function) | 20-Mar 17:06 |
| 385 | BrianH | It seems that union/case would make sense here. How do they compare in efficiency? | 20-Mar 13:52 |
| 384 | Oldes | I'm looking at r2-mezz-xml and why there is used function joinset: func [cset chars] [insert copy cset chars] instead of just 'union ? | 20-Mar 10:14 |
| 383 | BTiffin | CCITT is/was the world Telecommunications standards body. Now ITU? Out of France. They did ratify a bunch of compression standards; ulaw Alaw etc. T.3 and T.4 are the international standard for fax transmission and compression. Along with those there are hundreds more; T.5 to T.600 protocols. Probably a thousand more since I left the phone company. CCITT liked to "keep busy"; justified the fees all the Telco's have to pay. :) But to their credit, you can phone home from remote China to remote northern Canada and the signal works ... far more often than not. | 18-Mar 1:14 |
| 382 | Gabriele | CCITT is the fax compression, isn't it? (maybe I remember it wrong) if so, then I guess it's used a lot, if fax handling is what they use tiff for. | 17-Mar 8:58 |
| 381 | Oldes | The hard part is only to compress/decompress. If we would extend compress function to accept more types, like LZW, JPEG (to de/compress raw data), the rest we can do myself directly in Rebol. I'm not sure how often are used the other types of compressions used in tiff like the CCITT or PackBits | 17-Mar 8:40 |
| 380 | Giuseppe | It is ok libtiff for me. Mine is only an alternative if it is viable. | 16-Mar 22:14 |
| 379 | Graham | Why not libtiff ? | 16-Mar 21:59 |
| 378 | Giuseppe | http://www.imagemagick.org/script/license.php | 16-Mar 21:59 |
| 377 | Giuseppe | You should take a look a ImageMagik Licence. | 16-Mar 21:59 |
| 376 | Giuseppe | Pekr, tiff, expecially multipage compressend/uncompressed TIFF is used in most scanning machines connected via LAN. It is a format widely used and with many inner variations. | 16-Mar 21:53 |
| 375 | Oldes | It would be fine to have LZW in Rebol:) | 16-Mar 13:47 |
| 374 | Graham | LZW was used in both GIF and TIFF. | 16-Mar 7:11 |
| 373 | Graham | http://www.remotesensing.org/libtiff/misc.html ... libtiff is free and available on most platforms. I think the LZW compression algorithm is now out of copyright ( 20 years ). | 16-Mar 7:10 |
| 372 | Graham | Let's see ... scanning is done in TIFF, ocr uses TIFF primarily as the input format, and fax is always in TIFF. | 16-Mar 6:53 |
| 371 | Pekr | Even digital cameras are turning off from hungry TIFF introducing their own RAW compressed formats, with particular decoders. I thought that we are going to abstract via loaders/savers .... | 16-Mar 6:40 |
| 370 | Pekr | Who uses TIFF today, gee? | 16-Mar 6:39 |
| 369 | Pekr | I am not sure we should hardcode any other image formats in REBOL directly, just because one developer needs it, sorry ... | 16-Mar 6:39 |
| 368 | Carl | I want to have TIFF in 3.0 -- But, it depends entirely on if we can find code that has a compatible license. | 16-Mar 0:20 |
| 367 | Graham | release early, release often .. not every 2 years | 15-Mar 22:27 |
| 366 | BrianH | The REBOL 2.7.6 release made PLNews. | 15-Mar 22:24 |
| 365 | Graham | It's the most common image format with gif and jpg | 15-Mar 22:22 |
| 364 | Graham | but some way of handling faxes would be nice ... instead of relying on imagemagick | 15-Mar 22:18 |
| 363 | Graham | I know it's complex ... | 15-Mar 22:18 |
| 362 | Graham | Is there ever going to be support for tiff ? | 15-Mar 22:17 |
| 361 | Graham | Even better :) | 15-Mar 22:17 |
| 360 | Gabriele | i do hope to reuse a lot of my imap code in R3 :) | 15-Mar 22:16 |
| 359 | Graham | Nice. | 15-Mar 22:15 |
| 358 | Gabriele | we have a complete browser based imap client (that is, to say it differently, it's going to be as good as thunderbird at least, eventually. now it's very close) | 15-Mar 22:14 |
| 357 | Gabriele | i'm not using the standard imap:// - it's all done from scratch. | 15-Mar 22:13 |
| 356 | Gabriele | Qmail does not yet support POP (and it's still in debate whether it should) | 15-Mar 22:12 |
| 355 | Graham | I presume you're using imap so that qtask can find out which mail is new ? | 15-Mar 22:11 |
| 354 | Graham | Gabriele, are you using the standard imap implementation, or an enhanced version? | 15-Mar 22:11 |
| 353 | Carl | Will, I would assume there must be ssl libs in most OSes as default not. But, then I am always wrong on that (expecting an OS to provide the basic libs). | 15-Mar 22:09 |
| 352 | Carl | Well, it all comes down to testing. If there is some way to test for the problem, we can enable the ssl debugging and see if we can figure it out. I am not even close to an expert on SSL. That was Holger's expertise. | 15-Mar 22:08 |
| 351 | WillArp | http://www.smxrtos.com/rtos/tcpip/barracuda/sharkssl.htm ?? | 15-Mar 22:08 |
| 350 | Graham | ie. secure pop and secure smtp | 15-Mar 22:07 |
| 349 | Graham | well, it can using pop and smtp .. just not imap ? | 15-Mar 22:06 |
| 348 | Gabriele | this is not critical because Qtask can just set up stunnel and use that for gmail access... but of course if would be nice to have rebol able to connect to GMail :) | 15-Mar 22:06 |
| 347 | Gabriele | note, that if i tunnel the connection thru the openssl command, the same thing happens (that's why i was thinking that maybe it does not work with ssl but requires tls). with stunnel works fine. | 15-Mar 22:05 |
| 346 | Graham | ok | 15-Mar 22:04 |
| 345 | Gabriele | i need imap, not smtp. my code works with other ssl imap servers. | 15-Mar 22:04 |
| 344 | Graham | for smtp ... | 15-Mar 22:03 |
| 343 | Gabriele | what happens is this: i can connect to GMail, i get the server hello, I can send in the login command, but after that no more data is transmitted or received. | 15-Mar 22:03 |
| 342 | Graham | I used ssl to get to gmail. | 15-Mar 22:03 |
| 341 | Carl | The only other alternative for now is to get you a version with debug output and see where it fails. | 15-Mar 22:02 |
| 340 | Carl | SSL is more than 30000 lines of code. Now, in the REBOL domain, that is a lot. | 15-Mar 22:01 |
| 339 | Carl | Well, we need to come up with a nice lib interface to ssl... | 15-Mar 22:00 |
| 338 | Gabriele | (unless you want to try to debug it, that is ;) | 15-Mar 22:00 |
| 337 | Gabriele | ok. Qtask will have to go thru stunnel to get to GMail i guess :) | 15-Mar 22:00 |
| 336 | Carl | Probably because our SSL is quite old at this point. V2.3 | 15-Mar 21:59 |
| 335 | Carl | Checking... and it is not supported. | 15-Mar 21:58 |
| 334 | Gabriele | i'm asking because imap over ssl:// to GMail does not work, and i'm not sure if that's because i'm supposed to use tls:// and not ssl:// or if there's a bug in REBOL's ssl://. | 15-Mar 21:58 |
| 333 | Gabriele | i don't remember it ever working, but i don't think i tried before Command 2.5 or 2.6. | 15-Mar 21:56 |
| 332 | Carl | TLS, like other things, is configured via words. E.g. the algorithm specification. | 15-Mar 21:55 |
| 331 | Carl | Did it work before? | 15-Mar 21:54 |
| 330 | Gabriele | (this is on Wine) | 15-Mar 21:53 |
| 329 | Gabriele | >> p: open tls://motoko.rebol.it:993 >> set-modes p [secure: true] ** Script Error: Feature not available in this REBOL ** Where: halt-view ** Near: set-modes p [secure: true] | 15-Mar 21:52 |
| 328 | Gabriele | (tried this on linux a few days ago and got feature-na, and i remember getting the same on windows in the past) | 15-Mar 21:48 |
| 327 | Gabriele | tls://some-server:N | 15-Mar 21:47 |
| 326 | Gabriele | no, client, not server | 15-Mar 21:47 |
| 325 | Carl | Testing that never gained any interest, AFAIK. | 15-Mar 21:45 |
| 324 | Carl | Correct, disabled. | 15-Mar 21:45 |
| 323 | Carl | You are asking about TLS://:N, right? | 15-Mar 21:44 |
| 322 | Gabriele | (linux last time i checked, but i remember the same result on the other platforms) | 15-Mar 21:41 |
| 321 | Gabriele | i get feature na | 15-Mar 21:40 |
| 320 | Carl | Does it? | 15-Mar 21:40 |
| 319 | Gabriele | is tls:// supposed to work on /Command? | 15-Mar 21:40 |
| 318 | Carl | We need to figure the best method to do this for 3.0 as well. Supporting SSL directly ourselves (with our own source version) is a losing game. | 15-Mar 21:40 |
| 317 | Gabriele | Carl, while you're here... | 15-Mar 21:40 |
| 316 | Carl | If we want SSL on Core, you'll need to call a DLL. | 15-Mar 21:39 |
| 315 | Gabriele | i don't have auto login though. | 15-Mar 21:32 |
| 314 | Gabriele | i have different passwords (although they are the same for some of the worlds) | 15-Mar 21:32 |
| 313 | Ashley | Perhaps having the same password for all worlds avoids the problem? I do, and I've never encountered this bug. | 15-Mar 21:10 |
| 312 | Henrik | AltMe bug: It usually happens for me when I start AltME and it autologins to a world, and I press the home button to go back out again and select the other world. | 15-Mar 20:31 |
| 311 | Gabriele | i don't have a way to test esmtp specifically, so I can't say for sure, but they should be fixed. | 15-Mar 19:48 |
| 310 | WillArp | please upload Altme osx intel for testing 8) | 15-Mar 19:47 |
| 309 | Graham | I was under the impression SSL was also going to be freed at some stage?? | 15-Mar 19:46 |
| 308 | Graham | Which reminds me .. ssl is still only a /command feature .. so most users can not send/access gmail with Rebol/Core. | 15-Mar 19:45 |
| 307 | Graham | Not that I've had any ... but I do recall that there were some messages about it a while ago on the mailing list. | 15-Mar 19:42 |
| 306 | Graham | Gabriele .. are all the issues with esmtp now resolved? | 15-Mar 19:42 |
| 305 | Graham | I was just googling pricing .. haven't tested under vista | 15-Mar 19:40 |
| 304 | Carl | Doesn't the current Thawte authenticode cert work? | 15-Mar 19:39 |
| 303 | Graham | Hmm... one order of magnitude difference in prices | 15-Mar 19:35 |
| 302 | Graham | vs Verisign $1293 US for 3 years | 15-Mar 19:35 |
| 301 | Graham | for signing Vista apps | 15-Mar 19:34 |
| 300 | Graham | http://www.wintellect.com/cs/blogs/jrobbins/archive/2007/12/21/code-signing-it-s-cheaper-and-easier-than-you-thought.aspx 3 year certificate from Tucows for $195 | 15-Mar 19:34 |
| 299 | Graham | I'm sure I always click the same way | 15-Mar 19:30 |
| 298 | Gabriele | at least two it would seem ;) | 15-Mar 19:30 |
| 297 | Gabriele | no idea, i guess many ;) | 15-Mar 19:29 |
| 296 | Graham | How many ways are there to click on a drop down? | 15-Mar 19:29 |
| 295 | Gabriele | maybe it just depends on how you click on the drop down... it's strange that some of us have never seen it | 15-Mar 19:28 |
| 294 | Graham | Signed binaries ... is this expensive? | 15-Mar 19:27 |
| 293 | Graham | Well, noticeably annoying this last year. | 15-Mar 19:27 |
| 292 | Graham | Present for at least a couple of years | 15-Mar 19:27 |
| 291 | Gabriele | yes, currently, but I have used Windows for many years. is this a new bug? | 15-Mar 19:25 |
| 290 | Graham | Gabriele uses Wine ? | 15-Mar 19:24 |
| 289 | Carl | (copying those to the altme bug tracker - thanks for noting them) | 15-Mar 19:23 |
| 288 | BrianH | Yes, that is an AltME bug. | 15-Mar 19:21 |
| 287 | BrianH | I had to fix a bug in REBOL 3 that was caused by reacting incorrectly when R3 was called from AltMe with a "bad script". | 15-Mar 19:21 |
| 286 | Carl | Is that an AltME bug? | 15-Mar 19:21 |
| 285 | BrianH | I'm a bit more annoyed about the AltMe bug where it doesn't interpret the substitution variables in the command strings from the Windows registry right. That is why exes called by clicking on them in the file list get the string "%*" passed as a parameter, and it doesn't work to click on a zip file when you have the Windows compressed folders. | 15-Mar 19:20 |
| 284 | Gabriele | never happened to me either. | 15-Mar 19:19 |
| 283 | Carl | Ah, ok. Was not aware of that bug. | 15-Mar 19:13 |
| 282 | Graham | ( well, for those of us who have passwords of different lengths ! ) | 15-Mar 19:12 |
| 281 | Graham | If you use the drop down to select a world, and the password length does not change, you know you've got to type in the password | 15-Mar 19:12 |
| 280 | Graham | Btiffin, Paul, Sunanda, and Henrik have also reported it. | 15-Mar 19:10 |
| 279 | Graham | Quite a few of us experience this bug.. whereas Reichart says he's never seen it. | 15-Mar 19:08 |
| 278 | Graham | It's an intermittent bug. Some days it's fine ..other days .. drives us crazy. | 15-Mar 19:08 |
| 277 | Graham | And you can't login. | 15-Mar 19:07 |
| 276 | Graham | On Windows .. if you select another world quickly, the password does not change. | 15-Mar 19:07 |
| 275 | Carl | I've not typed in a password in... 3 years? | 15-Mar 19:06 |
| 274 | Carl | what bug? | 15-Mar 19:06 |
| 273 | Graham | very frustrating to keep losing the password | 15-Mar 19:06 |
| 272 | Graham | fixing this password bug would be good | 15-Mar 19:05 |
| 271 | Carl | Some new Altme test builds are expected. | 15-Mar 19:05 |
| 270 | Carl | I wonder if it makes them easier for vista usage. | 15-Mar 19:04 |
| 269 | Graham | Are you doing new builds for Altme? | 15-Mar 19:04 |
| 268 | Carl | BTW, should we sign the core and view exes? | 15-Mar 19:04 |
| 267 | Graham | Ahh.. thanks for the explanation. | 15-Mar 19:04 |
| 266 | Carl | Or, said another way, I don't know of any tool that will expand the icon color space of an exe. (Could be one out there, I just have not seen it.) | 15-Mar 19:03 |
| 265 | Carl | AFAIK, if you want to build an encap that needs a full quality XP icon, you need the exe to already be built that way. | 15-Mar 19:03 |
| 264 | Carl | The colorbar SDK icon is a full color-depth XP icon. The std REBOL icon is not. | 15-Mar 19:02 |
| 263 | Graham | It must have been an abberant version ! | 15-Mar 18:54 |
| 262 | Graham | Ahhh... encap 2.5.125 used the white R ... | 15-Mar 18:54 |
| 261 | Graham | enpro 1.1.2 has a similar one to now without the R. | 15-Mar 18:52 |
| 260 | Graham | Yeah .. before 2.6.2 | 15-Mar 18:48 |
| 259 | BrianH | I only have SDK 2.6.2 and 2.7.5 for Windows to compare to, and they have the same icon as 2.7.6. | 15-Mar 18:46 |
| 258 | Graham | In 2004 my encap apps used the old icon | 15-Mar 18:45 |
| 257 | BrianH | Nope, they used the same one. | 15-Mar 18:43 |
| 256 | Graham | except earlier sdk's used the white R icon | 15-Mar 18:41 |
| 255 | BrianH | That icon was only for SDK apps that haven't had their icon replaced. | 15-Mar 18:41 |
| 254 | Graham | 2.7.6 is now out for lInux .. great. | 15-Mar 18:40 |
| 253 | Graham | I was always under the impression that logo was for the betas/alphas | 15-Mar 18:40 |
| 252 | Graham | Really? And the tv color test bars ?? | 15-Mar 18:39 |
| 251 | Paul | I like the black R. | 15-Mar 11:16 |
| 250 | Graham | ie. the White R instead of the Black R | 15-Mar 6:10 |
| 249 | Graham | Can we change the encap icon back to the original REBOL "R" for 2.7.7 ? | 15-Mar 4:36 |
| 248 | Graham | I guess it's more for vid .. but right mouse clicks for field context menus, and access to the inking, and writing recognition. | 14-Mar 23:17 |
| 247 | Carl | I have a tablet here. What features? | 14-Mar 23:06 |
| 246 | Graham | This is more for R3 than R2 but what about supporting the unique features of Windows Tablet OS ? | 14-Mar 18:08 |
| 245 | BTiffin | Hear hear. Go Bri Go Long Live R2 hmm, D2 :) | 14-Mar 3:30 |
| 244 | BrianH | :) | 14-Mar 1:41 |
| 243 | Anton | Who doesn't think Brian is on fire at the moment ? | 14-Mar 1:41 |
| 242 | BrianH | It has had the added benefit of giving me another chance to evaluate the R3 stuff, which has led to improvements and fixes. | 13-Mar 22:17 |
| 241 | BrianH | I did the backporting so that people can get started on making their code forwards-compatible right away, and to act as a trailer for R3. | 13-Mar 22:16 |
| 240 | Anton | That backporting was cool. Now R2 and R3 are more similar. | 13-Mar 16:59 |
| 239 | Brent | Petr: Agreed that Carl's time spent on R2 can negatively affect R3's schedule, but backporting has been accomplished mostly by others like BrianH, I think. SHIFT (the only real enhancement to the core) was not a big time consumer. Carl's main efforts appear to have been on 1) some bug fixes, but also 2) on some important Linux and OSX issues that would have come up with R3 as well. That makes it not completely time wasted for R3. Also, one other thing to consider is if R3 came out today, it would be neat, but not have much immediate benefit, because it will take some time to learn, write new code, realize advantages, etc. Whereas with some new incremental versions of R2, there is immediate benefit to existing users. I would think these releases of R2 will help build userbase momentum going into R3's release. So IMO, don't delay R3 if not necessary, but some incremental R2 releases with easy bug fixes give a lot of bang for the buck and benefit R3 as well. | 13-Mar 16:34 |
| 238 | Paul | this should be documented better as it is a very useful feature. | 13-Mar 16:21 |
| 237 | Paul | It worked! Very cool. | 13-Mar 16:20 |
| 236 | Paul | I'll let you know it if works later. | 13-Mar 12:52 |
| 235 | Henrik | so you should be able to build this using contexts and 'self. | 13-Mar 12:07 |
| 234 | Paul | seems to be | 13-Mar 12:04 |
| 233 | Henrik | I made an error. It doesn't matter. It looks like the binding is correct, isn't it? | 13-Mar 12:03 |
| 232 | Paul | of what? | 13-Mar 12:03 |
| 231 | Henrik | what the context wrapped around | 13-Mar 12:02 |
| 230 | Henrik | ah, sorry, didn't see where the context wrapped around | 13-Mar 12:02 |
| 229 | Paul | I did it again and it did output the system object the second time. Not sure why it locked up REBOL the first time. I may have had 'd defined globally also in some way. | 13-Mar 12:02 |
| 228 | Henrik | >> a: context [self: context [b: 1 c: does [probe bind? 'b probe bind? 'd]] d: "rebol"]
>> a/c
make object! [
b: 1
c: func [][probe bind? 'b probe bind? 'd]
]
make object! [
d: "rebol"
] I thought 'a, 'b and 'c would be bound to the same object here. | 13-Mar 12:02 |
| 227 | Paul | No I had to kill the process. | 13-Mar 12:00 |
| 226 | Henrik | it isn't outputting the SYSTEM object? | 13-Mar 12:00 |
| 225 | Henrik | interesting | 13-Mar 12:00 |
| 224 | Henrik | if it delivers the correct value, then the binding is correct. | 13-Mar 11:59 |
| 223 | Paul | rut row - just tried that and it locked up REBOL. | 13-Mar 11:59 |
| 222 | Paul | I'm going to test it later today. | 13-Mar 11:58 |
| 221 | Henrik | you can check the binding of 'd inside the function with PROBE BIND? 'd | 13-Mar 11:58 |
| 220 | Henrik | I don't think the binding is affected, but don't quote me on that :-) | 13-Mar 11:58 |
| 219 | Paul | I'm ok with doing that - just concerned how that affects binding it into subfunctions and if they could break the REBOL does things with objects. | 13-Mar 11:54 |
| 218 | Paul | The the 'd can't be directl referenced. | 13-Mar 11:53 |
| 217 | Henrik | yes, something like that | 13-Mar 11:52 |
| 216 | Paul | a: context [self: context [b: 1 c: does [print d]] d: "rebol"] | 13-Mar 11:52 |
| 215 | Paul | Ingo, showed a trick on hiding that direct reference in the REBOL3 world. He did something similiar to: | 13-Mar 11:52 |
| 214 | Henrik | it might be possible if you put the block in a context inside the context and the functions related to that block inside that same context. | 13-Mar 11:51 |
| 213 | Paul | so in other words what we need is a way to make hide code in the object from being referenced directly. | 13-Mar 11:51 |
| 212 | Paul | I want to reference the block in other parts of the object. For example, I have a block called operations which has subblocks that get bound to others functions within the object but I don't want someone to reference the operations block directly. | 13-Mar 11:50 |
| 211 | Henrik | you mean, you can reference the block inside the object in other places than you want or you can reference the block from outside the object, which you don't want? | 13-Mar 11:47 |
| 210 | Paul | Actually, my example was a bad one. I have a scenario right now where I have a very large block defined in an object and bind subblocks from that large block into a function also defined in the object. However, someone can still reference the very large block which is what I don't want. | 13-Mar 11:43 |
| 209 | Henrik | and having USE return something usable for CONTEXT would break what it does now, namely run some code like DO, just with a few private words. | 13-Mar 9:35 |
| 208 | Henrik | and I don't know if you noticed:
a: context [a: 1 b: 2 set 'c 3] 'c is now public while 'a and 'b inside the context are not. | 13-Mar 9:25 |
| 207 | Henrik | does USE have any effect there? | 13-Mar 9:21 |
| 206 | Paul | Would be cool to do something like:
a: context use [a b c][a: 1 b: 2 c: 3] Do we have some functionality like that? | 13-Mar 2:38 |
| 205 | Paul | I think there was a discussion about doing that on the REBOL3 world not to long ago. | 13-Mar 2:26 |
| 204 | Paul | How significant of a change would it be to declare a variable in in the 'use block. Just as: use [a: [1 2 3][foreach item a [print item]] | 13-Mar 2:26 |
| 203 | Graham | And it is important to produce fixes given that R3 may not be production ready for a year. Nobody wants to alienate the existing customer base. | 13-Mar 2:03 |
| 202 | Graham | Personally I don't think we should be backporting .. just doing fixes. Unless the backporting is fairly trivial. | 13-Mar 2:02 |
| 201 | Henrik | and my customers would be quite upset if I did that. it's not wise betting on R3 for production environments now or in 2 months. perhaps a year from now. | 12-Mar 23:41 |
| 200 | Henrik | Useless, because I have large heaps of code that will not be R3 compatible unless I rewrite it. | 12-Mar 23:40 |
| 199 | Pekr | ... plus the obvious benefits .... | 12-Mar 23:39 |
| 198 | Pekr | Useless because of what? VID3 might be in good shape in 2 months, View adapted to Unicode in 2 days. Then we can move to protocols, and you are mostly on R2 level. | 12-Mar 23:39 |
| 197 | Pekr | So please, finish and release 2.7.6., and please let's move on ... | 12-Mar 23:38 |
| 196 | Henrik | well, whatever. I use R2 in production environments. R3 will be useless to me for a long time. | 12-Mar 23:38 |
| 195 | Pekr | I don't believe in "long time" as you say. With REBOL, we are used to use beta software for production for a long time .... | 12-Mar 23:36 |
| 194 | Pekr | e.g. we are backporting some functions from R3 - for whom and for what? We use R2 8 or more years without them. And in one year max we all will be using R3. I am strict on time spent on old stuff. If it would be my project, I would do necessary fixes with zero enhancements. All those requests are substracting time from R3. | 12-Mar 23:36 |
| 193 | Henrik | there's a long time before R3 is production ready. until then, we need R2 updates. | 12-Mar 23:34 |
| 192 | Pekr | As I said - it is not bad to have 2.7.6, nor to plan for 2.7.7., but we should remind ourselves, that our future is R3, not R2. | 12-Mar 23:32 |
| 191 | Pekr | Because it is always like that - 2 - 3 days, then ending in 2 - 3 months of constant work | 12-Mar 23:31 |
| 190 | Pekr | I think it will do. | 12-Mar 23:31 |
| 189 | Pekr | IMO radical freeze and release of 2.7.6 (with some plan for 2.7.7) is needed asap. | 12-Mar 23:31 |
| 188 | Henrik | I don't think 2 weeks will look that big in the overall schedule for R3 :-) | 12-Mar 23:31 |
| 187 | Pekr | I am not stating it is bad. I am glad, that guys got some bugfixes, some free stuff and enhancements after some time. But initial plan was few days. We have DevCond planned for September. And I would like to see R3 being in decent shape by that time. And really - R3 is the future. I don't understand all that sudden request! Those two wasted weeks could give us maybe finished modules and maybe R3 could start being decent enough sooner, to have host parts released to devbase and proceed further ... | 12-Mar 23:28 |
| 186 | Paul | But this is good - it stimulates us for 2.7.7 that while the rest of us look at ways to improvie 2.7.6 we can have Carl and others focused on R3 development. | 12-Mar 23:24 |
| 185 | Paul | And I agree Pekr, then get back onto R3. | 12-Mar 23:23 |
| 184 | Paul | I think a lot of good came out of it and we probably need to hit the street with something. | 12-Mar 23:23 |
| 183 | Pekr | This whole thing is starting to take away way too much time away from R3. We should rething R2's position. I know guys will not like it, but R2 is not the future anymore ... | 12-Mar 23:22 |
| 182 | Paul | Or I should say 2.7.7? | 12-Mar 23:19 |
| 181 | Paul | Can REBCODE be thrown into this release or is it to vastly different for that? | 12-Mar 23:19 |
| 180 | Carl | Oldes, if you are looking for REBOL to parse like ActionScript, there will be many other factors that will get in the way. REBOL's lexicon is quite complex, but its syntax is trivial. | 12-Mar 17:38 |
| 179 | Carl | (and add to that just a single "," with space on both sides) | 12-Mar 17:32 |
| 178 | Carl | Or, in just one line, parse this: [a, b,c 1, 1,2 1,2.3 %file,txt bob,smit@foo.com] | 12-Mar 17:31 |
| 177 | Carl | Oldes, this is a deep topic. I am not sure how much we want to get into it, but some of the issues are... 1. Adding comma is a lexical extension of REBOL. That is, we would recognize the "," char. 2. Do we allow it to be a semantic extension? (length? [a , b , c] = 3 or 5 ?) 3. Is it a strong delimiter? (E.g I write: http://www.rebol.com/page,here ?) 4. What to do about ambiguous cases (what is 1,2 and 1, 2 and just 1, )? 5. If it is a word, can it be within a word? (my,word -- is that 1 or 2 words?) | 12-Mar 17:29 |
| 176 | Gabriele | (don't forget that 1,2 is a decimal! in REBOL) | 12-Mar 16:41 |
| 175 | Gabriele | Oldes, if it's a rebol dialect, why would you need the comma? why not anything else? | 12-Mar 16:39 |
| 174 | Pekr | Oldes - does replace/all slows down your script so much? How large your scripts are? I think that two-pass aproach could work. And - even with string parsed dialect, you could still use REBOL like syntax, no? | 12-Mar 16:09 |
| 173 | BrianH | You can compile ActionScript to your internal dialect... | 12-Mar 16:08 |
| 172 | Oldes | Doing one pass just to convert commas I don't want as it would slow down a lot... I rather force myself to write extra parens in my dialect. | 12-Mar 16:06 |
| 171 | Oldes | I know, but if I start string based parsing, I don't have to do Rebol like language dialect and can parse for example directly ActionScript. I want Rebol like language but I found that I miss some delimiters if I don't do complex analysis which function requires which number of arguments etc. | 12-Mar 16:04 |
| 170 | BrianH | LOAD only works with dialect code (REBOL data syntax). With full language code (other language syntax) you need to do string parsing. | 12-Mar 15:52 |
| 169 | DocKimbel | You could do a first pass on your data replacing the comma with another valid character (or word), then LOAD it ? | 12-Mar 15:51 |
| 168 | Oldes | I need to be able to load the dialect. this doesn't help... I know I can use string based parsing, but that will be probably slower than native 'load | 12-Mar 15:49 |
| 167 | DocKimbel | Oldes: if you really need it, you may use a workaround : word! == type? to-word "," | 12-Mar 15:45 |
| 166 | Oldes | That's the reason why I want it to mature.. as an error it's useless:( | 12-Mar 15:39 |
| 165 | BrianH | It could be used as a delimiter, similar to space. We could make a comma! type that could be ignored or not by your dialect. | 12-Mar 15:37 |
| 164 | Oldes | Do you have any exact idea for what it can be used in future? | 12-Mar 15:35 |
| 163 | Oldes | And is it necessary? | 12-Mar 15:33 |
| 162 | Carl | It means nothing. It is being "held back" in case we want to use it to extend the language. | 12-Mar 15:28 |
| 161 | Oldes | what does it mean? | 12-Mar 15:27 |
| 160 | Carl | It is reserved. | 12-Mar 15:16 |
| 159 | Oldes | Is there still reason why 'comma cannot be a valid word? I would like to use it in block based dialect. >> load " , " ** Syntax Error: Invalid word -- , ** Near: (line 1) , >> load " . " == . | 12-Mar 15:03 |
| 158 | Carl | lol! | 11-Mar 20:52 |
| 157 | Graham | R3 is like a new drug .. ! But some of us have to wait to see the side effects, and wish to stick to the tried and faithful | 11-Mar 5:58 |
| 156 | Pekr | nice! What's next? Tasking? Modules? Initial plan was plugins, but those needed modules, and modules needed to count with Unicode. So logically now modules should be on the table? But tasking will not surely hurt too and Doc will welcome it for some test efforts of porting Uniserve .... | 11-Mar 5:45 |
| 155 | Carl | I actually have been doing some "multitasking" --- last night added logical flags. | 11-Mar 5:42 |
| 154 | Carl | I am so ready to return to R3 work. | 11-Mar 5:42 |
| 153 | Pekr | Police call: R2 guys will not like me, but I hope this whole effort will not exceed more than few additional days and we are back on R3, which is our future .... | 11-Mar 5:40 |
| 152 | Carl | That's REBOL's addition. | 10-Mar 23:17 |
| 151 | Oldes | in gzip is the size known, it's the last 4 bytes | 10-Mar 18:38 |
| 150 | Carl | That is, the decompressed data size is unknown without that information. | 10-Mar 17:43 |
| 149 | Carl | R adds the size so it can most efficiently handle the decompression blocks.... if I recall correctly. | 10-Mar 17:42 |
| 148 | Carl | I think there's one other thing... the blocking factor. | 10-Mar 17:42 |
| 147 | WillArp | hopefully Carl can add them | 10-Mar 16:58 |
| 146 | BrianH | Oldes, the difference between zlib, zip, gzip and REBOL compress is in headers and footers; the compressed data is the same. The main difference in the headers and footers is the checksum algorithm. If the necessary checksum algorithms were added to the CHECKSUM native, all of those formats could be implemented as a mezzanine using the existing COMPRESS and DECOMPRESS natives. | 10-Mar 16:28 |
| 145 | Oldes | Yes... there is.. uploaded test script into Tests | 10-Mar 15:20 |
| 144 | Pekr | good question, Giuseppe :-) | 10-Mar 14:56 |
| 143 | Giuseppe | But is there also sound in the new release of REBOL ? | 10-Mar 14:43 |
| 142 | Pekr | RT should use it for its marketing advantage, at least with recent developers :-) It is good preparation for what will R3 allow out-of-the-box ... | 10-Mar 14:40 |
| 141 | Pekr | no, not yet ... I know this is significant :-) | 10-Mar 14:39 |
| 140 | Henrik | this is quite significant. has it been announced publicly yet? | 10-Mar 13:53 |
| 139 | Pekr | I think too. It depends how is /library implemented for R3. I remember Carl considering wheter it should be added to R3 core, or done as one of plug-ins. The latter was more probable, IIRC. But plug-ins need modules, and modules needed Unicode. So I think we will now get modules, then plug-ins, in order to get /library component. Without it, there is no SQLite for R3. | 10-Mar 13:51 |
| 138 | Henrik | well, I think some of the work will make it easier to implement the same things for R3? | 10-Mar 13:49 |
| 137 | Pekr | it means free SQLite support, and maybe so guys will start to add some wrappers ... I fear it is too little too late, but a good starter nevertheless ... | 10-Mar 13:49 |
| 136 | Henrik | great | 10-Mar 13:48 |
| 135 | Pekr | yes ... | 10-Mar 13:48 |
| 134 | Henrik | as I understand it, we do get /library now? | 10-Mar 13:37 |
| 133 | Pekr | that is why I voted at least for /library to be released. I don't expect Carl putting https and ODBC into free Core, but maybe https should be there ... it is pretty common part of daily web-life, when you need to login somewhere ... | 10-Mar 13:15 |
| 132 | Henrik | between Pro and Command, I mean :-) | 10-Mar 12:55 |
| 131 | Henrik | ah yes, there is a difference | 10-Mar 12:54 |
| 130 | Paul | SSL is in /Command from what I remember. | 10-Mar 12:52 |
| 129 | Oldes | HTTPS is not even in PRO | 10-Mar 12:41 |
| 128 | Henrik | few people probably use https, because it has been a pro-only feature so far. an interesting side effect: pro features do not get as much testing exposure as non-pro features. | 10-Mar 12:06 |
| 127 | Pekr | Graham - https is very usefull imo. I once needed to automate download of e-invoices from our portal, and I needed https support definitely. Worked nicely. https should be part of Core, at least for R3 definitely ... | 10-Mar 11:43 |
| 126 | Graham | I suspect that no one else uses https! | 10-Mar 11:07 |
| 125 | Graham | Anyway, even though I checked with Curl .. I'm not sure if this is a bug, or a Cheyenne bug. | 10-Mar 11:07 |
| 124 | Graham | Gabriele, weren't you writing a new https prot ? | 10-Mar 11:04 |
| 123 | Graham | Why would I need https ? Because I was sending faxes to a soap service by https that would then fax them out for me. ie. provide my application with a desktop faxing ability. | 10-Mar 11:03 |
| 122 | Graham | I basically dropped using https because of that. | 10-Mar 11:01 |
| 121 | Graham | Also, Doc reports there that encmd crashes if 'title is used in the encap header. But enpro and enface were okay. | 10-Mar 11:01 |
| 120 | Graham | 28th June .. Cheyenne channel. | 10-Mar 11:00 |
| 119 | Graham | Reading what I wrote, I was sending up a file to cheyenne using https. 16376 is okay, 16377 bytes fails but it worked okay with Curl. | 10-Mar 10:59 |
| 118 | Graham | Let me check R3 world | 10-Mar 10:52 |
| 117 | Gabriele | SSL limit: i remember you talking about it... but do you have a test case to show the problem? is it SSL in general or HTTPS only? | 10-Mar 10:51 |
| 116 | Gabriele | Graham: Qtask runs on Command; I have verified it can be made to run on Pro but that requires loading some mezz files yourself. | 10-Mar 10:50 |
| 115 | Graham | I was posting large messages to Cheyenne and hit a limit with https before I hit a http limit .. maybe it was Cheyenne bug. | 10-Mar 10:50 |
| 114 | Graham | I'm hoping someone else will remember. | 10-Mar 10:49 |
| 113 | Pekr | So you ask us, what you reported back then? :-) We can't see to your head :-) But I do remember some related discussion. Can't remember its content though ... | 10-Mar 10:48 |
| 112 | Graham | Didn't I report some bug with https ... something about there being a limit that one can post ? | 10-Mar 10:45 |
| 111 | Graham | no odbc on linux | 10-Mar 10:43 |
| 110 | Pekr | fastCGI ... that is all that comes to my mind .... | 10-Mar 10:43 |
| 109 | Pekr | mySQL, Oracle .... | 10-Mar 10:43 |
| 108 | Pekr | SSL, ODBC? | 10-Mar 10:43 |
| 107 | Graham | Qtask uses ssl?? | 10-Mar 10:42 |
| 106 | Graham | What does Linux command add that pro doesn't? | 10-Mar 10:41 |
| 105 | Gabriele | Carl, if you get me a Linux Command build, i can test running Qtask on it. :) | 10-Mar 10:40 |
| 104 | Oldes | and some info http://www.websiteoptimization.com/speed/tweak/compress/ | 10-Mar 10:16 |
| 103 | Oldes | Here is something for test: print length? a: read http://www.webcompression.org ;== 3098 print length? b: read/custom http://www.webcompression.org [header [Accept-Encoding: "gzip,deflate"]] ;== 1168 | 10-Mar 10:15 |
| 102 | Oldes | For use with cheyenne you may use external gzip and cache the compressed content and serve files compressed/uncompressed according the http request header... using compression on the fly is not too wise as you would add cpu load. There are modules for appache which does the same.. just stores the compressed content in cache. | 10-Mar 10:11 |
| 101 | Oldes | And it would be fine to just get the zlib compress/decompress from the png source. With chance to compress/decompress just stream chunks it would be great. | 10-Mar 10:07 |
| 100 | Graham | Carl, any plans on releasing the source for the odbc scheme ? I would really really like to get odbc working on Linux. Unixodbc is the accepted standard. | 10-Mar 8:50 |
| 99 | Oldes | GZIP just add some additional info like modification date.. http://tools.ietf.org/html/rfc1952 | 10-Mar 8:47 |
| 98 | Oldes | the code is in the source for PNG I/O | 10-Mar 8:45 |
| 97 | Oldes | First we would need CRC32... Rebol's compress/decompress is based on ZLIB, just without the checksum | 10-Mar 8:41 |
| 96 | WillArp | for Cheyenne, what's needed can be found here http://www.cherrypy.org/wiki/GzipFilter | 10-Mar 1:36 |
| 95 | WillArp | I'm looking for Oldes.. | 10-Mar 1:27 |
| 94 | Carl | For that matter, even finding someone to do the testing. | 10-Mar 1:27 |
| 93 | Carl | It's not clear to me what's required or how we'd test it. | 10-Mar 1:27 |
| 92 | Carl | I deferred it because... | 10-Mar 1:26 |
| 91 | WillArp | tracker item#11, it would be really nice to have native gzip encoding to use with Cheyenne, if possible 8) | 10-Mar 1:25 |
| 90 | Carl | No. But, it does seem faster here for some strange reason. Maybe just due to CPU's being faster these days. | 10-Mar 1:24 |
| 89 | Robert | Carl, any real speed improvements to expect or is this just a placebo? | 9-Mar 21:53 |
| 88 | Carl | supercharged | 9-Mar 21:41 |
| 87 | Robert | No problems found so far. Feels a bit faster / snappier. Is this possible or just a self-fullfilling-prophecy of my own? | 9-Mar 21:34 |
| 86 | Robert | I tested 2.7.6 against my application: - uses SQLite DLL - 1.6 MB of rebol source-code - uses RebGUI - does a lot of calculations - uses DRAW extensively | 9-Mar 21:33 |
| 85 | Carl | I hope everyone is finding a bit of time to test 2.7.6 beta.... to be released quite soon, and we want it to be useful. Don't find out it crashes when you press escape key or something... the day after it is released.... ;) | 9-Mar 18:38 |
| 84 | Gabriele | i still think we should support only one distro and leave everything else to R3 and community effort. | 9-Mar 13:39 |
| 83 | Gabriele | for debian i guess it's iceweasel, BT may now better here :) | 9-Mar 13:38 |
| 82 | Gabriele | i think we can assume firefox in Ubuntu and derivatives. maybe if not there, try konqueror, but i guess that people not having firefox installed can as well put their own shell script to solve the issue. | 9-Mar 13:38 |
| 81 | Gabriele | linux browser: distribution dependent. some distros have a shell script that you can call and will fire the user selected browsers. others just make assumptions. | 9-Mar 13:36 |
| 80 | Henrik | is also in 2.7.6.3.1 | 9-Mar 9:08 |
| 79 | Henrik | sorry %20%22, not %20%20 | 9-Mar 9:06 |
| 78 | Henrik | hmm.. in AltME on Windows, if I have Safari as the default browser, %20%20 is inserted in front of the url when using BROWSE and the page won't load. | 9-Mar 9:03 |
| 77 | Graham | I normally code the browser as a preference in linux. | 9-Mar 8:20 |
| 76 | Paul | I'm getting the Worldmaster time sync warning on startup of Altme. I assume this is due to the time change happening tonight. | 9-Mar 4:24 |
| 75 | BrianH | That's my job :) | 9-Mar 4:20 |
| 74 | Carl | Next, it needs to fix bugs before I even open the editor. ;) | 9-Mar 4:19 |
| 73 | Paul | Ahhh that is the kinda performance I like to hear. | 9-Mar 4:19 |
| 72 | Carl | Wow, rebview on my linux box boots soooo fast. The viewtop is up on the screen before I release the enter key! | 9-Mar 4:18 |
| 71 | BrianH | Or for that matter, which desktop environments actually have a "default browser" as a concept? | 9-Mar 4:16 |
| 70 | Carl | Anyone in linuxville know how to find the default browser? | 9-Mar 4:15 |
| 69 | Carl | Hmm... well clicking on it in linux does nothing. poof. | 9-Mar 4:15 |
| 68 | Paul | If your looking to put status updates I really like how Jerry handled the notification of R3 Unicode status update - check out his very informative pic at: http://rebollovesjerry.blogspot.com/2008/03/rebol-30unicode_09.html | 9-Mar 4:14 |
| 67 | Carl | Yes, absolutely does include a .r site! | 9-Mar 4:12 |
| 66 | Paul | Woohoo. Does this include .net site also? | 9-Mar 4:10 |
| 65 | Carl | I decided to go ahead and start the rebolution of the rebolution, even before graphics. ;) | 9-Mar 4:08 |
| 64 | Carl | exactly | 9-Mar 4:08 |
| 63 | BrianH | Content first, style later? | 9-Mar 4:07 |
| 62 | Carl | www.rebol.com - the change begins | 9-Mar 4:06 |
| 61 | Carl | I've found the selection limited. But, then, I don't have time to do a really deep scan. | 9-Mar 3:54 |
| 60 | Carl | Ashley: I thought that was something different... and what we need is authenticode, microsoft's exe authenticating signatures. | 9-Mar 3:53 |
| 59 | BrianH | It may be a requirement if we want to do that iPhone support we've been talking about. | 9-Mar 2:42 |
| 58 | Ashley | Doesn't CAcert do this for free ( http://www.cacert.org/ )? I remember Kaj handing out their forms at the end of DevCon06 ... | 8-Mar 21:34 |
| 57 | Carl | Meaning, a requirement for certain apps to get stamp "this app is ok"... or at least, "we know where they live" ;) | 8-Mar 19:36 |
| 56 | Carl | In fact, I think it will become a requirement. | 8-Mar 19:35 |
| 55 | Carl | For REBOL, RT could do the same sort of thing, using existing R methods and code. | 8-Mar 19:35 |
| 54 | Carl | Yep. | 8-Mar 19:33 |
| 53 | Maarten | Carl: that's basically the trust busines, whether it's an accountant or a certificate issuer, or.... | 8-Mar 18:38 |
| 52 | Pekr | you can as well start your own CA - there is SW for it out there. You can certificate us, developers ... | 8-Mar 17:46 |
| 51 | BrianH | How would that work with the REBOL community? Do we have me program something, or start an argument with Pekr? :) | 8-Mar 17:42 |
| 50 | Pekr | Certificates are good, no? It is about one factor - some independent third party can confirm that you are you :-) | 8-Mar 17:40 |
| 49 | Carl | It seems to me that these certificate vendors just sit back and collect money. | 8-Mar 17:33 |
| 48 | Carl | Robert, well, for our own REBOL trust tree, I want to do that. But, my msg above refers to the signing keys/etc needed to be authentic in Win32. | 8-Mar 17:32 |
| 47 | Robert | IMO the whole certificate stuff is totally useless in that none can really cross-check if this is all true and valid or not. And, there were even problems with the source of the certificates which issued wrong ones or it was possible to fake those. | 8-Mar 8:55 |
| 46 | Robert | Carl, how about using a PGP signature that's based on the "web of trust" and those good known people from the Rebol community sign RTs key? | 8-Mar 8:53 |
| 45 | BrianH | Count on some lossage? :) | 7-Mar 22:51 |
| 44 | Carl | I think what you can do is route it through Ukiah. | 7-Mar 22:50 |
| 43 | Maarten | BTW: any tips on international shipping of alcohol (this seems to be hard with customs) apreciated. I have to ship some Geneva to L.A. | 7-Mar 22:15 |
| 42 | Carl | I am looking for some advice... In May 2008, REBOL's Authenticode certification will need to be renewed. We buy ours from Thawte, and its pricy. Also, the process of obtaining the code signing certificate is a big hassle. Every 2 years we do it, but we sure don't like doing it. So, I find myself wondering what other developers use? Surely there must be something easier, cheaper, and also as widely accepted as Thawte! | 7-Mar 20:36 |
| 41 | Henrik | no, the bugtracker holds the list of what is being considered for fixing. | 7-Mar 7:02 |
| 40 | Graham | Is there a list of what will be fixed? | 7-Mar 4:21 |
| 39 | Brent | added folders similar to folders for R3a release:
/releases/dev
/releases/release
/releases/test I'm assuming we'll use a similar process as was used for the R3a release. | 5-Mar 21:08 |
| 38 | Henrik | perhaps it's time to limit ourselves now and see what bugs can be fixed. so far it looks like there is only time for 5-10 bugfixes including testing, and there are only 2 days left. | 5-Mar 20:05 |
| 37 | BrianH | Please do. We need to plan the future releases too. | 5-Mar 19:53 |
| 36 | Henrik | do we keep this world after release? might be useful | 5-Mar 18:22 |
| 35 | Carl | I see why I did not fix it -- must ssh to the server to make the admin change -- security, got to love it. | 4-Mar 19:28 |
| 34 | Carl | Fixing it... | 4-Mar 19:24 |
| 33 | Carl | Pekr: wanted to do that, but it's been so long, I kept getting denied from upnews! | 4-Mar 19:24 |
| 32 | Carl | Graham: I think that's worth fixing. | 4-Mar 19:23 |
| 31 | Pekr | Carl, maybe you could post your 2.7.6 related blogs here: http://www.rebol.net/cgi-bin/upnews.r instead of Carl's blog? It would be good to have them in one place? Hmm, but due to low activity in R2 updates, maybe noone visits that section? :-) So maybe post blogs in above link, and refer to it in Carl's blog? Dunno .... | 4-Mar 11:46 |
| 30 | Graham | Was tested on sdk 2.7.5 | 4-Mar 6:50 |
| 29 | Graham | Anyone wish to confirm this? | 4-Mar 6:31 |
| 28 | Graham | and of course the black box appears. | 4-Mar 6:31 |
| 27 | Graham | launch/quit but in linux, I have to do call system/options/boot and then quit. | 4-Mar 6:30 |
| 26 | Graham | In windows I can do this | 4-Mar 6:30 |
| 25 | Graham | which reminds me that 'launch doesn't work on Linux platforms | 4-Mar 5:54 |
| 24 | Paul | Overlapping functionality but not overlapping internal code so it might add overhead (stats/evals) when doing thrown into loops but I don't suppose that would happen much for those function at least not for lengthly loops that I can think of. | 4-Mar 3:48 |
| 23 | Ashley | Maybe more of an issue for R3, but I noticed the convo on CALL and RUN which got me to thinking that we have at least 4 funcs with overlapping functionality: run call launch browse It'd be nice if we could normalize them down to a single function with refinements to handle the things that make them unique. | 4-Mar 3:27 |
| 22 | BrianH | Fix submitted. | 4-Mar 0:04 |
| 21 | BrianH | Carl was right about the problem with clean-path in R3. It was in the portion of the code that he wrote :) | 3-Mar 23:53 |
| 20 | BrianH | If you're doing that, profile. You find out some amazing things. For instance, I found that CASE is faster than two or more EITHERs. | 3-Mar 23:14 |
| 19 | Paul | yeah I'm thinking the same and going to go back and look at some code in my current apps. | 3-Mar 22:13 |
| 18 | Henrik | I think it's been there for quite a while. Meant for perl users, but I use it as it's faster than if not [] | 3-Mar 22:12 |
| 17 | Paul | Yeah I see. I should use it more. Was that available prior to 2.7.5? | 3-Mar 22:12 |
| 16 | Henrik | if | 3-Mar 22:11 |
| 15 | Henrik | nope, it's opposite of | 3-Mar 22:11 |
| 14 | Paul | Ah... I'm thinking it behaved opposite of while. Thanks henrik. | 3-Mar 22:10 |
| 13 | Henrik | unless works very well :-) use it all the time | 3-Mar 22:08 |
| 12 | Henrik | no block at the condition part | 3-Mar 22:08 |
| 11 | Paul | I never use it. But just looked at it and attempted the following: >> blk: [1 2 3 4 5] == [1 2 3 4 5] >> unless [tail? blk][print first blk blk: next blk] == none | 3-Mar 22:05 |
| 10 | Paul | Anyone know if unless is broke? | 3-Mar 22:03 |
| 9 | Paul | We discussed this not to long ago but do we think that 'case should support break? Since it is similiar to a loop when /all is used. | 3-Mar 21:55 |
| 8 | Carl | In R3, there seemed to be a lot (and I mean a lot!) more to it. | 3-Mar 21:06 |
| 7 | Carl | Is it? | 3-Mar 21:05 |
| 6 | Pekr | Dunno - if it is just one parameter change, what's the problem OTOH? :-) Well, if it is not the priority, ok then ... | 3-Mar 21:04 |
| 5 | Carl | This does not look like an R2 priority. | 3-Mar 21:02 |
| 4 | BrianH | If we can't really support Unicode in R2, what's the point of using RegisterClassW? | 3-Mar 21:01 |
| 3 | Pekr | Dunno if it would make sense, but I digged up following. While I know we can't get Unicode for R2, there were some attempts by Jerry, and maybe we could help to improve situation for R2 few bits with registering different window class. Look at older Jerry's message: IsWindowUnicode: make routine! [ hWnd [ integer! ] return: [ integer! ] ] User32.dll "IsWindowUnicode" The IsWindowUnicode returns false, which means REBOL/view windows are not Unicode-aware, because they're registered using RegisterClassA() instead of RegisterClassW. I came to a dead end. What should I do for fun on weekends? Luckily I still have the bi-directional text render problem to solve, it's my top priority now. | 3-Mar 21:00 |
| 2 | Carl | BB in 30. | 3-Mar 20:43 |
| 1 | Carl | New ALL group. | 3-Mar 20:42 |