| # | User | Message | Date |
| 374 | Gabriele | eg. >> view layout [box snow effect [draw [fill-pen black pen none shape [move 0x6 'line 6x6 'line 0x-12]]]] Segmentation fault | 20-Feb-10 15:43 |
| 373 | Gabriele | DRAW seems to segfault on Linux in 2.7.7. (same draw block works on 2.7.6.) | 20-Feb-10 15:43 |
| 372 | BrianH | Got through finally. Perhaps someone can tell me where to collect them. | 2-Jan-10 20:34 |
| 371 | BrianH | Trying to post the release notes, but the blog posting limits are stopping me. One more doc post to go once it lets me post again. | 2-Jan-10 13:32 |
| 370 | Graham | you guys forever dwelling in the past ... | 1-Jan-10 3:20 |
| 369 | BrianH | You and your time traveling :) | 1-Jan-10 3:18 |
| 368 | BrianH | Weirdo New Zealanders :) | 1-Jan-10 3:18 |
| 367 | BrianH | Interesting - I can wait til next year to try it :) | 1-Jan-10 3:17 |
| 366 | Graham | next year for me long while now! l8r | 1-Jan-10 3:17 |
| 365 | Graham | Ahh... but it is redirecting writes ... | 1-Jan-10 3:17 |
| 364 | BrianH | In any case, must go now. Happy New Year, Be Back Next Year :) | 1-Jan-10 3:16 |
| 363 | Graham | I'm now running as non-admin and it's not objecting to me writing to the current directory | 1-Jan-10 3:16 |
| 362 | BrianH | I've been following the project though, and I dowenload the VMs of the new versions as they come out to test. | 1-Jan-10 3:15 |
| 361 | BrianH | There's a lot of stuff that I've been just getting up and running recently. ReactOS is on my list, on principle. | 1-Jan-10 3:15 |
| 360 | Graham | Ok, assumed you were running it .. | 1-Jan-10 3:14 |
| 359 | BrianH | Haven't checked the latest version yet. Ask me in a week, or download a VM yourself and try it. | 1-Jan-10 3:13 |
| 358 | Graham | How does it compare with say XP ? | 1-Jan-10 3:12 |
| 357 | BrianH | ReactOS has improved drasticly in the last year. I didn't say supported (without someone paying me), I said tested :) | 1-Jan-10 3:12 |
| 356 | BrianH | Graham, I'm going to put in the aliases next month after I go over the R3 versions and improve them as needed. | 1-Jan-10 3:11 |
| 355 | Graham | I couldn't get it to run my encapped application a year or so ago. | 1-Jan-10 3:11 |
| 354 | Graham | Is reactos that stabe now?? | 1-Jan-10 3:10 |
| 353 | BrianH | Btw, as as I am concerned ReactOS will be a tested platform for the Windows builds of R2 and R3, just on principle. I want to make an HX host of R3 too if I can get away with it, and maybe SanOS :) | 1-Jan-10 3:10 |
| 352 | Graham | really need those aliases in there by default .. cd, pwd, ls etc | 1-Jan-10 3:10 |
| 351 | Carl | HNY. BBNY. | 1-Jan-10 3:09 |
| 350 | Carl | We had it running 64 bits in 2001. | 1-Jan-10 3:08 |
| 349 | Graham | They've dropped the 32 bit version | 1-Jan-10 3:08 |
| 348 | Carl | It will be interesting to try. R2 used to build that way. | 1-Jan-10 3:08 |
| 347 | Graham | Ubuntu server is only 64 bit now | 1-Jan-10 3:08 |
| 346 | BrianH | Particularly of R3. | 1-Jan-10 3:07 |
| 345 | BrianH | Well, glad to hear it. I want a 64bit version too :) | 1-Jan-10 3:07 |
| 344 | Graham | we all want 64 bit versions ... | 1-Jan-10 3:07 |
| 343 | Carl | Website will be rebuilt very soon. | 1-Jan-10 3:07 |
| 342 | Graham | none of those is sold now ... | 1-Jan-10 3:06 |
| 341 | Carl | In fact, we were contacted just today... someone wants a 64 bit version. | 1-Jan-10 3:06 |
| 340 | Graham | the installer links to this page http://www.rebol.com/products.html | 1-Jan-10 3:06 |
| 339 | Carl | Sure, yeah. | 1-Jan-10 3:06 |
| 338 | Carl | Ah, products. | 1-Jan-10 3:06 |
| 337 | Carl | Any? | 1-Jan-10 3:06 |
| 336 | Graham | Or, change the link to the SDK instead ... | 1-Jan-10 3:06 |
| 335 | BrianH | Carl, want to chime in here? Are there any? | 1-Jan-10 3:06 |
| 334 | Graham | The installer should remove that link to "Commercial Rebol products" as they don't exist now. | 1-Jan-10 3:05 |
| 333 | BrianH | Cool, that is better than I was expecting. Fixing the installer next month should be easy :) | 1-Jan-10 3:05 |
| 332 | Graham | same as non admin | 1-Jan-10 3:04 |
| 331 | BrianH | Sorry, I'll try to trigger a UAC prompt in the next version. What does it say is the appdata directory when you run as admin? | 1-Jan-10 3:04 |
| 330 | Graham | Most windows apps do not require that you run as admin to install though | 1-Jan-10 3:03 |
| 329 | Graham | Ok, trying run as admin ... | 1-Jan-10 3:03 |
| 328 | Graham | Windows 7 professional | 1-Jan-10 3:02 |
| 327 | BrianH | Yes. | 1-Jan-10 3:02 |
| 326 | Graham | view-277-2.exe | 1-Jan-10 3:02 |
| 325 | Graham | Do I have to be admin to install? | 1-Jan-10 3:02 |
| 324 | Graham | can not copy rebol.exe to home directory error | 1-Jan-10 3:02 |
| 323 | BrianH | Too painful to think about - must go out and drink now :) | 1-Jan-10 3:01 |
| 322 | Carl | Yep, and I have a vista around here somewhere, just for such suffering cases. | 1-Jan-10 3:01 |
| 321 | BrianH | I'd say it's good enough for the final build. | 1-Jan-10 3:00 |
| 320 | BrianH | I'll check out the registry later, see what needs tweaking, and test on Win7 too. | 1-Jan-10 3:00 |
| 319 | BrianH | Works for me :) | 1-Jan-10 2:59 |
| 318 | Carl | So, that'll do it for now.... it it all works ok, we'll do final build tomorrow. | 1-Jan-10 2:56 |
| 317 | Carl | there it is. | 1-Jan-10 2:55 |
| 316 | Carl | will upload it now... | 1-Jan-10 2:54 |
| 315 | BrianH | I am really behind on that stuff. Once I start testing alternate host builds I'm going to have a crash course (with real crashing no doubt). | 1-Jan-10 2:53 |
| 314 | Carl | So windows inherits those createprocess settings. | 1-Jan-10 2:52 |
| 313 | Carl | yep, good one | 1-Jan-10 2:51 |
| 312 | BrianH | CALL/show perhaps? | 1-Jan-10 2:51 |
| 311 | Carl | That CALL line starts REBOL, but it does not open any windows! | 1-Jan-10 2:48 |
| 310 | Carl | Here's the problem: == %/C/Program%20Files/rebol/view/ >> call a/rebol.exe | 1-Jan-10 2:47 |
| 309 | Carl | REBOL is it's own installer. Quite the trick. | 1-Jan-10 2:47 |
| 308 | BrianH | Wait, why is the exe in Program Files running when the installer is running? They should be separate exes. | 1-Jan-10 2:43 |
| 307 | Carl | The biggest problem is you cannot write an exe if it's running. | 1-Jan-10 2:41 |
| 306 | Carl | Well, I rewrote it for 2.7... it's actually pretty good most of the time. | 1-Jan-10 2:41 |
| 305 | BrianH | Well, if you can't work it out I say disable the installer and we'll rewrite it next month. | 1-Jan-10 2:38 |
| 304 | Carl | hard to say | 1-Jan-10 2:36 |
| 303 | BrianH | CALL/show ? | 1-Jan-10 2:34 |
| 302 | Carl | When you install, and it runs REBOL.exe the first time... it's running in background. No windows open. | 1-Jan-10 2:33 |
| 301 | Carl | A bit odd... | 1-Jan-10 2:33 |
| 300 | BrianH | Funny :) | 1-Jan-10 2:33 |
| 299 | Carl | Was a rebol.exe in background. | 1-Jan-10 2:29 |
| 298 | Carl | XP | 1-Jan-10 2:29 |
| 297 | BrianH | What OS version? | 1-Jan-10 2:19 |
| 296 | Carl | Has a problem copying the exe. | 1-Jan-10 2:18 |
| 295 | BrianH | Is is as bad as before? If not, an improvement might be better than nothing. | 1-Jan-10 2:17 |
| 294 | Carl | The installer runs now... but some kind of problem installing. | 1-Jan-10 2:12 |
| 293 | BrianH | And what I learn can be used to build REBOL app installers, including for R3. | 1-Jan-10 2:02 |
| 292 | BrianH | And none at runtime. | 1-Jan-10 2:01 |
| 291 | BrianH | With a month to work on it, I can really get the new installer working well. I want a UAC prompt when appropriate on Vista/7/2008 :) | 1-Jan-10 2:01 |
| 290 | Carl | I think so. More testing. | 1-Jan-10 1:59 |
| 289 | BrianH | Yay? | 1-Jan-10 1:59 |
| 288 | Carl | Looks like commenting out the flag above did the job. | 1-Jan-10 1:59 |
| 287 | BrianH | I did half of it in two days, the rest incrementally. The compatibility testing took the longest, given that I had to se the scientific method to determine the subtleties of R3's behavior. | 1-Jan-10 1:59 |
| 286 | Carl | R2/Forward is quite impressive work. | 1-Jan-10 1:57 |
| 285 | BrianH | There's quite a few :) | 1-Jan-10 1:57 |
| 284 | Carl | And, thanks for 2.7.7 enhancements! | 1-Jan-10 1:57 |
| 283 | Carl | Ok, have a good night/morning. | 1-Jan-10 1:56 |
| 282 | BrianH | I'm going to leave this on so I can check in before I leave. | 1-Jan-10 1:56 |
| 281 | BrianH | You take a look at that, and I'll get breakfast and clean up to go out :) | 1-Jan-10 1:55 |
| 280 | Carl | That can't be good. | 1-Jan-10 1:54 |
| 279 | Carl | #define NO_AUTO_INSTALL 1 | 1-Jan-10 1:54 |
| 278 | BrianH | Put figuring that out on the priority list. We've been going on faith so far, and some people haven't been contributing for that reason. | 1-Jan-10 1:54 |
| 277 | Carl | So... anything critical can get hung up that way. | 1-Jan-10 1:53 |
| 276 | Carl | Well... it's problematic if there are many copyrights, because each country has different laws on it. | 1-Jan-10 1:52 |
| 275 | BrianH | You should reevaluate the licensing policy if you want contributions from those who are more paranoid than I am. | 1-Jan-10 1:50 |
| 274 | BrianH | I hope the license addendum I put on the files meets with your lawyer's approval. It was easier than splitting out the R2/Forward changes, especially since it's MIT licensed. I didn't want to assign copyright. | 1-Jan-10 1:48 |
| 273 | BrianH | done. | 1-Jan-10 1:46 |
| 272 | Carl | ah... so system/options/install is set to noinstall | 1-Jan-10 1:42 |
| 271 | BrianH | Not quite yet :) | 1-Jan-10 1:41 |
| 270 | Carl | but, there's not rush... if you have to get going | 1-Jan-10 1:40 |
| 269 | Carl | ok | 1-Jan-10 1:40 |
| 268 | Carl | we can do that after 2.7.7 | 1-Jan-10 1:40 |
| 267 | BrianH | I'll do that now. | 1-Jan-10 1:40 |
| 266 | Carl | yes | 1-Jan-10 1:40 |
| 265 | BrianH | Oh, if you want to split mezz-series, splitting off the string functions might be best: ASCII? LATIN1? UTF? ENLINE DELINE | 1-Jan-10 1:39 |
| 264 | Carl | It looks like it isn't even calling check-install. | 1-Jan-10 1:38 |
| 263 | Carl | I'll give it a quick go. If it's just a line or two shouldn't be a problem. | 1-Jan-10 1:37 |
| 262 | BrianH | Should that be tomorrow's task? Or next month's? | 1-Jan-10 1:36 |
| 261 | BrianH | Right now it just starts up in desktop mode, strewing the current directory with desktop files. It makes it difficult to use for scripts. | 1-Jan-10 1:35 |
| 260 | BrianH | It gets set by the installer. Plus, the exe should realize it's not installed. | 1-Jan-10 1:33 |
| 259 | Carl | Grepping, I find 7 places where view-root is set! | 1-Jan-10 1:33 |
| 258 | BrianH | There is a view-root registry entry that you should be checking. | 1-Jan-10 1:32 |
| 257 | Carl | No pun intended. | 1-Jan-10 1:32 |
| 256 | Carl | And an extra good one to boot. | 1-Jan-10 1:32 |
| 255 | Carl | Yes, party! | 1-Jan-10 1:32 |
| 254 | BrianH | view-root <> boot path (or at least it shouldn't) | 1-Jan-10 1:31 |
| 253 | BrianH | Yup. And me going out and partying tonight :) | 1-Jan-10 1:31 |
| 252 | Carl | That is, the view-root problem. | 1-Jan-10 1:30 |
| 251 | Carl | So, only thing left is installer, right? | 1-Jan-10 1:30 |
| 250 | Carl | Note: >> 1 !== 2 == true >> 1 !== 1 == false | 1-Jan-10 1:29 |
| 249 | Carl | Ok, view-277 uploaded. | 1-Jan-10 1:29 |
| 248 | BrianH | I look forward to spending a week reviewing migrated RAMBO tickets in CureCode :) | 1-Jan-10 1:27 |
| 247 | BrianH | Yay, to both! | 1-Jan-10 1:26 |
| 246 | Carl | BTW, removed RAMBO from the help panel. | 1-Jan-10 1:26 |
| 245 | Carl | Boots! | 1-Jan-10 1:25 |
| 244 | Carl | yep, I have to type nn then nf ;) | 1-Jan-10 1:25 |
| 243 | BrianH | Oh, wait, got it :) | 1-Jan-10 1:24 |
| 242 | BrianH | You have nf | 1-Jan-10 1:24 |
| 241 | Carl | btw, need nf to check for updates like nn | 1-Jan-10 1:23 |
| 240 | Carl | ok | 1-Jan-10 1:23 |
| 239 | BrianH | Done. Try again :) | 1-Jan-10 1:23 |
| 238 | Carl | build worked | 1-Jan-10 1:21 |
| 237 | Carl | ok | 1-Jan-10 1:19 |
| 236 | BrianH | Wait, if we accept that overhead I can reduce the memory usage of types-of. Give me a minute :) | 1-Jan-10 1:19 |
| 235 | BrianH | Has a little more overhead: one reduce rather than a copy, a case statement, an intermediate block for each pseudo-typeset in types-of | 1-Jan-10 1:17 |
| 234 | BrianH | Done. | 1-Jan-10 1:15 |
| 233 | Carl | The original post said "by the end of the week" (not "end of the year" ;) | 1-Jan-10 1:13 |
| 232 | Carl | BTW... we've got 1 more day. | 1-Jan-10 1:12 |
| 231 | BrianH | I'm almost done. | 1-Jan-10 1:12 |
| 230 | Carl | So, we leave out mezz-reflect? | 1-Jan-10 1:11 |
| 229 | Carl | got it | 1-Jan-10 1:09 |
| 228 | BrianH | mezz-error is new. mezz-control has changed, now doesn't have *-ERROR | 1-Jan-10 1:08 |
| 227 | Carl | not seeing any diffs | 1-Jan-10 1:08 |
| 226 | Carl | ok | 1-Jan-10 1:06 |
| 225 | BrianH | I sbmitted a changed mezz-control and mezz-error to chat too. mezz-error can go before everything or after mezz-fnc. | 1-Jan-10 1:05 |
| 224 | Carl | ok | 1-Jan-10 1:04 |
| 223 | BrianH | I have a solution. Already fixed to-typeset, now fixing types-of. | 1-Jan-10 1:04 |
| 222 | Carl | Updating version # and dates. | 1-Jan-10 1:00 |
| 221 | BrianH | Give me a sec. | 1-Jan-10 0:50 |
| 220 | Carl | Also, if nothing else uses this code, we can put it in 2.7.8 | 1-Jan-10 0:49 |
| 219 | Carl | If you change: switch/default :value [ to switch/default to-word :value [ doesn't that fix it? | 1-Jan-10 0:49 |
| 218 | BrianH | No, it's that part that is the problem. | 1-Jan-10 0:48 |
| 217 | Carl | Me to, on rebin. | 1-Jan-10 0:47 |
| 216 | Carl | The only problem is the #[ ] code. | 1-Jan-10 0:47 |
| 215 | Carl | That part is ok. | 1-Jan-10 0:47 |
| 214 | BrianH | I look forward to rebin... | 1-Jan-10 0:47 |
| 213 | BrianH | The result of the reduces in the parens of to-typeset and types-of. | 1-Jan-10 0:46 |
| 212 | Carl | I only see #[datatype ...] are there others? | 1-Jan-10 0:45 |
| 211 | BrianH | Give me a sec and I'll see it there is an easy refactoring. | 1-Jan-10 0:44 |
| 210 | BrianH | The compose/only tricks there was to make the reduce only happen once. | 1-Jan-10 0:44 |
| 209 | BrianH | No, it won't - there will be huge reduce overhead at runtime if we can't have inline datatypes without reducing them. | 1-Jan-10 0:42 |
| 208 | Carl | For #[] the scanner has to call the datatype MAKE function anyway. | 1-Jan-10 0:41 |
| 207 | Carl | Performance should be the same. | 1-Jan-10 0:41 |
| 206 | BrianH | Yeah, that's going to be a performance problem. | 1-Jan-10 0:40 |
| 205 | Carl | So, the code must construct all datatypes, no scanner shortcuts. | 1-Jan-10 0:40 |
| 204 | Carl | The problem is that the system does not expect any #[type] values in its boot. | 1-Jan-10 0:40 |
| 203 | BrianH | Ah, got it. Give me a sec - it must be choking on the compose. | 1-Jan-10 0:39 |
| 202 | Carl | Ah... that's it... | 1-Jan-10 0:38 |
| 201 | Carl | So, it's likely that MOLD is not doing what we need. | 1-Jan-10 0:38 |
| 200 | Carl | Here's how to reproduce the error in mezz-reflect: >> do load mold load %mezz-reflect.r ** Script Error: head expected series argument of type: series port | 1-Jan-10 0:37 |
| 199 | Carl | yes | 1-Jan-10 0:32 |
| 198 | BrianH | mezz-error.r perhaps? After mezz-func.r, before mezz-series.r | 1-Jan-10 0:31 |
| 197 | BrianH | Yeah, I keep telling people about the improved design. | 1-Jan-10 0:31 |
| 196 | Carl | mezz-series, mezz-series2. | 1-Jan-10 0:30 |
| 195 | Carl | What we should do is div up the series functions. | 1-Jan-10 0:30 |
| 194 | Carl | We could move them to mezz-func | 1-Jan-10 0:30 |
| 193 | Carl | Fixed problem with operators. R3 is so much better designed! | 1-Jan-10 0:29 |
| 192 | BrianH | That wouldn't cause the script error above though. | 1-Jan-10 0:29 |
| 191 | BrianH | A lot of the series functions use throw-error or cause-error internally. They should be defined before they are called. | 1-Jan-10 0:28 |
| 190 | Carl | no. | 1-Jan-10 0:24 |
| 189 | BrianH | Did you put the error handling code (cause-error, throw-error) in mezz-func? | 1-Jan-10 0:22 |
| 188 | Carl | The init goes in phases. | 1-Jan-10 0:21 |
| 187 | Carl | #include %mezz-func.r #include %mezz-series.r #include %mezz-control.r #include %mezz-file.r #include %mezz-input.r | 1-Jan-10 0:21 |
| 186 | Carl | Those are in the boot mezz section, listed above. | 1-Jan-10 0:20 |
| 185 | BrianH | Where in that list is mezz-file, mezz-control and mezz-series? | 1-Jan-10 0:19 |
| 184 | Carl | #include %mezz-help.r #include %mezz-web.r #include %mezz-config.r #include %mezz-xml.r #include %mezz-spoof.r ;#include %mezz-reflect.r #include %mezz-cvs.r #include %mezz-boot.r | 1-Jan-10 0:18 |
| 183 | BrianH | What order do you build the mezz files in in the bild where you get the error in types-of? | 1-Jan-10 0:12 |
| 182 | BrianH | Rejected by the scanner? | 1-Jan-10 0:11 |
| 181 | Carl | != works | 1-Jan-10 0:11 |
| 180 | BrianH | Didn't work? | 1-Jan-10 0:09 |
| 179 | Carl | operators rejected | 1-Jan-10 0:06 |
| 178 | Carl | yes | 1-Jan-10 0:05 |
| 177 | BrianH | Are the words unset! and error! defined at that point? | 1-Jan-10 0:04 |
| 176 | BrianH | I added a missing compose/only, but I didn't even get the script error without it. | 1-Jan-10 0:04 |
| 175 | Carl | hmmm. | 1-Jan-10 0:03 |
| 174 | BrianH | I don't get the script error. | 31-Dec-09 23:58 |
| 173 | Carl | added != !== debugging !== for integer | 31-Dec-09 23:57 |
| 172 | Carl | (here in testing) | 31-Dec-09 23:53 |
| 171 | Carl | It's uploaded. | 31-Dec-09 23:53 |
| 170 | BrianH | If you post a build without mezz-reflect, I can test mezz-reflect locally. | 31-Dec-09 23:52 |
| 169 | BrianH | Wait, that should have worked before, both those words are in any-type!. | 31-Dec-09 23:50 |
| 168 | BrianH | Got it. Change head remove find head remove find to-typeset any-type! error! unset! to head remove find to-typeset any-type! unset! | 31-Dec-09 23:48 |
| 167 | Carl | I need some time anyway to fix those other things listed above. | 31-Dec-09 23:46 |
| 166 | Carl | Ok, NP. Will upload a build here with all except mezz-reflect. | 31-Dec-09 23:45 |
| 165 | BrianH | Give me a sec to think. | 31-Dec-09 23:45 |
| 164 | Carl | valtype: ( head remove find head remove find to-typeset any-type! error! unset! ) ; R2 excludes error! here where R3 doesn't | 31-Dec-09 23:44 |
| 163 | Carl | mezz-reflect | 31-Dec-09 23:44 |
| 162 | BrianH | Sorry, mezz-func, mezz-control, the rest. Do you know where that error comes from? | 31-Dec-09 23:43 |
| 161 | Carl | ** Script Error: head expected series argument of type: series port ** Near: head remove find head remove | 31-Dec-09 23:42 |
| 160 | Carl | ok, building... | 31-Dec-09 23:41 |
| 159 | BrianH | Mezz-func first, the rest are standalone | 31-Dec-09 23:40 |
| 158 | Carl | what order? | 31-Dec-09 23:39 |
| 157 | BrianH | Done, mezz-control and mezz-reflect updated. Try it now. | 31-Dec-09 23:38 |
| 156 | Carl | I tried various combinations, but nothing worked. | 31-Dec-09 23:37 |
| 155 | Carl | Yes, go ahead. | 31-Dec-09 23:36 |
| 154 | BrianH | If you like I can rearrange it in the mezz source on DevBase. | 31-Dec-09 23:33 |
| 153 | BrianH | It's at the top of the file in R2/Forward, just after the function stuff. | 31-Dec-09 23:32 |
| 152 | BrianH | Crap, that's it. Move the typeset spoofing to the top of the file and it will be fine. | 31-Dec-09 23:31 |
| 151 | Carl | That compose will happen at boot time, but to-typeset is defined after this. | 31-Dec-09 23:30 |
| 150 | BrianH | It's funny, those reflectors are well tested. I had to replace references to them in the other mezz changes to inlined equivalent code. When I've been using R2/Forward they work fine. | 31-Dec-09 23:30 |
| 149 | Carl | I've not studied the code... so, just a guess: types-of: funct [ "Returns a copy of the types of a function." value ] compose/only [ ; Returns types as blocks of datatype! values anytype: (to-typeset any-type!) ^^^^^ | 31-Dec-09 23:30 |
| 148 | BrianH | The order in the function, or the order in the script? The reflectors aren't called before the typeset spoofing funcs are, they're just created. I am not married to the order specified though. | 31-Dec-09 23:28 |
| 147 | Carl | Ok... good. Removed mezz-reflect from build, removed ASSERT. Built succeeds. | 31-Dec-09 23:27 |
| 146 | Carl | Something seems wrong with the order of to-typeset in mezz-reflect.r | 31-Dec-09 23:26 |
| 145 | BrianH | Ah, that will do too. I would prefer a native assert though for future releases (as I mentioned earlier) :) | 31-Dec-09 23:23 |
| 144 | Carl | I think for now, will just move assert. | 31-Dec-09 23:21 |
| 143 | BrianH | It can be moved. | 31-Dec-09 23:21 |
| 142 | BrianH | The typeset spoofing is in mezz-reflect.r | 31-Dec-09 23:21 |
| 141 | Carl | assert uses to-typeset | 31-Dec-09 23:20 |
| 140 | Carl | Looks like an init order problem. | 31-Dec-09 23:20 |
| 139 | Carl | Running... boot error: ** Script Error: to-typeset has no value ** Near: to-typeset any-type! | 31-Dec-09 23:19 |
| 138 | Carl | Build complete, no errors. | 31-Dec-09 23:19 |
| 137 | BrianH | Ah, got it. I hadn't tried that hack for stripping things out of prebol. Remove that whole thing. | 31-Dec-09 23:17 |
| 136 | Carl | Found it: #do [comment [ ; So this section is not loaded by prerebol. | 31-Dec-09 23:15 |
| 135 | BrianH | Ooo, that is interesting. | 31-Dec-09 23:08 |
| 134 | Carl | I've never seen it happen before... in fact. | 31-Dec-09 23:07 |
| 133 | Carl | It comes from one of the build preprocessing scripts. | 31-Dec-09 23:07 |
| 132 | BrianH | Where's that error coming from? | 31-Dec-09 23:07 |
| 131 | Carl | quite strange | 31-Dec-09 23:07 |
| 130 | Carl | Looks like a : needed. | 31-Dec-09 23:06 |
| 129 | Carl | Build crash.... Performing Custom Build Step on ..\..\src\boot.r ------ Building: boot.r ** Script Error: data has no value ** Where: do-expr ** Near: if unset? data [ print ["***" cmd "must return a value or none:" mold file] ] | 31-Dec-09 23:04 |
| 128 | BrianH | OK, that's cool. I really want the rest of the mezz code in R3 chat so we can sic the REBOL optimizer on it :) | 31-Dec-09 23:03 |
| 127 | Carl | Core and View include all the mezz. | 31-Dec-09 23:01 |
| 126 | Carl | I've got an idea... we'll do Core and View today, but SDK next week. | 31-Dec-09 23:01 |
| 125 | BrianH | I think? Unless the error management code went into mezz-control.r - perhaps we need a mezz-error.r for cause-error, throw-error and attempt. | 31-Dec-09 23:00 |
| 124 | BrianH | The mezz-func.r should go before mezz-series.r - there is error management code there too. | 31-Dec-09 22:58 |
| 123 | BrianH | Rearrange as you see fit. I didn't know how to rewrite mezz.r. | 31-Dec-09 22:57 |
| 122 | Carl | The boot core uses: #include %mezz-series.r #include %mezz-control.r #include %mezz-file.r #include %mezz-input.r | 31-Dec-09 22:57 |
| 121 | BrianH | Is there a problem with putting the rest of the mezzanine source in #41? I mean the stuff that is built into /Base and such. | 31-Dec-09 22:57 |
| 120 | Carl | So, we may need to split up some of the files. | 31-Dec-09 22:56 |
| 119 | Carl | Yes, but, the SDK does not include all mezz-*.r | 31-Dec-09 22:56 |
| 118 | BrianH | Ooo, interesting! I warned you that you's need to put mezz-func.r first. | 31-Dec-09 22:55 |
| 117 | Carl | Dependency problem... FUNCT used before defined. | 31-Dec-09 22:54 |
| 116 | Carl | I looked at it... seems good to me! Will put it in the std build. | 31-Dec-09 22:51 |
| 115 | BrianH | I really hope you include the spoofed closures and typeset functions in the build. Spoofing datatypes in mezz code is my favorite trick :) | 31-Dec-09 22:49 |
| 114 | Carl | Me too! | 31-Dec-09 22:46 |
| 113 | BrianH | I don't think anything will break, but I'm keeping my fingers crossed :) | 31-Dec-09 22:45 |
| 112 | Carl | Ok. Almost ready to build with mezz changes. | 31-Dec-09 22:45 |
| 111 | BrianH | If you have any questions or concerns, ask here. I'll be available. | 31-Dec-09 22:44 |
| 110 | Carl | Ok. I am wrapping up some wiki work, and will start the merge in about 20 min. | 31-Dec-09 21:38 |
| 109 | BrianH | Graham has a good http revamp that can make it into next version, with some tweaks. Also, I didn't want to port over most of the mezz-file changes without some discussion and testing. I did add a COPY to WHAT-DIR though. | 31-Dec-09 21:36 |
| 108 | Carl | Ok, thanks Brian. Have a great NYE/D. | 31-Dec-09 21:34 |
| 107 | BrianH | Submitted proposed 2.7.7 changes: #6591, #6593, #6595, #6596, #6597, #6598. Notes: - mezz-func.r is required by -control, -series and -file. - mezz-reflect.r isn't strictly required, but would be useful to build in. - mezz-spoof.r is not required but might be good to include in the SDK. | 31-Dec-09 20:39 |
| 106 | BrianH | OK, I am done integrating the R2/Forward stuff that I feel safe integrating. Time to check and post the changes. | 31-Dec-09 20:10 |
| 105 | Graham | Release today? | 31-Dec-09 20:10 |
| 104 | BrianH | Carl, what is EVAL used for in R3? | 31-Dec-09 19:16 |
| 103 | BrianH | But the STRICT-NOT-EQUAL? bug is a killer, and the reason noone mentioned it is because we didn't have an op! for that function so it was rarely used. Once we add an op! the bug will be glaring. | 31-Dec-09 19:03 |
| 102 | BrianH | That and the INSERT/dup/part bug are my major pet peeves in R2. And an obscure PARSE BREAK bug that is tough to replicate. | 31-Dec-09 19:02 |
| 101 | Carl | It is possible to make the major fixes in it.... but it is unlikely to be as good as R3. | 31-Dec-09 19:00 |
| 100 | BrianH | It doesn't compare integer! values, just their types. For decimal: >> strict-not-equal? 1.0 2.0 == true | 31-Dec-09 18:57 |
| 99 | BrianH | >> strict-not-equal? 1 2.0 == true >> strict-not-equal? 1 2 == false ; should be true | 31-Dec-09 18:56 |
| 98 | BrianH | Give me a sec, I'll track it down in CureCode. You fixed it in R3. | 31-Dec-09 18:54 |
| 97 | Carl | what bug? | 31-Dec-09 18:53 |
| 96 | BrianH | If you can add the new ops, please do. And fix the STRICT-NOT-EQUAL? bug (if it is still there). | 31-Dec-09 18:52 |
| 95 | Carl | DEFAULT may be good. FUNCO not needed. | 31-Dec-09 18:52 |
| 94 | Carl | Maybe != and !== | 31-Dec-09 18:51 |
| 93 | BrianH | The version of MAP-EACH I'm adding assumes that bug#709 will get implemented. The /into option is a two-comparison fix to the R2 version once it is structured to support list! values. | 31-Dec-09 18:42 |
| 92 | BrianH | I'm not sure about DEFAULT. Also FUNCO, FUNCT, TO-FUNCTION, CLOSURE, CLOSURE!, CLOSURE?, TO-CLOSURE. | 31-Dec-09 18:24 |
| 91 | BrianH | I'll list the new functions here when I'm done. So far: CAUSE-ERROR, THROW-ERROR (only needed for R2), ASSERT, !, ++, --, ALSO, QUOTE, TRUE?, FIRST+, AJOIN, TAKE, SWAP. | 31-Dec-09 18:22 |
| 90 | BrianH | I'll put in APPLY and ASSERT but we really need native versions of those, and COLLECT-WORDS and RESOLVE as well. I have an outdated RESOLVE that I want to update before I put it in, so not in this version. | 31-Dec-09 18:19 |
| 89 | BrianH | Already included :) | 31-Dec-09 18:17 |
| 88 | Carl | Lol! | 31-Dec-09 18:17 |
| 87 | BrianH | Yes, I also have a fake closure! type :) | 31-Dec-09 18:17 |
| 86 | Carl | I think it would be a good idea to get R2er's using those. Preps them for R3. | 31-Dec-09 18:17 |
| 85 | BrianH | I was going to post the source to the REFLECT set in a different file. | 31-Dec-09 18:16 |
| 84 | BrianH | I have full fake versions of APPLY, MAP-EACH, REFLECT and the -OF functions. Should I include them? | 31-Dec-09 18:16 |
| 83 | Carl | It's close, maybe a few small tweaks. | 31-Dec-09 18:15 |
| 82 | BrianH | No REWORD - the behavior of that isn't final, and I'm not yet sre it will make the mezz cut in R3. | 31-Dec-09 18:14 |
| 81 | Carl | BTW, that view-root problem must be from newer releases... because the appdata/rebol has desktop in it - but from a couple years ago | 31-Dec-09 18:14 |
| 80 | Carl | If includes REWORD, hold off on that until 2.7.8 | 31-Dec-09 18:13 |
| 79 | Carl | Ok, good. | 31-Dec-09 18:11 |
| 78 | BrianH | I'm currently copying over some stuff from R2/Forward to the mezz files. Mostly better error handling, though some new functions. | 31-Dec-09 18:10 |
| 77 | BrianH | If it's more than a 2-line fix I'll be shocked. | 31-Dec-09 18:08 |
| 76 | BrianH | There are two installation dirs: The dir of the exe and the desktop root, and they should be different dirs. The view-root should be that %APPDATA%\REBOL dir. | 31-Dec-09 18:07 |
| 75 | Carl | checking... view-root comes up as the boot dir | 31-Dec-09 18:06 |
| 74 | Carl | Does it put the desktop files in the current dir or in the installation dir? | 31-Dec-09 18:05 |
| 73 | BrianH | The big issue can wait for next time. | 31-Dec-09 18:02 |
| 72 | BrianH | For this round, just checking and using the view root would be enough. Right now it just puts the desktop files in the current directory, regardless of where the view root in the registry says to put them. The view-root variable is not getting set. | 31-Dec-09 18:02 |
| 71 | Carl | On install, do we have a specific issue to fix, or is it mainly that big issue about per-user installation? | 31-Dec-09 15:29 |
| 70 | Carl | BrianH: on APPENDs, no. We don't want to upset the apple cart too much. But things like FUNCT, yes. That's about all I use these days. | 31-Dec-09 15:26 |
| 69 | Carl | Checking in. Are we set to build? | 31-Dec-09 15:25 |
| 68 | Graham | though the protcols are quite limited in the way they allow access | 31-Dec-09 9:40 |
| 67 | Graham | Ok, I had the wrong impression then. | 31-Dec-09 9:39 |
| 66 | Gabriele | Graham: I do, it's fully tested on GMail, but it is lower level than imap:// (though, a new imap:// could be built on top of it). it will be released eventually, but i don't know when, and definitely I don't see that happening for 2.7.7. | 31-Dec-09 9:30 |
| 65 | Graham | what's wrong with the current 'append? | 31-Dec-09 8:25 |
| 64 | BrianH | mninmal -> minimal | 31-Dec-09 6:09 |
| 63 | BrianH | For future R2 versions, I have a few big-bang-for-buck natives that could use backporting: APPLY, ASSERT, COLLECT-WORDS, and RESOLVE. These should be portable with mninmal impact and no model changes. | 31-Dec-09 6:09 |
| 62 | BrianH | Quick decision questions: Should we include the backports of APPEND and REPEND ? They are much more complex, i.e.: append: funco [ "Inserts a value at tail of series and returns the series at head. (Modifies)" series [series! port! bitset!] "Series at point to insert" value [any-type!] "The value to insert" /part "Limits to a given length or position" length [number! series! port! pair!] /only "Inserts a series as a series" /dup "Duplicates the insert a specified number of times" count [number! pair!] ][ ; Nasty, but the best you can do without native APPLY head do pick pick pick [[[ [insert tail series get/any 'value] [insert/part tail series get/any 'value length] ] [ [insert/only tail series get/any 'value] [insert/part/only tail series get/any 'value length] ]] [[ [insert/dup tail series get/any 'value count] [insert/part/dup tail series get/any 'value length count] ] [ [insert/dup/only tail series get/any 'value count] [insert/part/dup/only tail series get/any 'value length count] ]]] not dup not only not part ] | 31-Dec-09 6:06 |
| 61 | BrianH | R2 2.7.6 SDK sources posted to R3 chat in the appropriate headers. Time to integrate the backwards-compatible R2/Forward changes. | 31-Dec-09 2:41 |
| 60 | Graham | Releasing SSL for view was a great move ... | 30-Dec-09 21:45 |
| 59 | Graham | And can your imap protocol use port 993 ( secure ) ? | 30-Dec-09 21:45 |
| 58 | Graham | GAbriele .. I understand you have a fully functional imap protocol .. that you or qtask were considering open sourcing. Is this something that can go into 2.7.7 ( since it's been already fully tested at qtask ) ? | 30-Dec-09 21:43 |
| 57 | BrianH | As long as we move the registry entries to the right places, that's cool. Although that fix could also be put in 2.7.8. | 30-Dec-09 20:13 |
| 56 | Carl | It's quite possible that it is broken, because I normally upgrade.... not do a fresh install. | 30-Dec-09 20:13 |
| 55 | Carl | Ok, so let's do this... confirm that the basic install works (system, not per user). That's an easy check, and easy fix (I think.) | 30-Dec-09 20:12 |
| 54 | BrianH | It is sandboxed, but I'm not sure how much is by the OS and how much is by Chrome. Vista/7 offers some OS sandboxing but I haven't needed to look into it yet. This is all kind of sudden, so I'm getting up to speed as fast as I can. Most of the installer issues I I know from memory, since I had to do multiuser R2 installs before. | 30-Dec-09 20:10 |
| 53 | BrianH | If I get the installer working properly on 2000+, we might want to integrate the fix for testing purposes. I have VMs here, but not enough of them installed for the cut-off time. We should put off the installer until next month. I want to test on Win9x, 2000, XP/2003, x64 builds, and 7 (which will cover Vista). And ReactOS :) | 30-Dec-09 20:07 |
| 52 | Carl | Ah... so is it sandboxed by the OS? | 30-Dec-09 20:04 |
| 51 | Carl | I don't have a virtualized system here... so can only test it in a few specific cases. | 30-Dec-09 20:04 |
| 50 | BrianH | For an application running as a least-privileged user, there is no problem with it being installed per-user. It can't do anything bad anyways. | 30-Dec-09 20:04 |
| 49 | Carl | The problem with modifying installer is that you've got to test it on various user accounts and OS versions. | 30-Dec-09 20:03 |
| 48 | Carl | Ok, well... this could require a lot of debugging in a short period of time. We may want to defer it for later. | 30-Dec-09 20:03 |
| 47 | BrianH | Win7 uses the same rules, it just enforces them. | 30-Dec-09 20:02 |
| 46 | Carl | What kind of security does that offer? | 30-Dec-09 20:02 |
| 45 | Carl | I've got to imagine that V7 must vomit on such installs, no? | 30-Dec-09 20:01 |
| 44 | BrianH | A bad one. | 30-Dec-09 20:01 |
| 43 | BrianH | Local Settings/Application Data. So it doesn't roam. | 30-Dec-09 20:01 |
| 42 | Carl | What kind of wreck of an OS is this anyway?? | 30-Dec-09 20:01 |
| 41 | Carl | Wow, it installs in Application Data. | 30-Dec-09 20:01 |
| 40 | BrianH | Yes. They learned the trick from MySpace IM (I guess), which did it to let students get around admin restrictions on school computers. | 30-Dec-09 20:00 |
| 39 | Carl | Chome installs in local?? | 30-Dec-09 19:59 |
| 38 | Carl | I think it's no longer a requirement to work on very old Win32, so we can make such a change (per user). | 30-Dec-09 19:58 |
| 37 | Carl | Yes, originally, the registry values were setup for the entire system, not per user. | 30-Dec-09 19:58 |
| 36 | Carl | It could be that the installer GUI isn't setting the view root if that field is modified. (Which I did not change yesterday.) | 30-Dec-09 19:57 |
| 35 | Carl | Well, the 2.7 installer must work to some degree... I installed yesterday's build, and it looks correct. Of course, this was an upgrade, so the dirs already existed. | 30-Dec-09 19:56 |
| 34 | BrianH | The question is whether we want to go central, or least-privilege. Chrome uses the latter model - it puts its program files under local settings along with its data files, and still maintains file associations. | 30-Dec-09 19:56 |
| 33 | BrianH | Quick policy question: Do we want to support per-user R2 exe's? Program Files is for central install by an admin. We could support that as a default, but also support per-user exe files, file associations and such on later versions of Windows (2000+). | 30-Dec-09 19:54 |
| 32 | BrianH | And the registry entries are still going in the wrong section of the registry, which is why admin is needed and multiuser doesn't work. | 30-Dec-09 19:51 |
| 31 | BrianH | Yes, but the 2.7 installer doesn't work, as of 2.7.6 - it doesn't detect the View root and puts the desktop files in the current directory. | 30-Dec-09 19:49 |
| 30 | Carl | I should note that 2.7 fixed many issues with the installer. So, it might be that very little is needed. | 30-Dec-09 19:47 |
| 29 | BrianH | I agree that the installer might take some time, which is why I put it last on the list. If it can't be done in time, move it to next month. | 30-Dec-09 19:25 |
| 28 | BrianH | No summary doc yet, not without going back to the mailing list archive from 2000. | 30-Dec-09 19:24 |
| 27 | Carl | Also, is there a summary doc of what's wrong with the installer? | 30-Dec-09 16:07 |
| 26 | Carl | Brian... you may want to recruit a couple folks on some of the tasks. Changing the installer alone could take a lot of time. It's time consuming to test it, and some of those source files are "wordy." | 30-Dec-09 16:05 |
| 25 | Carl | Uploaded testing/cmd.exe -- test verifies fix. Thanks for writing such a simple test. That helps a lot! | 30-Dec-09 15:58 |
| 24 | Carl | That fact was useful. Found the problem. Fixing it. | 30-Dec-09 15:50 |
| 23 | Carl | Gabriele... that information is *very* interesting. When it comes to LOAD, there should be no diff between Core and Command! | 30-Dec-09 15:45 |
| 22 | Carl | Ok. It would be good to keep it minimal... in order to get it out without taking a lot of your time... and to have enough time to test any changes (because there's no beta test.) | 30-Dec-09 15:44 |
| 21 | BrianH | I'll be able to work fulltime on this for most of Wednesday and Thursday. My priorities are:
- Getting the SDK source into R3 chat so we can modify the files
- Reviewing R2/Forward to figure out the least disruptive changes we can adapt, and updating FUNCT
- Evaluating and integrating the patches from the community, as appropriate
- Fixing the installer Perhaps we can migrate some changes from the old DevBase but I suspect that all of the non-VID changes are in R2/Forward. | 30-Dec-09 9:40 |
| 20 | Gabriele | (you may need to add an "halt" at the end of the script on windows...) | 30-Dec-09 6:00 |
| 19 | Gabriele | REBOL/Command 2.7.6.4.2 (14-Mar-2008)
Copyright 2008 REBOL Technologies
REBOL is a Trademark of REBOL Technologies
All rights reserved. Finger protocol loaded Whois protocol loaded Daytime protocol loaded SMTP protocol loaded ESMTP protocol loaded POP protocol loaded IMAP protocol loaded HTTP protocol loaded FTP protocol loaded NNTP protocol loaded HTTPS protocol loaded Licensed to: Gabriele Santilli (commercial) <g.santilli@tiscalinet.it> License ID: 5-50021-1 Script: "Test LOAD with pass-thru handlers" (none) test protocol loaded >> read test://test == "some string" >> load test://test == "some string" | 30-Dec-09 6:00 |
| 18 | Gabriele | this is the "wrong" one I get with Command: | 30-Dec-09 5:59 |
| 17 | Gabriele | REBOL/Core 2.7.6.4.2 (14-Mar-2008)
Copyright 2008 REBOL Technologies
REBOL is a Trademark of REBOL Technologies
All rights reserved. Finger protocol loaded Whois protocol loaded Daytime protocol loaded SMTP protocol loaded ESMTP protocol loaded POP protocol loaded IMAP protocol loaded HTTP protocol loaded FTP protocol loaded NNTP protocol loaded Script: "Test LOAD with pass-thru handlers" (none) test protocol loaded >> read test://test == "some string" >> load test://test == [some string ] | 30-Dec-09 5:59 |
| 16 | Gabriele | this is the "correct" output I get with Core: | 30-Dec-09 5:59 |
| 15 | Gabriele | uploaded on Tests/load-on-pass-thru.r | 30-Dec-09 5:59 |
| 14 | Gabriele | ah, it is getting more interesting. REBOL/Core 2.7.6.4.2 (14-Mar-2008) does not show the bug, but REBOL/Command 2.7.6.4.2 (14-Mar-2008) does! | 30-Dec-09 5:57 |
| 13 | Carl | Gab: did you upload it? I think I have the fix, but need a test for it. | 29-Dec-09 23:57 |
| 12 | Carl | Brian, on R2/Forward... the ones that we know make the most difference and are well tested. E.g. FUNCT would be really nice to have. | 29-Dec-09 21:27 |
| 11 | BrianH | Thanks! | 29-Dec-09 19:43 |
| 10 | Gabriele | i'll upload a test script here asap. | 29-Dec-09 13:55 |
| 9 | Gabriele | added as #4380 | 29-Dec-09 13:55 |
| 8 | Henrik | I will not attempt any VID stuff for this release, although #4371 is simple. | 29-Dec-09 9:28 |
| 7 | BrianH | Should the R2-compatible mezzanine fixes from R2/Forward be integrated as well? The answer to that depends on the timeframe. | 29-Dec-09 9:14 |
| 6 | BrianH | I will attempt to fix the installer in time for the release. It should be easy enough. | 29-Dec-09 9:12 |
| 5 | Carl | The 2.7 test release has been built. This is the base build for the next release. It contains SSL, ODBC, DES, etc., and no-license key is required. In addition, I've added an install checkbox to the Prefs (User panel) and an Uninstall to the Help panel. These are just shortcuts to existing features. The download is www.rebol.com/downloads/view277-test1.exe Note that it's version # is 2.7.6, but it has a new system/build date. Don't mix it up with prior versions, as it's not at all tested. | 29-Dec-09 4:55 |
| 4 | Carl | RAMBO is best, because it will get ported to CureCode. | 29-Dec-09 4:54 |
| 3 | Gabriele | should I post my problem with LOAD with pass-thru scheme handlers to RAMBO? Or here? (in case you want to fix it in the future) | 29-Dec-09 4:36 |
| 2 | Carl | A good place to post notes about next 2.7 release? | 29-Dec-09 4:06 |
| 1 | Carl | Created this group. | 29-Dec-09 4:06 |