| # | User | Message | Date |
| 537 | Oldes | Replace all code in %prot-http.r: port/port-id <> 80 to: not find [80 443] port/port-id to avoid a bug (infinite loop) with redirection from HTTPS scheme to HTTP scheme like in this example: read https://sourceforge.net/ | 27-Sep-10 9:34 |
| 536 | Henrik | I've now seen it twice. Both crashes occuring in different places after making some heavy modifications to the program. I expect to see it more, so I'll try to see if I can narrow it down. | 30-Dec-09 7:17 |
| 535 | Carl | Most likely cause is a bad memory pointer. Do you see this error often? | 29-Dec-09 23:56 |
| 534 | Carl | That error indicates a block was being interpreted as a string. | 29-Dec-09 23:56 |
| 533 | Henrik | Unfortunately it's a large program, but I've not seen this error before, so I wonder if there are any pointers on what is going on? | 29-Dec-09 21:12 |
| 532 | Henrik | ---------------------------
REBOL Error
---------------------------
REBOL Internal Error: Invalid series width 1 was 16 (type 39) Program terminated abnormally. This should never happen. Contact www.REBOL.com with details. --------------------------- OK --------------------------- | 29-Dec-09 21:12 |
| 531 | Henrik | Just had an interesting crash in 2.7.6: | 29-Dec-09 21:12 |
| 530 | Oldes | I found the reason: in net-utils is used global word 'slash for #"/" which is redefined in the script above. So probably not a bug. The script above should be enclosed in context or maybe not using global word in so important part of code like net-utils | 13-Jan-09 20:39 |
| 529 | Oldes | So No... it's not my bug.. it's the kind of strange bugs: REBOL/View 2.7.6.3.1 10-Mar-2008 Copyright 2000-2008 REBOL Technologies. All rights reserved. REBOL is a trademark of REBOL Technologies. WWW.REBOL.COM >> do http://box.lebeda.ws/~hmm/rebol/expr-tests.r Expression: [1 + 2 * 3 / 4] =============================================== Standard priority/associativity evaluator [add 1 divide multiply 2 3 4] Simple right-associative expression evaluator [add 1 multiply 2 divide 3 4] REBOL-like expression evaluator [divide multiply add 1 2 3 4] >> do http://box.lebeda.ws/~hmm/rebol/expr-tests.r ** Script Error: Invalid argument: / ** Where: parse-url ** Near: either parse/all url url-rules [ if net-watch [print ["URL Parse:" reduce vars]] if user [port/user: user] ... >> | 13-Jan-09 20:19 |
| 528 | Oldes | No... forget it... it's bug in one of mine scripts which modifies the http scheme | 13-Jan-09 19:57 |
| 527 | Oldes | use read instead of DO above (that the error is not from the content but from url parser). | 13-Jan-09 19:52 |
| 526 | Oldes | In 2.7.6: >> do http://box.lebeda.ws/~hmm/rebol/expr-tests.r ** Script Error: Invalid argument: / ** Where: parse-url ** Near: either parse/all url url-rules [ if net-watch [print ["URL Parse:" reduce vars]] if user [port/user: user] | 13-Jan-09 19:49 |
| 525 | Henrik | confirmed under OSX | 18-Sep-08 11:08 |
| 524 | Gabriele | just discovered this: http://www.rebol.net/cgi-bin/rambo.r?id=4346& | 18-Sep-08 10:14 |
| 523 | Cyphre | Carl, ok. | 5-Sep-08 14:36 |
| 522 | Carl | Creating separate chat for "fixes" | 4-Sep-08 21:28 |
| 521 | Carl | Cyphre: sound source code for 2.7 uploaded in source folder. | 4-Sep-08 21:26 |
| 520 | Pekr | Hi, Cyphre is currently developing base for our advertising presentation platform. We wrapped ffmped library, which means we can play 30+ formats of video, audio, images, etc., cross-platform. Cyphre even managed to get Overlay mode working, so you can run video as a background to your REBOL/View app, if you wish to :-) However - there is some difficulcy with sound! IIRC Carl knows the situation, as Kurt did some demo in the past and got in some trouble too? My question is, if it is possible to get bug fixed? The sound is fine, except its continuations - when you insert new part of the stream to the port, the sound is not continuous, but interrupted. It would be nice to get it fixed, or we will have to use external sound library, which imo, would be a pity .... | 22-Aug-08 7:59 |
| 519 | Oldes | I forgot... it's already fixed in R3:) | 21-Aug-08 8:44 |
| 518 | Oldes | Do we want to fix parse (to enable escape) not to halt Rebol by using rules like this: parse "" [any []] I know I can do: parse "" [any [()]] But what about newbies (if there are any:). | 21-Aug-08 8:38 |
| 517 | Anton | Posted the workaround to Rambo. | 22-Jun-08 16:19 |
| 516 | Anton | which creates a symbolic link "netscape" pointing to "firefox", and in rebol adds a double-quote character, which successfully works around an error sh: Syntax error: Unterminated quoted string | 22-Jun-08 15:55 |
| 515 | Anton | I'm on Kubuntu 7.04 - I did this: sudo ln -s /usr/bin/firefox /usr/bin/netscape and in Rebol/View 2.7.6.4.2: browse "http://rebol.com^(22)" | 22-Jun-08 15:53 |
| 514 | Henrik | browse does not work correctly in Windows with Safari. | 22-Jun-08 15:48 |
| 513 | Anton | I just got BROWSE working in linux with View 2.7.6.4.2 Is that news to anyone ? | 22-Jun-08 15:28 |
| 512 | Gabriele | just make face won't do much. the layout call will at least do a size-text, which might be what is needed. if that's not enough, we'll have to check if there's anything else in the text style init or in the layout itself that may be the actual workaround. (Carl?) | 20-Jun-08 9:31 |
| 511 | Pekr | And this one for the other Linux SDK version (sorry if the output is messed-up, but it was posted to icq which trimmed line terminators): 000 08:01 183149 /usr/lib/libXaw7.so.7.0.0 b7dc6000-b7dcd000 rw-p 00053000 08:01 183149 /usr/lib/libXaw7.so.7.0.0 b7dcd000-b7dce000 rw-p b7dcd000 00:00 0 b7dce000-b7e1b000 r-xp 00000000 08:01 182474 /usr/lib/libXt.so.6.0.0 b7e1b000-b7e1f000 rw-p 0004d000 08:01 182474 /usr/lib/libXt.so.6.0.0 b7e1f000-b7e2c000 r-xp 00000000 08:01 182470 /usr/lib/libXext.Aborted | 19-Jun-08 12:46 |
| 510 | Pekr | he got something like this: *** glibc detected *** ./rebface: double free or corruption (!prev): 0x082b0298 *** ======= Backtrace: ========= /lib/libc.so.6[0xb7ada4b6] /lib/libc.so.6(cfree+0x89)[0xb7adc179] /usr/lib/libX11.so.6(_XF86BigfontFreeFontMetrics+0x5b)[0xb7e29edb] /usr/lib/libX11.so.6(XFreeFont+0xad)[0xb7e29f9d] ./rebface[0x806bd1a] .... | 19-Jun-08 12:42 |
| 509 | Pekr | I suggested my friend to try make face [text: "workaround" font: make face/font []] ... but I will post him also your suggestion ... | 19-Jun-08 12:34 |
| 508 | Pekr | Gabriele - if we comment it out, rebol crashes ... both linux SDK distros ... | 19-Jun-08 12:33 |
| 507 | Gabriele | size-text make face [text: "imanXwin kludge"] | 19-Jun-08 9:30 |
| 506 | Gabriele | if so, you might be able to replace it with something like: | 19-Jun-08 9:30 |
| 505 | Gabriele | Petr, do fonts stop working if you comment it out? (I never tried commenting that out) | 19-Jun-08 9:30 |
| 504 | Pekr | Could anyone explain plese, what does following hack in Linux SDKs mean? ;-- Must be done prior to loading anything that requires fonts on Linux. layout [text "imanXwin kludge"] ;-throw this one away soon-- okay? open-events It actually prevents SDK from rebface + rebgui usage. So instead of fixing crashes we do like above? | 18-Jun-08 13:38 |
| 503 | Oldes | not = to | 13-Jun-08 11:41 |
| 502 | Oldes | Or maybe no? Maybe it's more complicated. Maybe it should be fixed to-set-word not to dehex what's possible. | 13-Jun-08 11:41 |
| 501 | Oldes | There should be: name: to-set-word dehex name Instead of: name: to-set-word name In decode-cgi function. | 13-Jun-08 11:34 |
| 500 | Henrik | recycle does not alway reduce memory usage if he is looking in the task manager. | 11-Jun-08 11:33 |
| 499 | Pekr | quicker memory consumption is with layout [view scroller] | 11-Jun-08 11:19 |
| 498 | Pekr | Bobik reports, that recycle does not help. I can't reproduce memory concumption when not clicking though .... | 11-Jun-08 11:17 |
| 497 | Henrik | although without recycling it does eat memory. | 11-Jun-08 10:16 |
| 496 | Henrik | what's the pattern after recycling? I can't quite repeat the problem here after a recycle. | 11-Jun-08 10:15 |
| 495 | Pekr | Today I was contacted by Bobik, reporting following: view layout [choice 100 "1" "2" "3" text-list 200x200 "aaa" "bbb" "ccc"] Code above, when you click anything, consumes memory. He reports, it happens with RebGUI too. He also reports, that at some stage, the memory goes up even if you do nothing, but I am not able to reproduce latter. Is there any memory leak? Or any reason, why should memory consumption go higher? | 9-Jun-08 9:57 |
| 494 | Oldes | More than position is important that long paths are truncated - invisible - which is pretty bad. | 20-May-08 16:17 |
| 493 | Henrik | Is there anything to do about the security requester sometimes placing itself at the back of a layer of windows in Windows? It also places itself sometimes in the upper left corner, sometimes in the right side, sometimes in other places. I think the position of it is related to the position of View windows or the console window, but I would much rather have it at the front and centered on screen. | 19-May-08 8:44 |
| 492 | Oldes | Submited RAMBO #4124 set-modes doesn't work on file directories into tracker here. | 28-Apr-08 20:46 |
| 491 | Gabriele | i would suggest RAMBO, actually. | 10-Apr-08 6:28 |
| 490 | BrianH | Put it in the tracker. | 9-Apr-08 19:50 |
| 489 | Oldes | I found it while decoding html entities, where the leading zero is not required | 9-Apr-08 15:02 |
| 488 | BrianH | That would make sense. ENBASE would never return such a string, but other functions could. | 9-Apr-08 13:56 |
| 487 | Oldes | Wouldn't it be better to have: debase/base "3" 16 ;== #{03} instead of none? without it I had to write such an ugly code: ent: debase/base (head insert/dup ent #"0" (length? ent) // 2) 16 | 9-Apr-08 11:20 |
| 486 | Graham | REBOL Internal Error: corrupt datatype: 67 at 78456536 | 4-Apr-08 9:53 |
| 485 | Graham | I'm still getting corrupt datatype errors with async networking on 2.7.6 | 4-Apr-08 9:52 |
| 484 | BTiffin | Thanks for the info. I'll track it here as a reminder, tagged it as wish | 1-Apr-08 15:14 |
| 483 | Gabriele | (probably a good idea to make it work with get-modes though) | 1-Apr-08 14:48 |
| 482 | Gabriele | i don't remember if you could get async-modes. maybe it is set only. can be accessed as port/async-modes though. | 1-Apr-08 14:47 |
| 481 | BTiffin | Need a little verification before tracking; >> tp: open tcp://:9097 >> get-modes tp 'network-modes == [type-of-service keep-alive receive-buffer-size send-buffer-size no-delay interf aces async-modes] >> get-modes tp get-modes tp 'network-modes ** Script Error: Invalid argument: async-modes ** Near: get-modes tp get-modes tp 'network-modes | 1-Apr-08 12:58 |
| 480 | RobertS | We do from Dolphin Smalltalk ... I'll try to find a moment ... the other thiing is to be running as an NT Service, again, if my memory serves .. this fblack box was a plague for old Visual Smalltalk on Windows ... I am still hoping to wrap a Rebol3 DLL with Dolphin Smalltlak as it is so well-behaved on Windows. And then there is my plan to wrap the R3 DLL with Curl ... | 1-Apr-08 12:39 |
| 479 | Gabriele | Can a Win32 programmer come up with a sane way to call exes without the black window popping up? | 1-Apr-08 8:07 |
| 478 | BTiffin | And - if you call "notepad" you end up with a zombie notepad. On '98 at least. | 1-Apr-08 4:34 |
| 477 | BTiffin | It turns out the magic call refinement change is /show - it was a change to stop the black windows console from flashing. Calc works, as far as I can guess, as it has no input boxes, only buttons. /shell works as that allows a console. - I think. | 1-Apr-08 2:45 |
| 476 | RobertS | This is also the case with just core as in REBOL/Core 2.7.6.3.1 (14-Mar-2008) One must call into a win32 DLL which the other does not ?? user.dll is sometimes an issue if memory serves ... | 1-Apr-08 1:20 |
| 475 | RobertS | BUT. using call/shell succeeds with "notepad.exe" under 2.7.6 as in call/shell "notepad.exe" | 1-Apr-08 1:18 |
| 474 | RobertS | in R3 alpha both work fine (but return none, not 0 ) | 1-Apr-08 1:16 |
| 473 | RobertS | REBOL/View 1.3.2.3.1 5-Dec-2005 Core 2.6.3 is OK | 1-Apr-08 1:13 |
| 472 | RobertS | Just running do %vid-usage.r in aan old beta shows the difference: the call to both app's works under old 2.7.3.1 and same under | 1-Apr-08 1:13 |
| 471 | RobertS | I first saw this on a sweep through do %vid-usage.r | 1-Apr-08 1:07 |
| 470 | RobertS | call "notepad.exe " and call "calc.exe" as both prompt for permission and both open the app. Not so for 2.7.6 where the latter works fine but the former simply returns 0 | 1-Apr-08 1:05 |
| 469 | RobertS | for | 1-Apr-08 1:04 |
| 468 | RobertS | in old versions od View on WIn XP there is no difference in result | 1-Apr-08 1:04 |
| 467 | BrianH | It's been like that since to-local-file was added as a native. | 29-Mar-08 22:47 |
| 466 | Graham | Actually I knew this all along .. my code has lots of checking for trailing "\" and "/" depending on whether it is windows or linux | 29-Mar-08 22:47 |
| 465 | Graham | Has it been rambo'ed? | 29-Mar-08 22:44 |
| 464 | BrianH | I had to write to-relative-file differently in R2 and R3 because of that. | 29-Mar-08 22:43 |
| 463 | Graham | pity we can't fix it for 2.7.7 as well. | 29-Mar-08 22:42 |
| 462 | Graham | Ah ... ok. | 29-Mar-08 22:42 |
| 461 | BrianH | That has always been the case. R3 fixes it. | 29-Mar-08 22:42 |
| 460 | Graham | >> read to-rebol-file to-local-file what-dir ** Access Error: Cannot open /d/rebol ** Near: read to-rebol-file to-local-file what-dir >> read dirize to-rebol-file to-local-file what-dir == [ .. list of file s... ] | 29-Mar-08 22:42 |
| 459 | Graham | the trailing "/" is lost | 29-Mar-08 22:41 |
| 458 | Graham | this is not symmetrical >> change-dir %/d/rebol == %/d/rebol/ >> to-local-file what-dir == "d:\rebol" >> to-rebol-file to-local-file what-dir == %/d/rebol | 29-Mar-08 22:40 |
| 457 | Anton | Gab. Done. | 26-Mar-08 15:09 |
| 456 | Gabriele | please add this to RAMBO | 26-Mar-08 14:31 |
| 455 | Gabriele | i'm not aware of any differences between ext3 and ext2 except for the journaling | 26-Mar-08 14:30 |
| 454 | Gabriele | does not hang if there's a trailing / | 26-Mar-08 14:30 |
| 453 | Anton | Somebody can confirm, or compare with ext2 filesystem ? | 26-Mar-08 13:59 |
| 452 | Anton | I have a new bug, I think. (I checked RAMBO). 27-Mar-2008 View 2.7.6.4.2 on Kubuntu linux 7.04 with ext3 filesystem close make port! %~ ; <- can be a file or directory too ; rebol hangs, max cpu | 26-Mar-08 13:58 |
| 451 | Gabriele | Updated RAMBO. If you see a bug reported there that has been fixed but not marked as so, please report it to me. Thanks! | 25-Mar-08 11:12 |
| 450 | Graham | creates it's own bugs | 24-Mar-08 22:37 |
| 449 | Graham | Two bug systems now ... | 24-Mar-08 22:37 |
| 448 | Henrik | I see that some bugs fixed in 2.7.6 have not been marked fixed in RAMBO, such as #4319. | 24-Mar-08 10:00 |
| 447 | BTiffin | Tracked here, RAMBOed the copy/part crash. | 22-Mar-08 14:53 |
| 446 | Gabriele | Brian, make sure they are in RAMBO or at least in the tracker here. Thanks! | 22-Mar-08 8:26 |
| 445 | BTiffin | Secure demo from viewtop/rebol/tests reports create-links is failing security sandbox test | 22-Mar-08 4:58 |
| 444 | BTiffin | org: "abcdef" s: skip org 4 e: tail org remove/part org 5 ;; s and e now refer to past tail org copy/part s e ;; crashes REBOL 2.7.6 and 1.3.2 but R3 handles it well. | 21-Mar-08 18:48 |
| 443 | Graham | >> call "explorer c:\"
== 0
>> call/show "explorer c:\"
== 0
>> call/show "explorer c:\"
== 0 In the first case, the explorer window no longer appears as it used to in < 2.7.6. It now needs the /show to make it work. And I don't see any black box .. not that I wish to! | 20-Mar-08 22:45 |
| 442 | Carl | (no time to dive in -- maybe it's ok) | 20-Mar-08 17:18 |
| 441 | Carl | Above patch, not sure. ;) | 20-Mar-08 17:18 |
| 440 | Carl | Will, interesting bug (decloak). Reduce it down to just a very short set of binary values, like #{a00a}. Then put it in tracker, thanks. | 20-Mar-08 17:17 |
| 439 | Oldes | Fixed the 'page in devbase. | 20-Mar-08 9:50 |
| 438 | WillArp | so what is the shortest version of above patch? | 20-Mar-08 9:37 |
| 437 | WillArp | REBOL/Core 2.7.6.2.4 (15-Mar-2008) >> a: load decloak/with load "#{3C92F2E9F5EB928DEC15EE12680DF68F708A72170CF2EF166E0D3C0D52A29B6942734172}" "123" ** Script Error: a needs a value ** Near: a: load decloak/with load {#{3C92F2E9F5EB928DEC15EE12680DF68F708A72170CF2EF166E0D3C0D52A29B6942734172}} >> | 19-Mar-08 22:16 |
| 436 | WillArp | REBOL/Command 2.7.6.2.5 (18-Mar-2008) >> a: load decloak/with load "#{3C92F2E9F5EB928DEC15EE12680DF68F708A72170CF2EF166E0D3C0D52A29B6942734172}" "123" == #{ 789C532A2E2D482D52C8C9CF4B57C848CDC9C95728CF2FCA495102006DC308C0 18000000 } >> | 19-Mar-08 22:14 |
| 435 | WillArp | have a lapsus, tell me please 8) | 18-Mar-08 21:45 |
| 434 | Carl | Is that the simplest? | 18-Mar-08 17:52 |
| 433 | WillArp | temp patch: system/schemes/http/handler/open: func append first get in system/schemes/http/handler 'open 'page second get in system/schemes/http/handler 'open | 18-Mar-08 17:40 |
| 432 | Carl | It would be great if someone can submit a fix to DevBase, so 2.7.7 can fix this bug. | 18-Mar-08 17:19 |
| 431 | Carl | This appears to be a very old bug according to source control archives. Dating back to at least 2004. | 18-Mar-08 17:15 |
| 430 | Carl | (You will not see it on normal "success" pages.) | 18-Mar-08 17:10 |
| 429 | Carl | So, this only happens on FORWARD actions. | 18-Mar-08 17:10 |
| 428 | Carl | Will has pointed out PAGE local is global. The code position is: forward: [ page: copy "" while [ ( str: pick port/sub-port 1 ) <> "" ][ append page reduce [str newline] ] headers: Parse-Header HTTP-Header page | 18-Mar-08 17:09 |
| 427 | Paul | not seeing the problem on windows. 'page still has no value on windows | 18-Mar-08 16:01 |
| 426 | WillArp | osx/intel | 18-Mar-08 15:56 |
| 425 | Paul | WillArp what platform is that? Windows? Linux? | 18-Mar-08 15:49 |
| 424 | Anton | To the bug tracker, then :) | 18-Mar-08 15:43 |
| 423 | WillArp | I would say that need an urgent fix 8) | 18-Mar-08 15:35 |
| 422 | Anton | Yes, same here on REBOL/View 2.7.6.3.1 10-Mar-2008 | 18-Mar-08 15:33 |
| 421 | WillArp | sorry for the line ends.. here is a more readable version http://reboot.ch/rebol/bug.txt | 18-Mar-08 15:32 |
| 420 | WillArp | can anyone check this please, looks like 'page get set globally? REBOL/Core 2.7.6.2.5 (11-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 >> page ** Script Error: page has no value ** Near: page >> p: read http://apple.com connecting to: apple.com connecting to: www.apple.com == {<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd"> <html> <head> ^-<title>A... >> page == {Date: Tue, 18 Mar 2008 15:25:41 GMT Server: Apache/1.3.33 (Darwin) mod_perl/1.29 Location: http://www.apple.com/ Connection: cl... >> | 18-Mar-08 15:29 |
| 419 | BTiffin | I mentoned that I'd vote for carry on the 2.7.6 release blog entry. So, when do we start chatting up 2.7.7? I can't believe I missed out on pestering Carl and Brian to backport FORMAT. 2.7.6 will still end up with a thousand and one PAD functions. | 18-Mar-08 14:14 |
| 418 | Anton | I'm not afraid to complain about that byte-limited shift. I think it's not as useful as multi-byte carry. But both can be useful. | 18-Mar-08 14:10 |
| 417 | BTiffin | I don't mind being confused with Mr. Hawley. Good for my street cred. Can't say the other Brian would agree. :) But thanks for the shift info. Makes sense. | 18-Mar-08 0:08 |
| 416 | Gabriele | (ah i was referring to Brian T not Brian H - sorry for the confusion...) | 17-Mar-08 19:34 |
| 415 | BrianH | The original design request was for the shift to carry, and later not to carry, both from the same person. Decisions about carrying were deferred to a later version. | 17-Mar-08 18:44 |
| 414 | Gabriele | Brian, shift on binary shifts each byte independently. | 17-Mar-08 18:40 |
| 413 | BrianH | Normally I would prefer an error. | 17-Mar-08 18:05 |
| 412 | Henrik | to-time "xx" produces an error under R3 | 17-Mar-08 17:48 |
| 411 | Henrik | an empty string produces an error. I'm not sure this is intentional. | 17-Mar-08 17:47 |
| 410 | Oldes | Is this better result than error? >> to-time "xx" == none | 17-Mar-08 17:41 |
| 409 | BrianH | I don't think shift carries yet. | 17-Mar-08 15:47 |
| 408 | BTiffin | What's the deal with the new SHIFT native? shift/left 1 10 1024 but shift/left #{0001} 10 goes zero. Only an 8bit result for shifting binary! ? Not complaining, just want to know what the intent is. Jerry mentioned on the release blog entry but there was no answer. | 17-Mar-08 15:23 |
| 407 | Anton | It's the 32-bit integer filesize limit. IIRC, this was fixed on Windows such that at least the file could be opened (even if the size was reported incorrectly), but I don't know about linux. | 17-Mar-08 14:42 |
| 406 | Henrik | there's no limitation in there that prevents rebol from opening files larger than 2 GB? | 17-Mar-08 9:40 |
| 405 | Anton | The large file is Fedora install DVD, at 3.6GB. | 17-Mar-08 8:27 |
| 404 | Anton | So I am able to checksum a large file using the Windows build but not the Linux build. | 17-Mar-08 8:26 |
| 403 | Anton | On linux: rebview 2.7.6.4.2 9-Mar-2008 - can't open or query large files, but rebview 2.7.6.3.1 10-Mar-2008 (via Wine) - can. | 17-Mar-08 8:25 |
| 402 | Gabriele | yes, osx! ;) | 11-Mar-08 22:19 |
| 401 | Carl | Move to db group. Thanks. I got to get back to OSX. | 11-Mar-08 22:19 |
| 400 | Carl | yes, indeed | 11-Mar-08 22:18 |
| 399 | Gabriele | it's like Brian said, you have to go to the site and buy like the rest of us ;) | 11-Mar-08 22:18 |
| 398 | Carl | Actually, I think maybe there was one for me on Irix. | 11-Mar-08 22:18 |
| 397 | BrianH | I'm running on Windows Server 2003, which has a different ODBC version than Windows XP. What does the 'columns command return? | 11-Mar-08 22:18 |
| 396 | Carl | No, I think it is because the only license key that I could find was for cindy@rebol -- I don't rate high enough for my own license key ;) | 11-Mar-08 22:17 |
| 395 | Carl | lol. | 11-Mar-08 22:16 |
| 394 | Gabriele | i think it's Ukiah. we should all move there, then all REBOL bugs would disappear. | 11-Mar-08 22:16 |
| 393 | Carl | Create a "DB 2.7.7" group. Move discussion on this topic there. | 11-Mar-08 22:16 |
| 392 | BrianH | I switch the type to decimal and it worked. | 11-Mar-08 22:16 |
| 391 | Carl | Timeout. Sorry, must move back to 2.7.6 release. | 11-Mar-08 22:16 |
| 390 | Carl | nope, same result. worked. | 11-Mar-08 22:15 |
| 389 | BrianH | I'm having the problem with MSSQL 2005. Here is the the columns report: [["GREYLIST" "dbo" "blah" "t" 6 "float" 15 8 none 10 1 none none 6 none none 1 "YES" none none none none none none 109]] The table is named "blah" in my test. | 11-Mar-08 22:13 |
| 388 | Carl | trying something different... 1 min... | 11-Mar-08 22:12 |
| 387 | Graham | I'm going home to test on a different PC. | 11-Mar-08 22:12 |
| 386 | Graham | well, at least Brian sees something. | 11-Mar-08 22:11 |
| 385 | Graham | 2.03 | 11-Mar-08 22:11 |
| 384 | Graham | not when I've been pulling it out! | 11-Mar-08 22:10 |
| 383 | Carl | What FB vers are you on? | 11-Mar-08 22:10 |
| 382 | Carl | lol, is that a good thing? | 11-Mar-08 22:10 |
| 381 | Graham | I have no hair left now | 11-Mar-08 22:10 |
| 380 | Carl | I used the example in the bug report | 11-Mar-08 22:09 |
| 379 | Carl | float | 11-Mar-08 22:09 |
| 378 | BrianH | What is your column type? | 11-Mar-08 22:08 |
| 377 | Carl | Well, I am using latest FB distros. | 11-Mar-08 22:08 |
| 376 | Carl | I do not see the bug. | 11-Mar-08 22:08 |
| 375 | Graham | what's different? | 11-Mar-08 22:08 |
| 374 | Carl | [[1.0] [1.1] [1.0] [1.1]] | 11-Mar-08 22:07 |
| 373 | Carl | lol | 11-Mar-08 22:07 |
| 372 | Gabriele | :) | 11-Mar-08 22:06 |
| 371 | BrianH | Ask Cindy for one, Carl. That's what the rest of us had to do. | 11-Mar-08 22:06 |
| 370 | Graham | Brian , that is encouraging. | 11-Mar-08 22:06 |
| 369 | Graham | I can send you mine! | 11-Mar-08 22:05 |
| 368 | BrianH | Graham, that is confirmed. | 11-Mar-08 22:05 |
| 367 | Carl | Back.... sorry about delay. I needed a license.key. (I *hate* licence.key) | 11-Mar-08 22:05 |
| 366 | Pekr | I wonder where does Command ODBC code comes from? | 11-Mar-08 22:04 |
| 365 | Pekr | Graham - really old? I remember the situation with fastcgi ... I talked to fastcgi author about not supporting dynamic mode under Windows, and the author said it is long time fixed. Then, and only then Holger confirmed, that RT is using old implementation and he was not willing to correct that, as it was not a priority .... | 11-Mar-08 22:03 |
| 364 | Graham | and the correct value for float values that were "integer"? | 11-Mar-08 22:02 |
| 363 | Graham | So, in your testing, you got none back for float values that were not an "integer" ? | 11-Mar-08 22:02 |
| 362 | BrianH | Float isn't part of the SQL standard - it's an extension. | 11-Mar-08 22:00 |
| 361 | Graham | How old would that have to be? | 11-Mar-08 22:00 |
| 360 | BrianH | It could be that REBOL is using an older odbc version that doesn't support the float type. | 11-Mar-08 21:59 |
| 359 | Graham | except under Wine | 11-Mar-08 21:58 |
| 358 | Graham | decimal works | 11-Mar-08 21:58 |
| 357 | Pekr | OK, good to know ... then it might be related to dtype conversions? | 11-Mar-08 21:58 |
| 356 | BrianH | I'll try the decimal type and see if that helps. | 11-Mar-08 21:58 |
| 355 | BrianH | I am testing with MSSQL - I don't have Firebird here. | 11-Mar-08 21:57 |
| 354 | Graham | ie. I didn't cross test with a native tool | 11-Mar-08 21:57 |
| 353 | Pekr | Couldn't it be problem of particular ODBC driver implementation? What abou another sources as MS SQL? | 11-Mar-08 21:57 |
| 352 | Graham | Brian - ahh .. I didn't test that. | 11-Mar-08 21:57 |
| 351 | Pekr | but right, this is bug channel, not chit-chat .... | 11-Mar-08 21:56 |
| 350 | BrianH | Graham, the bug seems to have no effect on whether I can connect later. The behavior in the tracker happens though. Looking at the table, the values were inserted correctly - the problem seems to be on select. | 11-Mar-08 21:56 |
| 349 | Pekr | No, but R3 will bring those features for free anyway, no? How many potential Command buyers are out there? As for me, I paid few SDKs, IOS, Pro, so I am ok .... | 11-Mar-08 21:56 |
| 348 | Graham | Do you want free wine as well?? | 11-Mar-08 21:55 |
| 347 | Pekr | And then we can ask Carl, if he could just accidentally release ODBC for free too :-) | 11-Mar-08 21:54 |
| 346 | Graham | Carl needs to confirm... | 11-Mar-08 21:54 |
| 345 | Pekr | Bug was confirmed, no? So I expect it will be corrected now? | 11-Mar-08 21:54 |
| 344 | Graham | can we concentrate please on this bug? | 11-Mar-08 21:53 |
| 343 | Graham | Pekr, your db preferences are not on topic | 11-Mar-08 21:53 |
| 342 | Pekr | Graham - not in my standards, period. | 11-Mar-08 21:53 |
| 341 | Pekr | No, let's say I want client app without installing a client driver. Or at least its automated configuration ... | 11-Mar-08 21:52 |
| 340 | Graham | Pekr, everything needs a driver | 11-Mar-08 21:52 |
| 339 | Carl | not there yet... pasting your test, Graham. | 11-Mar-08 21:52 |
| 338 | BrianH | If you want a client app without installing a server, use SQLite. | 11-Mar-08 21:52 |
| 337 | Graham | Thanks Brian. | 11-Mar-08 21:52 |
| 336 | Graham | and then Rebol can't open the odbc connection? | 11-Mar-08 21:52 |
| 335 | BrianH | Bug confirmed. | 11-Mar-08 21:51 |
| 334 | Carl | ah, I see test... yes worked | 11-Mar-08 21:51 |
| 333 | Pekr | BrianH - so no client app without the need to install ODBC driver first? That sucks, no? | 11-Mar-08 21:51 |
| 332 | Carl | I would think so. | 11-Mar-08 21:51 |
| 331 | Graham | Brian, you confirmed the bug?? | 11-Mar-08 21:51 |
| 330 | Pekr | user dsn should be ok, no? Can you open it in rebol then? | 11-Mar-08 21:50 |
| 329 | BrianH | Confirmed on SQL Server 2005 over the SQL Native Client odbc driver. | 11-Mar-08 21:50 |
| 328 | Carl | hmm... I added it as a user DSN. | 11-Mar-08 21:50 |
| 327 | Graham | the "Test Connection" button on the system dsn setup | 11-Mar-08 21:49 |
| 326 | Carl | test? | 11-Mar-08 21:49 |
| 325 | Graham | and did it test okay? | 11-Mar-08 21:49 |
| 324 | Carl | already linked to this fdb | 11-Mar-08 21:49 |
| 323 | Carl | finished that | 11-Mar-08 21:49 |
| 322 | Carl | C:\Program Files\Firebird\Firebird_2_0\examples\empbuild.fdb | 11-Mar-08 21:48 |
| 321 | Graham | You're creating a system dsn right? | 11-Mar-08 21:48 |
| 320 | Graham | what is the path? | 11-Mar-08 21:48 |
| 319 | Carl | Error: windows cannot open .fdb file | 11-Mar-08 21:47 |
| 318 | Graham | or just paste in the path | 11-Mar-08 21:47 |
| 317 | Graham | browse to the employee.fdb on the c drive | 11-Mar-08 21:46 |
| 316 | Carl | (you don't mean the file, right, you mean via tcp or odbc url?) | 11-Mar-08 21:46 |
| 315 | Carl | what's the url? | 11-Mar-08 21:45 |
| 314 | Carl | browse to it? | 11-Mar-08 21:45 |
| 313 | Graham | and browse to the database | 11-Mar-08 21:44 |
| 312 | Carl | ok | 11-Mar-08 21:43 |
| 311 | Graham | any name you want | 11-Mar-08 21:43 |
| 310 | Carl | employee? | 11-Mar-08 21:43 |
| 309 | Graham | Database account is SYSDBA, and password is masterkey | 11-Mar-08 21:43 |
| 308 | Carl | so, what's the dsn for above? "employees" | 11-Mar-08 21:43 |
| 307 | Graham | when creating the system dsn leave all the defaults | 11-Mar-08 21:42 |
| 306 | Carl | I see it. | 11-Mar-08 21:42 |
| 305 | Carl | checking... | 11-Mar-08 21:41 |
| 304 | Graham | well, your path will be different | 11-Mar-08 21:41 |
| 303 | Graham | C:\Program Files\Firebird\Firebird_1_5\examples\employee.fdb | 11-Mar-08 21:41 |
| 302 | Graham | yes you need to setup a dsn | 11-Mar-08 21:40 |
| 301 | Graham | see if you got the employees.fdb | 11-Mar-08 21:40 |
| 300 | Carl | it is running, fbserver, don't I need to setup a db on it to add it to odbc? | 11-Mar-08 21:40 |
| 299 | Graham | the installer would have started it .. if you agreed | 11-Mar-08 21:38 |
| 298 | Carl | I guess I need to init an instance of FB. | 11-Mar-08 21:38 |
| 297 | Carl | oh wait... got to add it. | 11-Mar-08 21:38 |
| 296 | Graham | :( | 11-Mar-08 21:38 |
| 295 | Carl | y | 11-Mar-08 21:37 |
| 294 | Graham | You're on XP? | 11-Mar-08 21:37 |
| 293 | Carl | yes | 11-Mar-08 21:37 |
| 292 | Graham | you installed odbc? | 11-Mar-08 21:37 |
| 291 | Carl | not there | 11-Mar-08 21:37 |
| 290 | Graham | Firebird/Interbase driver | 11-Mar-08 21:36 |
| 289 | Graham | yes | 11-Mar-08 21:36 |
| 288 | Carl | G: shouldn't FB show up in ODBC ctrl panel now? | 11-Mar-08 21:36 |
| 287 | Carl | nice install -- (I think) | 11-Mar-08 21:35 |
| 286 | Carl | ok, so got them (engine and odbc) | 11-Mar-08 21:31 |
| 285 | Graham | whoever is funding sourceforge I guess | 11-Mar-08 21:30 |
| 284 | Carl | why is sourceforge using doubleclick anyway? who there is making the $? | 11-Mar-08 21:29 |
| 283 | Carl | I hate that.... my download connection blocks out on my firewall doubleclick filter. | 11-Mar-08 21:29 |
| 282 | Carl | ahhhh! downloads thru sourceforge... | 11-Mar-08 21:28 |
| 281 | BrianH | The SQL language support of Firebird is quite limited - very old-school SQL. | 11-Mar-08 21:28 |
| 280 | Graham | FB is on all platforms .. not sure why it isn't used more. | 11-Mar-08 21:28 |
| 279 | Graham | that'll do | 11-Mar-08 21:27 |
| 278 | Carl | I like it! | 11-Mar-08 21:26 |
| 277 | Carl | Hey, these guys are just like RT... everything over last few years is beta. | 11-Mar-08 21:26 |
| 276 | Carl | Version 2.0 Beta? 5 Sep 2007 (OdbcJdbc v2-0-Beta Ver 02.00.00.0144) | 11-Mar-08 21:26 |
| 275 | Carl | So, what do you want me to DL? | 11-Mar-08 21:26 |
| 274 | Graham | It's not GPL | 11-Mar-08 21:25 |
| 273 | Carl | I thought anyone could buy anything... mostly GPL. | 11-Mar-08 21:25 |
| 272 | Graham | though Borland of course own Interbase | 11-Mar-08 21:24 |
| 271 | Graham | Can't buy FB ... it's not controlled by anyone | 11-Mar-08 21:24 |
| 270 | Graham | I don't have sql-server | 11-Mar-08 21:24 |
| 269 | Carl | Maybe next, Sun... no HP, will by FB for $1B. Lol. | 11-Mar-08 21:24 |
| 268 | Graham | Perhaps it's FB specific?? | 11-Mar-08 21:23 |
| 267 | Graham | Brian .. you able to confirm this bug? | 11-Mar-08 21:23 |
| 266 | Carl | You obviously love DB's, Pekr. | 11-Mar-08 21:23 |
| 265 | Pekr | hmm, my friend told me, FB is decent DB though. Dunno what quality it is .... | 11-Mar-08 21:23 |
| 264 | Carl | Anyway to conclude for today... 2.7.6 finalization is priority. | 11-Mar-08 21:23 |
| 263 | BrianH | ODBC and OLE DB _are_ the native drivers for MSSQL - there are no intermediate layers. | 11-Mar-08 21:22 |
| 262 | Graham | I've never looked at the sources | 11-Mar-08 21:22 |
| 261 | Carl | Lol, so... if I look at the source, will I vomit? ;) | 11-Mar-08 21:22 |
| 260 | Pekr | BrianH - is there native driver for MSSQL or are they mostly using ODBC? Can ODBC driver be installed and configure automatically? | 11-Mar-08 21:21 |
| 259 | Graham | No restrictions on the license | 11-Mar-08 21:21 |
| 258 | Graham | FB is totally free | 11-Mar-08 21:21 |
| 257 | Carl | B: So, RT cannot take the backend and use it for RIF, right? | 11-Mar-08 21:21 |
| 256 | Pekr | yes, ODBC ... | 11-Mar-08 21:20 |
| 255 | Carl | The fix is ODBC, so let's focus on that. | 11-Mar-08 21:20 |
| 254 | BrianH | Firebird is commercial open source based on formerly semi-open source from another product. There are limits. | 11-Mar-08 21:20 |
| 253 | Carl | G: Totally true. | 11-Mar-08 21:20 |
| 252 | Pekr | Anyway - interesting to hear there was FireBird driver, just not released. I remember Jeff Kreis did even BerkeleyDB wrapper. But that DB is strange too - only library wrapper, no TCP. The same goes for mySQL embedded ... | 11-Mar-08 21:20 |
| 251 | Graham | Anyway this surely is besides the point. | 11-Mar-08 21:20 |
| 250 | Carl | I'm actually a bit with Pekr on this one... it seems odd not to have a tcp interface these days. But, then, I'm not much of a DB guy. | 11-Mar-08 21:20 |
| 249 | Graham | well, if you can read the dll source ... | 11-Mar-08 21:19 |
| 248 | Pekr | Graham - so what are they thinking? Having server with undocumented TCP and library wrapper? Come on .... I know you can't choose DB sometimes ... | 11-Mar-08 21:19 |
| 247 | Graham | As I understand it, the tcp interface is not trivial. | 11-Mar-08 21:18 |
| 246 | Pekr | I know .... of course it is good to have every possible DB with REBOL. I just thought that in edge cases we should be fine with ODBC. The trouble is ODBC is Command and not available for REBOL Linux IIRC? | 11-Mar-08 21:18 |
| 245 | Graham | You're supposed to use either odbc, or the fbclient.dll ... the latter is also undocumented. | 11-Mar-08 21:18 |
| 244 | Carl | IMHO, a tcp interface is trivial, why would it be missing? | 11-Mar-08 21:18 |
| 243 | BrianH | In general, I tend to build servers rather than sell apps. Different needs. | 11-Mar-08 21:17 |
| 242 | Graham | tcp interface is not documented | 11-Mar-08 21:17 |
| 241 | Pekr | If not, that is what I don't like. | 11-Mar-08 21:17 |
| 240 | Graham | thisis a medical system I am using | 11-Mar-08 21:17 |
| 239 | Pekr | Is there TCP interface to FireBird? | 11-Mar-08 21:17 |
| 238 | Pekr | yes, it feels like old DOS aproach - shared DB on filesystem. Even while you use tcp, you pass filesystem path or so. It is often used with medical systems, restaurant systems, etc., as a first level escape from Delphi own DB etc. | 11-Mar-08 21:16 |
| 237 | Graham | postgresql is very large | 11-Mar-08 21:16 |
| 236 | Carl | Huh??? "The driver source is available under the Initial Developers Public Licence (IDPL), a variant of the InterBase Public Licence (IPL) that is a variant of the Mozilla Public License 1.1. " | 11-Mar-08 21:16 |
| 235 | Graham | pekr, licensing is very different | 11-Mar-08 21:16 |
| 234 | BrianH | I would go with PostgreSQL or SQL Server Express over Firebird, but that is more of a professional opinion. Too many restrictions. | 11-Mar-08 21:16 |
| 233 | Carl | "strangely install it" ? | 11-Mar-08 21:15 |
| 232 | Pekr | no. FireBird is special beast. As for me, I don't like it. You have to strangely install it. But I know why those ppl like it. You can install it with your app. I would go with mysql anyday. | 11-Mar-08 21:14 |
| 231 | BrianH | No, it's not smaller than SQLite. | 11-Mar-08 21:13 |
| 230 | Carl | ah. | 11-Mar-08 21:13 |
| 229 | BrianH | SQLite doesn't have ODBC support. | 11-Mar-08 21:13 |
| 228 | Graham | that's why I use it | 11-Mar-08 21:13 |
| 227 | Carl | Smaller than sqlite? | 11-Mar-08 21:13 |
| 226 | BrianH | You don't have a test VM that you can trash when needed? | 11-Mar-08 21:13 |
| 225 | Graham | Firebird is the smallest | 11-Mar-08 21:12 |
| 224 | Carl | Yes, but looking for minimal footprint. This is a dev system. I avoid loading anything on it... like the plague. | 11-Mar-08 21:12 |
| 223 | BrianH | Download SQL Server Express, Firebird, PostgreSQL, or whatever. Lots of free databases for Windows. | 11-Mar-08 21:11 |
| 222 | Graham | I can't see that any of those db drivers are free | 11-Mar-08 21:10 |
| 221 | Carl | The problem is, I need to be able to reproduce it here too. Or the test cycle will take forever. | 11-Mar-08 21:09 |
| 220 | Graham | 45 | 11-Mar-08 21:09 |
| 219 | BrianH | I can test with databases here. What is the problem, or tracker #? | 11-Mar-08 21:08 |
| 218 | Carl | (I'm trying to avoid booting our old db test machine-- which brings up 5 dbs at same time, including oracle, and is just way too slow.) | 11-Mar-08 21:08 |
| 217 | BrianH | I use odbc a lot. | 11-Mar-08 21:08 |
| 216 | Carl | So, of the 5 drivers XP comes with, do you happen to know if any of those db's are free with XP, and I can just test with those. | 11-Mar-08 21:08 |
| 215 | Carl | Ok looks like it. | 11-Mar-08 21:07 |
| 214 | Carl | Let me see if this XP has odbc... 1 min... | 11-Mar-08 21:06 |
| 213 | Graham | It doesn't on linux either. | 11-Mar-08 21:03 |
| 212 | Graham | odbcdriver is http://www.firebirdsql.org/index.php?op=files&id=odbc | 11-Mar-08 21:02 |
| 211 | BrianH | Not on Windows it doesn't. | 11-Mar-08 21:02 |
| 210 | Graham | this is windows | 11-Mar-08 21:02 |
| 209 | Carl | So, won't rewrite the OS kernel with dll patches? | 11-Mar-08 21:01 |
| 208 | Graham | straight forward | 11-Mar-08 21:01 |
| 207 | Graham | 2 mins | 11-Mar-08 21:01 |
| 206 | Carl | So, Firebird? What's the install like? | 11-Mar-08 21:00 |
| 205 | Graham | I didn't try any other db | 11-Mar-08 21:00 |
| 204 | Carl | That is, for all db? E.g. excel? | 11-Mar-08 21:00 |
| 203 | Carl | what db? | 11-Mar-08 20:59 |
| 202 | Graham | I'm curious .. what exactly was done with Interbase? | 11-Mar-08 20:59 |
| 201 | Graham | For me it's fairly critical | 11-Mar-08 20:58 |
| 200 | Carl | checking... | 11-Mar-08 20:58 |
| 199 | Carl | We can release 2.7.7 in a week, if we think it's critical. But, I cannot be the judge on that. You must be the judge. I don't use ODBC. | 11-Mar-08 20:58 |
| 198 | Graham | 3 lines | 11-Mar-08 20:58 |
| 197 | Graham | I posted the test code to the tracker | 11-Mar-08 20:58 |
| 196 | Carl | (I mean a few lines, not a page.) | 11-Mar-08 20:57 |
| 195 | Pekr | Graham - now as there is free library component, maybe someone will wrap some dll/so in the meantime .... | 11-Mar-08 20:57 |
| 194 | Carl | But, fix can be made in early 2.7.7 -- is there some test code? | 11-Mar-08 20:57 |
| 193 | Graham | I see. | 11-Mar-08 20:57 |
| 192 | Carl | We closed submissions for 2.7.6 a few days ago. | 11-Mar-08 20:57 |
| 191 | Graham | You saying it won't make 2.7.6 ? | 11-Mar-08 20:56 |
| 190 | Carl | The fix above. | 11-Mar-08 20:56 |
| 189 | Graham | huh? | 11-Mar-08 20:56 |
| 188 | Carl | Ok, so on ODBC, cannot be in 2.7.6. | 11-Mar-08 20:56 |
| 187 | Carl | G: never released. | 11-Mar-08 20:55 |
| 186 | BrianH | Through ODBC he does. | 11-Mar-08 20:55 |
| 185 | Graham | Carl, you never supported Interbase! | 11-Mar-08 20:55 |
| 184 | Carl | Ok. | 11-Mar-08 20:55 |
| 183 | Carl | Pekr, ah, Interbase -- I think we dropped that a few years ago. | 11-Mar-08 20:54 |
| 182 | Graham | this is RT's ODBC. | 11-Mar-08 20:54 |
| 181 | Carl | Ah, so his are not via odbc. Direct. | 11-Mar-08 20:53 |
| 180 | Pekr | RT's. Is there any Doc's version? I don't remember - he has mysql, postgres | 11-Mar-08 20:52 |
| 179 | BrianH | DocK has an ODBC scheme, or Firebird? I thought he just had MySQL and PostgreSQL schemes. | 11-Mar-08 20:52 |
| 178 | Carl | RT's or DocK's version? | 11-Mar-08 20:50 |
| 177 | Graham | Carl, is there any chance ODBC will be looked at? | 11-Mar-08 20:48 |
| 176 | Graham | This is with Firebird - can download the database from http://www.ibphoenix.com/main.nfs?a=ibphoenix&page=ibp_download_21 | 11-Mar-08 20:10 |
| 175 | Graham | Added as tracker item 45. | 11-Mar-08 20:01 |
| 174 | Graham | If I try and store other than non integer values into a float, it gets stored as null instead. | 11-Mar-08 20:01 |
| 173 | Graham | This is a nasty one ... I've been trying to get round a bug in wine, and I found a fault with the float datatype in odbc. | 11-Mar-08 20:00 |
| 172 | Oldes | 1. if the port is open, you cannot open it again 2. I don't know what's the problem, it's just don't play at all if you run the script again from console | 10-Mar-08 18:42 |
| 171 | Carl | That sound port stays open? | 10-Mar-08 17:45 |
| 170 | Carl | So, the bug is what? | 10-Mar-08 17:45 |
| 169 | Oldes | anyway.. I'm not able to run the script 2x to hear the sound.. I'm not author and I don't know how sound port works so I give up.. as there is no Rambo ticket, nobody need it now | 10-Mar-08 15:32 |
| 168 | Oldes | it's not a bug.. the port must be first closed | 10-Mar-08 15:27 |
| 167 | Oldes | I'm not sure if it's bug.. but if you run Testing/sound-test.r >> do %sound-test.r Script: "Untitled" (none) Play a ping sound... (escape) >> do %sound-test.r Script: "Untitled" (none) Play a ping sound... ** Access Error: Cannot open sound ** Near: sound-port: open sound:// insert sound-port >> | 10-Mar-08 15:25 |
| 166 | Gabriele | Ok. let me know if there are still problems with prot-http.r and i'll look into them. | 9-Mar-08 12:12 |
| 165 | Carl | (actually, cassette tape) | 8-Mar-08 17:29 |
| 164 | Carl | Gabriele: yes, eventually remembered both of those -- once the disk swapped those brain pages back in. ;) | 8-Mar-08 17:28 |
| 163 | Gabriele | Carl: i think we had agreed to default /seek to binary because newline conversion did not work correctly. | 8-Mar-08 11:00 |
| 162 | Gabriele | Carl, i think we had fixed #4039 last year. The initial fix I made had a problem since I added a refinement to open and the native code did not like that. So IIRC we had to decide whether you wanted to change the native code or use a different method in the mezz code. I think i had uploaded a new prot-http.r to workaround the native problem, but if not, I can do that. it's just a matter of removing that refinement and using a flag in port/locals. | 8-Mar-08 10:58 |
| 161 | BrianH | I like the test processor. | 8-Mar-08 7:28 |
| 160 | BrianH | Submitted as r2-mezz-series #387. | 8-Mar-08 7:22 |
| 159 | BrianH | I did port tests earlier, though it probably doesn't work that well with /direct ports. | 8-Mar-08 7:21 |
| 158 | BrianH | Fixed TAKE without /part - all tests pass, though there are no image or port tests. | 8-Mar-08 7:20 |
| 157 | BrianH | How should it behave? | 8-Mar-08 7:01 |
| 156 | BrianH | There are no tests for take on a port or image. | 8-Mar-08 7:01 |
| 155 | Carl | Yes, tomorrow there will be some for testing. | 8-Mar-08 7:01 |
| 154 | Graham | Are linux builds also being created for testing? | 8-Mar-08 7:00 |
| 153 | Carl | Thanks BrianH for your help on 2.7.6 and DevBase changes too. | 8-Mar-08 6:59 |
| 152 | Carl | I'll work on the Linux / OSX fixes next while hopefully a few other folks other than us try out the test version a bit more. | 8-Mar-08 6:58 |
| 151 | BrianH | Or I could fix it right now, based on the tests :) | 8-Mar-08 6:58 |
| 150 | Carl | Well, ok... we can let it simmer in the pot for several more hours and see what cooks up. | 8-Mar-08 6:57 |
| 149 | Carl | >> take d: copy "abc" == "a" | 8-Mar-08 6:55 |
| 148 | BrianH | I got it. I remember that part. | 8-Mar-08 6:55 |
| 147 | Carl | The tests are written such that you can pretty much copy and paste them into the console to investigate their failure reasons. | 8-Mar-08 6:55 |
| 146 | Carl | FAIL: [#"a" = take d: copy "abc"] | 8-Mar-08 6:54 |
| 145 | Carl | 52 Tests Evaluated 12 Tests failed: take-string take-string take-string take-string take-st ring take-string take-block take-block take-block take-block take-block take-block | 8-Mar-08 6:54 |
| 144 | BrianH | That reminds me, DevBase could use automated testing. | 8-Mar-08 6:51 |
| 143 | Carl | Uploading test driver... | 8-Mar-08 6:50 |
| 142 | Carl | misc-tests | 8-Mar-08 6:49 |
| 141 | Carl | Put TAKE test in tests folder. | 8-Mar-08 6:49 |
| 140 | BrianH | I mean, I wrote the others too but they were already tested. | 8-Mar-08 6:48 |
| 139 | BrianH | It's the only one I wrote from scratch, so tests for it would be nice. | 8-Mar-08 6:48 |
| 138 | Carl | Only for TAKE. | 8-Mar-08 6:47 |
| 137 | Carl | Let me check... | 8-Mar-08 6:45 |
| 136 | BrianH | I was having to do a lot of behavioral comparisons for TAKE, but there would be some areas that I missed. | 8-Mar-08 6:44 |
| 135 | BrianH | Do you have tests for the backported R3 functions that could also be backported? | 8-Mar-08 6:42 |
| 134 | Carl | Yes, feature. | 8-Mar-08 6:41 |
| 133 | BrianH | Yup. Perhaps we should update the tests, call it a feature. | 8-Mar-08 6:40 |
| 132 | Carl | Adding it to tracker here. | 8-Mar-08 6:40 |
| 131 | Carl | Note the response line. | 8-Mar-08 6:39 |
| 130 | Carl | http://www.rebol.net/cgi-bin/rambo.r?id=3598& | 8-Mar-08 6:39 |
| 129 | BrianH | I seem to recall the latter. | 8-Mar-08 6:39 |
| 128 | Carl | I think there are 2 likely possiblities: 1. Either a mode flag somewhere is messed up. 2. We removed STRING from /SEEK mode -- perhaps because it is problematic (you cannot seek AND at the same time ignore CR chars). | 8-Mar-08 6:38 |
| 127 | BrianH | Same in the pre-mezz-merge version. | 8-Mar-08 6:37 |
| 126 | Carl | All of the errors assume the port is STRING based. | 8-Mar-08 6:36 |
| 125 | BrianH | Same here. Going to try the pre-mezz version. | 8-Mar-08 6:36 |
| 124 | Carl | I noticed it because the REBOL test suite fails, showing 20 errors related to seek. | 8-Mar-08 6:36 |
| 123 | Carl | Result is binary. | 8-Mar-08 6:33 |
| 122 | Carl | write file: %tst.txt "abcdefg" port: open/seek file print string? probe copy port close port | 8-Mar-08 6:33 |
| 121 | Carl | Example: | 8-Mar-08 6:32 |
| 120 | BrianH | Please, demonstrate :) | 8-Mar-08 6:24 |
| 119 | Carl | I am seeing a strange problem with seeking in file I/O. | 8-Mar-08 6:23 |
| 118 | Carl | Pekr: No, in fact I have removed the black window now. There is a /show option. | 8-Mar-08 5:23 |
| 117 | Pekr | eh, surprised by #4 not working :-( The ugly black console window appearing with the call is pretty much disturbing, especially when used with View apps | 8-Mar-08 3:26 |
| 116 | Graham | This bug is on 2.7.5 but not earlier versions. | 8-Mar-08 2:52 |
| 115 | Graham | Ok, done. I've used your code to show the problem. | 8-Mar-08 2:50 |
| 114 | Carl | Graham, if you can add some sample code to #20, I'll take a closer look at it. | 8-Mar-08 2:44 |
| 113 | Brent | Carl's working on the pendings right now, and a couple have been surprises and probably won't make it in this round, but there is another release planned soon. | 8-Mar-08 2:44 |
| 112 | Brent | they're not in until it says built, though. | 8-Mar-08 2:43 |
| 111 | Brent | That's right, Graham. | 8-Mar-08 2:43 |
| 110 | Graham | In which case I'm happy as 3/4 of the bugs I entered are pending. | 8-Mar-08 2:43 |
| 109 | Graham | Pending means it's being worked on for build ? | 8-Mar-08 2:42 |
| 108 | Carl | Ok, #4 has been implemented, however... there appear to be some other major problems in CALL. | 8-Mar-08 2:40 |
| 107 | Carl | Ok, found the problem. The console pops up because of an AllocConsole() call. | 8-Mar-08 2:24 |
| 106 | Carl | For #4: the posted fix does not work. Setting wShowWindow = SW_HIDE does not hide the shell window. | 8-Mar-08 2:19 |
| 105 | Carl | If I recall correctly, I think it created a new bug, and was withdrawn from the last test build. | 8-Mar-08 2:05 |
| 104 | Carl | It is not clear to me the status of RAMBO #4039 http://www.rebol.net/cgi-bin/rambo.r?id=4039 It is not listed here, but I see some code related to it in my current build files. I need confirmation regarding status. I think Gabriele was working on it. | 8-Mar-08 2:05 |
| 103 | Carl | Someone needs to provide an example or test code for #20. This is a way to help. If not, I don't think it will be fixed in 2.7.6. | 8-Mar-08 1:59 |
| 102 | Carl | This bug is something that 3.0 took special care to fix: >> p: 20x20 == 20x20 >> p/x: p/y: 10 == 10 >> p == 10x10 | 8-Mar-08 1:56 |
| 101 | Carl | This bug cannot be fixed for 2.7.6 -- it has major complexities associated with it. | 8-Mar-08 1:55 |
| 100 | Henrik | >> p: 20x20 == 20x20 >> p/x: p/y: 10 == 10 >> p == 10x20 | 8-Mar-08 0:17 |
| 99 | Henrik | I added #3389, path evaluation problem. | 8-Mar-08 0:16 |
| 98 | BrianH | It means that a a change in the doc string of the query function should be requested as a bug report. | 6-Mar-08 17:57 |
| 97 | Paul | So does that mean would should add 'query to the wish list or its functionality as it currently is? | 6-Mar-08 17:53 |
| 96 | BrianH | I miss librebol.so - it was fun. Too bad it didn't work that well. The fast-cgi support in /Command was better. | 6-Mar-08 14:47 |
| 95 | Henrik | perhaps some time later it can be useful in R3 for general detection of changes in objects. | 6-Mar-08 14:24 |
| 94 | Gabriele | so, consider it an accident that you can still use it. it comes useful in some cases though :) just don't rely on it for future compatibility. | 6-Mar-08 14:13 |
| 93 | Gabriele | Carl was not even aware it was still there, last time I talked to him about it :) | 6-Mar-08 14:13 |
| 92 | Gabriele | QUERY on objects was implemented for the librebol.so Apache module, as it was necessary for its functionality. since that project never saw the light of the day, the functionality was never documented. | 6-Mar-08 14:12 |
| 91 | Paul | Yeah seems to be a very useful function. Probably the reason it hasn't been utilized more is the lack of documentation regarding it. | 6-Mar-08 14:05 |
| 90 | Henrik | too bad | 6-Mar-08 13:54 |
| 89 | Anton | Henrik, wrt QUERY,I noticed that the View system was clearing the "changed" status of an object, so you couldn't really use it with faces. | 6-Mar-08 13:51 |
| 88 | Henrik | I see that in R3, object! as an argument has been removed. | 6-Mar-08 11:09 |
| 87 | Henrik | Here's a quickie and correct me if I'm wrong: I've noticed the QUERY function before, but never realized it could be used for detecting changes in an ordinary object. This is due to its less than useful help text. I understand that it's mainly used for ports, but would you use it generally for detecting changes in any objects or are there caveats to this that would explain the "misleading" help text? | 6-Mar-08 11:01 |
| 86 | Gabriele | #4319 | 5-Mar-08 21:30 |
| 85 | Henrik | not until it's accepted | 5-Mar-08 18:27 |
| 84 | Paul | Yeah no dump using integer. | 5-Mar-08 18:27 |
| 83 | BTiffin | tracked it. But I didn't have a rambo # | 5-Mar-08 18:27 |
| 82 | BTiffin | Yeah, no dump? I tested a few types. I RAMBOed. | 5-Mar-08 18:22 |
| 81 | Henrik | after RAMBO it could be put in here. | 5-Mar-08 18:22 |
| 80 | Paul | Brian you need to submit that to RAMBO as it might not get fixed in this next release. | 5-Mar-08 18:20 |
| 79 | Paul | reduce/only 1 [] | 5-Mar-08 18:19 |
| 78 | Paul | Crashes a bit differently if you do: | 5-Mar-08 18:19 |
| 77 | Paul | XP | 5-Mar-08 18:16 |
| 76 | Paul | Same here Brian. | 5-Mar-08 18:16 |
| 75 | BTiffin | Don't know if this is new; reduce/only "string" [] dumps and crashes. Any non block data with an empty set. | 5-Mar-08 18:03 |
| 74 | Gabriele | btw, is tls:// supposed to work with /Command ? | 5-Mar-08 17:26 |
| 73 | Henrik | I wonder if that clipboard is depending on the window manager used, i.e. if it's available with really old WMs like TWM.. | 5-Mar-08 12:26 |
| 72 | Henrik | That's bug #4318 | 5-Mar-08 12:21 |
| 71 | Henrik | Carls' blog has a comment on that it would be a good idea to use the "real clipboard" in X11 instead of the selection clipboard. Don't know how much code it would take to fix that. | 5-Mar-08 12:20 |
| 70 | BrianH | I had been looking for that earlier :) I put the best parts of #21 into #4 and deleted #21. | 3-Mar-08 20:52 |
| 69 | Pekr | Brain, your input on console Window is the same as number 4. But your is more descriptive, so delete mine, if you want to ... | 3-Mar-08 20:49 |
| 68 | BrianH | Done. | 3-Mar-08 20:14 |
| 67 | BrianH | Wait, they just discussed it. Fixing... | 3-Mar-08 20:07 |
| 66 | BrianH | Yes. | 3-Mar-08 20:06 |
| 65 | Graham | Has someone listed the black screen when doing call ? | 3-Mar-08 19:54 |
| 64 | Graham | Listed my favorite 3 bugs - all under linux: call, vectorial fonts and function keys | 3-Mar-08 19:53 |
| 63 | BrianH | The install bug is a Vista-compatibility issue as well, since the rules that were being broken in 2000 are being enforced in Vista now. | 3-Mar-08 14:48 |
| 62 | Gabriele | About the CALL linux bug: there is a task for you on Qtask with all the info I have about it. let me know if I should copy that here. | 3-Mar-08 14:46 |
| 61 | BrianH | Or my report from after View 1.3.0 came out, after 3562 was supposed to solve the problem but didn't go far enough. | 3-Mar-08 14:44 |
| 60 | BrianH | If you like, I can dig out my bug report from 2000 and post it here. | 3-Mar-08 14:41 |
| 59 | BrianH | RAMBO #s 3562, 3775, 4236 (that I can find) relate to the installation bug. 3562 is marked as built, but still isn't fixed. | 3-Mar-08 14:38 |
| 58 | Henrik | >> 2147483648
== 2147483648.0
Probably not correct behavior in that negate above. | 3-Mar-08 14:35 |
| 57 | Henrik | damn, stupid OSX paste :-/ oh well. this one number causes math overflow. if you use negate -2147483649 it turns into decimal. is this correct behavior? | 3-Mar-08 14:33 |
| 56 | Henrik | >> negate -2147483648 ** Math Error: Math or number overflow ** Near: negate -2147483648 | 3-Mar-08 14:32 |
| 55 | BrianH | There are installation bugs that go back to 2000. They keep being marked as low priority, but are still a problem. Can we finally fix this? | 3-Mar-08 14:31 |
| 54 | Paul | >> abs -2147483648 == -2147483648 | 3-Mar-08 14:29 |
| 53 | Paul | I noticed that absolute is broke on big negative numbers if that helps pinpoint the problem. | 3-Mar-08 14:29 |
| 52 | Paul | And at least you get some form of error message to target where the error is. | 3-Mar-08 14:26 |
| 51 | BrianH | 4234 and 4233 seem like they are trying to change the value to a positive number before use. | 3-Mar-08 14:26 |
| 50 | Paul | I'm more concerned about the ones crashing the interpreter instead of the overflow ones. There could be workarounds for those. | 3-Mar-08 14:26 |
| 49 | Henrik | there are at least 10 bugs related to large number crashes. | 3-Mar-08 14:25 |
| 48 | Paul | 4234 seems to be possibly related to large numbers? | 3-Mar-08 14:24 |
| 47 | BrianH | Nevermind, I misread 4226 - it should probably be done too. | 3-Mar-08 14:23 |
| 46 | Paul | oops sorry bout that | 3-Mar-08 14:21 |
| 45 | Paul | : blk: at [1 2 3 4 5] 3 remove/part blk -2147483648 | 3-Mar-08 14:21 |
| 44 | BrianH | There's a whole series of math overflow bugs around that area of RAMBO. Though 4226 is off, the rest should be tested. | 3-Mar-08 14:21 |
| 43 | Paul | AppName: rebview.exe AppVer: 2.7.5.3 ModName: rebview.exe ModVer: 2.7.5.3 Offset: 00064c88 | 3-Mar-08 14:20 |
| 42 | Paul | 4229 crashed the interpreter here on XP also with debug info as: | 3-Mar-08 14:20 |
| 41 | Henrik | 4234 gives no problems on WinXP and OSX | 3-Mar-08 14:20 |
| 40 | Henrik | same here. Just returns a good error in OSX. | 3-Mar-08 14:19 |
| 39 | Paul | Gives a REBOL internal error: Corrupt datatype: 200 at 201 | 3-Mar-08 14:19 |
| 38 | Paul | #4233 crashed on XP here | 3-Mar-08 14:18 |