00:00:00 --- log: started retro/10.01.07 00:16:44 Wow, can't believe I left that swap in. Yet even more still betterer string compare... 00:16:57 : comp ( $$-f ) 00:16:57 [ here ] 00:16:57 dup @ push 1+ swap 00:16:57 dup @ push 1+ pop dup pop 00:16:57 !if drop drop dup xor ;; then 00:16:57 0 [ 12 , , ] 00:16:59 drop drop -1 00:17:01 ; 00:20:53 [js4/retro10] 16e58e: Remove useless swap in 'compare' 00:21:21 "swap !if" is always wrong. 03:16:53 [crcx/retro10] 5026c6: Remove useless swap in 'compare' 03:24:06 --- join: Mat2 (i=5b43e0e4@gateway/web/freenode/x-clagdvgyehqzkadp) joined #retro 03:24:12 G'Day 03:25:03 --- part: Mat2 left #retro 03:26:17 crc: crcx: Are you there? 03:26:45 yes 03:26:49 getting ready for work 03:27:15 I have a github question, hopefully it's easy... 03:27:32 I have two changes from you in red on the fork queue for retro10 03:27:44 I checked my source and they have not been applied. 03:28:02 I have no changes that should stop it from being applied... I think. 03:28:16 How do I find out why it's red, how do I force it to apply? 03:28:30 you can try applying it, if there's a problem, it'll let you know 03:28:39 or you can do it manually 03:28:45 git pull git://github.com/crcx/retro10.git master 03:28:57 then "git checkout " for any files with conflicts 03:29:08 ok, cool thanks. 03:29:18 and then add/commit the changes (only needed if there are conflicts) 03:39:19 oh... that's worse... now I have added a conflict with myself. I must be doing something wrong. 03:40:09 And this is probably a clue of what: 03:40:11 $ git status 03:40:11 # On branch master 03:40:11 # Your branch is ahead of 'origin/master' by 18 commits. 03:40:11 # 03:40:11 nothing to commit (working directory clean) 03:41:02 --- quit: virl (Remote closed the connection) 03:56:22 --- join: crcx (i=d8012b82@gateway/web/freenode/x-fufpttxjjwcilprt) joined #retro 04:13:36 * crcx wonders why your repo is out of sync... 04:16:15 with a fresh clone of your retro10 from github I get a conflict with 'core.retro' when pulling from mine 04:23:31 js4: do fresh clone of your retro10 directory 04:23:35 js4: then: 04:23:37 git pull git://github.com/crcx/retro10.git master 04:23:44 git checkout source/core.retro 04:23:52 git add source/core.retro 04:24:09 git commit -m 'pull from crc' 04:24:11 git push 04:24:19 and you should be clean 04:54:20 --- join: Mat2 (i=5b43e0e4@gateway/web/freenode/x-hqajcvxrsywhpzut) joined #retro 04:54:58 crcx, are you there ? 04:55:38 yes 04:56:20 fine, have you take a look at the avm ? 04:56:42 only briefly :( 04:57:18 ah ok, it would interrest me if you find some ideas useful 04:58:50 if you have questions, ask, I donj't had the time to full documentate the instruction set 04:58:57 ok 04:59:21 I won't be able to look at it in detail until this weekend 04:59:26 ok 04:59:38 see you 05:00:30 (hmm, probably the wrong sentece in english, sorry -> poor english speaker) 05:00:47 ciao 05:00:54 --- part: Mat2 left #retro 05:09:48 crcx: What is a "fresh clone"? Delete the old directory the do git clone? 05:11:56 yes 05:12:02 or rename the old directory 05:12:28 * crcx will be applying your change to handle backspace (127) under windows today 05:12:44 $ git clone git@github.com:js4/retro10.git 05:12:45 Initialized empty Git repository in /home/Jay/src/forth/retro/git/retro10/.git/ 05:12:45 Enter passphrase for key '/home/Jay/.ssh/id_rsa': 05:12:45 remote: Counting objects: 2550, done. 05:12:45 objects: 100% (985/985), done. 05:12:45 fatal: The remote end hung up unexpectedly0 KiB | 231 KiB/s 05:12:47 fatal: early EOF 05:12:49 fatal: index-pack failed 05:12:51 eep! 05:15:38 4th time worked. 05:15:58 github may be having server issues 05:16:19 * crcx cloned yours a little while ago and it worked fine for me 05:18:06 Jay@black ~/src/forth/retro/git/retro10 05:18:06 $ git pull git://github.com/crcx/retro10.git master 05:18:06 remote: Counting objects: 87, done. 05:18:06 remote: Compressing objects: 100% (75/75), done. 05:18:06 7remote: Total 76 % (6(delta 51), reused 0 (delta 0) 05:18:06 Unpacking objects: 100% (76/76), done. 05:18:08 From git://github.com/crcx/retro10 05:18:10 * branch master -> FETCH_HEAD 05:18:12 Auto-merging source/core.retro 05:18:14 CONFLICT (content): Merge conflict in source/core.retro 05:18:16 Auto-merging source/meta.retro 05:18:18 Automatic merge failed; fix conflicts and then commit the result. 05:18:20 05:18:22 Jay@black ~/src/forth/retro/git/retro10 05:18:24 $ git checkout source/core.retro 05:18:26 error: path 'source/core.retro' is unmerged 05:19:20 --- quit: crcx (Ping timeout: 180 seconds) 05:29:54 [js4/retro10] 878d21: add code style notes 05:30:05 --- join: crcx (i=d8012b82@gateway/web/freenode/x-alhipzkgwprjrrrc) joined #retro 05:31:04 I think it's fixed now. I have no idea what I did wrong though. 05:56:54 [crcx/retro10] 9e6e84: remap key 127 to key 8 (backspace fix for windows) 06:57:00 [crcx/retro10] 15e8a8: cleaner tempString, keepString, and $ 07:13:04 don't pull yet. seems to be some breakage :( 07:16:48 I think I know what got github/git confused. I applied a bunch of your changes. but there are two I made that you never did. the 127 remap and the ^H ^H in forth. 07:18:24 well, I'll bbiab. Have fun! 07:48:53 [crcx/retro10] 19ae6d: reduce return stack depth a bit with selective tai... 07:49:24 it's safe to pull now; bugs from string words have been fixed 08:02:57 --- quit: crcx ("Page closed") 08:54:27 --- join: crcx (i=d8012b82@gateway/web/freenode/x-utevpracffnuenkc) joined #retro 13:41:56 --- quit: crcx ("Page closed") 15:10:26 --- join: crcx (i=d8012b82@gateway/web/freenode/x-qeiyelqriiuaprqj) joined #retro 15:27:37 --- join: virl (n=virl__@chello062178085149.1.12.vie.surfer.at) joined #retro 15:28:52 --- join: erider (n=chatzill@unaffiliated/erider) joined #retro 15:50:24 --- quit: crcx ("Page closed") 16:07:40 hi 16:07:48 hi 16:15:32 hi 16:38:17 what is new 16:49:08 * crc is cleaning up small things in retro 17:04:50 the core is down to 2400 cells after a rewrite of several string handling words 17:12:49 nice! 17:18:56 cool 17:21:01 [crcx/retro10] 22edb3: match indention with surrounding code [core] 17:27:53 [crcx/retro10] a38fae: improvements to string->number code 17:28:54 a38fae reduces core size by 22 cells 17:37:54 [crcx/retro10] ecff06: modify default "word not found" error message 17:39:53 [crcx/retro10] cbf66b: eliminate an unused variable 17:53:01 chao 17:53:08 --- quit: erider ("ChatZilla 0.9.86 [Firefox 3.5.7/20091221164558]") 19:13:56 [crcx/retro10] 30be8e: rewrote "accept" (and related words) to eliminate ... 19:14:25 core is 2363 cells with this latest patch 19:15:39 at this point, the buffers (scratch, tib) can be altered to reduce the minimal image size by 512 cells. Should I do this? 19:26:37 I'm not sure. 19:27:16 Are we trying to get it under a certain size? 19:27:24 not really 19:27:41 total size is capped at 4k. including the 1k of buffer space 19:27:55 Niice 19:28:25 my main goal is just to clean up the code in core.retro and make small improvements to overall performance while doing so 19:29:47 Why does the metacompiler code look so different from the regular forth code? 19:30:00 all the extra punctuation, etc 19:31:27 because metacompilers are tricky :( 19:31:45 it's an adaption of the original source, which was cross-compiled to ngaro 19:31:50 first using retro 9.x 19:31:52 then toka 19:31:56 and now metacompiled 19:32:14 the same core.retro has been moved across the various hosts with only minor changes 19:32:30 though it is ugly to read 19:33:24 (basically, the core is written in a ngaro assembly language, with the metacompiler being the assembler. a few "macros" exist in the metacompiler to make it feel more like forth) 19:33:49 interesting 19:35:22 m, is an interesting word 19:36:10 basically data stack->target, just as , is data stack -> heap 19:36:35 yup 19:39:29 and t: is : with target instead of heap... and so forth I presume 19:39:34 yes 19:40:22 most words ending in , are vm assembley aliases, right? 19:40:47 at least the ones listed as n vm: word, 19:40:55 yes 20:38:44 cool, it only emits a ? instead of "Word not found" 20:38:55 * docl prefers it this way :) 20:41:02 more concise 20:41:12 and it tells you what word it couldn't id 20:41:27 cool 20:44:44 ah, the new (accept) inputs directly to the tib pointer instead of returning a char to be placed there. 20:45:11 yup 20:45:29 * crc likes this new 'accept' and friends better 20:46:10 eat-leading takes the space? 20:46:34 yes 20:47:05 eat-leading consumes input until something other than the delimiter character is encountered 20:47:36 ahhh 20:47:54 a muncher :) 20:50:52 ( ))))))))) test ) 20:52:53 nifty 21:31:16 this is very frustrating. I can't figure out how to tell github "throw away everything I did, make a new repository based on crc's current code" 21:32:39 my attempts usually involve git checkout, git pull, git pull (link) master, possibly rm -rf 21:33:27 1. delete current working dir and use git clone to get a fresh copy 21:33:43 2. git pull (crcx link) master 21:33:59 usually that will do the trick 21:35:12 another one is copy each file from a clone of the crcx repo and push them to your repo (starting with a fresh clone of it) 21:35:22 er commit then push 21:43:51 I diffed my retro10 against one I cloded fromt he public crc link. No differences. but my retro10 queue shows a buncha red patches that won't apply cleanly. 21:45:09 Though "This data is out of date. We are showing you a cached version. Please refresh in a minute for more up to date data." and even after many minutes is still out of date... 21:48:30 did you try starting with fresh clones of each repo and copying the files across, then committing that to your repo? 21:49:12 I think so. Cept I didn't bother copying since the files are identical. 21:53:03 docl: what do you do to get those "merge" check ins? 21:53:30 1. git pull git://github.com/crcx/retro10 master 21:53:35 2. git push 21:53:54 but that only works if it doesn't have a conflict... 21:56:40 well... maybe it's just the "out of date" problem... maybe it will go away in a while... 22:40:22 http://gist.github.com/271895 22:42:49 now with stack diagrams 22:45:05 oops, that is not going to be reliable with the variable being set like that 22:48:27 there we go 22:50:54 variable name makes more sense 23:08:32 and now much more readable 23:25:37 can you make "vocab foo (( ... ))" work? 23:25:53 I am sugesting a different syntax slightly, same behaviors 23:26:52 I don't know the vocabulary context search order words in retro at all 23:27:04 I don't know (( and )) either 23:35:10 I think you can if you make )) manipulate the dictionary order. It moves the entry for the item foo from before the words in to, to after the words in foo by messing with link fields. 23:36:29 And you wouldn't need a hiden variable, the new word created for the vocabulary would already have all 3 fields allocated... anyway, just an idea. 23:38:26 how does forget foo behave? 23:59:59 --- log: ended retro/10.01.07