| 00:28.19 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/ws.sh: Remove the Done |
| 00:30.13 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/ws.sh: this script doesn't do indentation, there are better tools for that job (e.g. brl-cad's indent.sh emacs hook script). |
| 00:32.09 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/regress/ (6 files): ws, test two |
| 00:45.56 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/ws.sh: too many changes by expanding internal tabs for my comfort zone, disable it by default for now. causes massive changes as brl-cad extensively uses tabs atm. (not visual, but massive nonetheless) |
| 01:02.14 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/include/ (93 files): ws, test three |
| 01:08.01 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/ (NEWS HACKING INSTALL AUTHORS BUGS configure.ac): ws, last test..looking good ;) |
| 01:12.05 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/ws.sh: ignore our own files |
| 01:15.52 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/ws.sh: obey WS_PROGRESS |
| 01:31.44 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/ws.sh: run ourself on ourself |
| 01:33.53 | brlcad | *boooom* |
| 01:40.02 | ``Erik | I presume "whitespace" means "tab and space", not the usual compiler/parser notion of "tab, space, and newline" |
| 01:42.02 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/ (1084 files in 111 dirs): ws. lots and lots of ws. see sh/ws.sh for details (cases 'abcdeg'). |
| 01:42.44 | ``Erik | perl where sed would've done? |
| 01:51.07 | brlcad | still not completely common to have a sed that supports in-place-editing, and full-file (-0777) matches are royal pita |
| 01:52.42 | ``Erik | yeah, linux lacks on the sed... the "mv sed rm" loop is lame |
| 01:52.51 | brlcad | futzing with sed and redirects, and my pedantic check of file inputs/outputs before and after would have just added bloat |
| 01:53.25 | brlcad | i had to add one in there for 'expand' and didn't really want to (though it in the end, that's the case that's disabled |
| 01:54.01 | *** join/#brlcad digitalfredy (n=digitalf@200.71.62.161) | |
| 01:55.36 | brlcad | that probably cut the line count down by a few hundred/thousand at least, no more spans of blank lines (bob had some tcl files with dozens at the end of file) |
| 01:55.45 | brlcad | at least no more than two now |
| 01:56.25 | brlcad | particularly trailing newlines after the footer.. that's just annoying |
| 01:58.48 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/ws.sh: s/one lines/on lines/ typo |
| 01:59.46 | ``Erik | hrmmmm |
| 02:00.37 | ``Erik | 'portlint' complains if there is more than one contiguous blank line... is that a feature we should concern ourselves with? does it ever make sense to have three consecutive newline characters? |
| 02:01.44 | brlcad | yeah, i can see the case for distinguishing between one and two .. separate out sections in file, each section might have breaks of one |
| 02:02.09 | brlcad | that's got to be configurable with portlint I'd imagine too |
| 02:02.24 | ``Erik | no, portlint is insanely anal on things like that |
| 02:02.33 | ``Erik | it's a pretty basic perl script last I looked |
| 02:02.52 | ``Erik | sections are seperated by exactly one blank line. necessarily. |
| 02:03.01 | brlcad | don't see why we'd need to care really |
| 02:03.07 | ``Erik | aight *shrug* |
| 02:03.29 | ``Erik | just throwing it out there :) |
| 02:03.31 | brlcad | i mean it does make sense to allow a couple, denotes a break more than one level |
| 02:04.00 | brlcad | at least, I see a use for it |
| 02:04.07 | ``Erik | probably be better to do a quick&dirty grep type search and manually fix it, anyways |
| 02:04.12 | brlcad | now the difference between 3 and 4 or 3 and 8, not so much |
| 02:04.15 | ``Erik | if apporpriate |
| 02:05.00 | brlcad | heh |
| 02:05.04 | brlcad | "why" |
| 02:05.16 | ``Erik | page break... to seperate high level sections... |
| 02:05.25 | brlcad | i know, but .. "why" |
| 02:05.26 | ``Erik | opposed to blank line, then /********/ |
| 02:06.03 | brlcad | mostly only useful if you shovel the file off to a printer |
| 02:06.06 | ``Erik | (also; shows up a pretty dark blue in vim) |
| 02:06.08 | ``Erik | yeah |
| 02:06.16 | ``Erik | *shrug* |
| 02:06.21 | brlcad | most of the time, you're just looking at stupid ^L's everywhere :) |
| 02:06.31 | ``Erik | pretty dark blue ^L's |
| 02:06.33 | brlcad | hehe |
| 02:07.08 | ``Erik | nonprintable characters are colored in vim, and the default color theme doens't use that color for anything else :) |
| 02:07.48 | brlcad | hm, the addition of opennurbs caused by 1.5 hour G4 compile to go to about 3 hours now |
| 02:08.36 | ``Erik | with wc.sh; you do progress for every single file? do any take long enough to worry about that? would it be worth building the perl program in a string, then executing perl once to lower the startup/teardown overhead? |
| 02:08.54 | brlcad | nah, the whole thing flies |
| 02:09.31 | brlcad | processed all of the sources and other files in like two minutes |
| 02:09.47 | brlcad | which is like .1sec/file |
| 02:10.06 | ``Erik | hm, seems like a better progress indicator would be to make total % done after each file, then collapse the perl |
| 02:10.43 | brlcad | meh, cleaning up ws is just a helper task .. and now it's done.. |
| 02:11.41 | brlcad | it's enough wrapper around 7 lines of actual work as it is |
| 02:12.03 | brlcad | the only nasty part is the collapse of whitespace at the beginning of a line |
| 02:12.45 | brlcad | could not for the life of me get the regexp to behave without iterating.. so it iterates -- and collapses one each iteration |
| 02:13.22 | brlcad | annoying, but again, meh it's done and did the job well |
| 02:13.46 | ``Erik | hum |
| 02:13.50 | ``Erik | ummm |
| 02:14.03 | ``Erik | did you give it the /e flag at the end to tell it extended regex? |
| 02:14.35 | ``Erik | (and, actually, your replacement string can be a perl program...) |
| 02:15.15 | ``Erik | the only perl in my toolbox these days is: |
| 02:15.16 | ``Erik | perl -pi -e 's/%([0-9A-Z][0-9A-Z])/chr(hex($1))/eig' |
| 02:16.56 | brlcad | i know, I tried a variety of junk |
| 02:17.35 | ``Erik | *shrug* if it works, it works :) hopefully, we'll all have enough discipline to never need it again... |
| 02:17.39 | brlcad | still couldn't get iteration without explicitly iterating .. something non-greedy that rematched on a global |
| 02:17.47 | brlcad | heh, riiiight |
| 02:18.16 | brlcad | it probably should be set up to run weekly |
| 02:18.49 | brlcad | later though, after the website (and after ef) |
| 02:35.43 | ``Erik | if run weekly, should it attempt to commit, or should it send an email reporting its 'findings' so a human can verify and commit? |
| 03:26.40 | *** join/#brlcad IriX64 (n=IriX64@bas3-sudbury98-1168057688.dsl.bell.ca) | |
| 03:29.23 | *** join/#brlcad digitalfredy (n=digitalf@200.71.62.161) | |
| 04:51.15 | brlcad | for that kind of script, probably just commit .. for others, depends |
| 06:21.22 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/ws.sh: no need to find configure.ac -- just let it use the current directory if no arguments are specified |
| 07:00.34 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/indent.sh: no need to find configure.ac -- just let it use the current directory if no arguments are specified. the helper script can still be searched for as needed. |
| 07:05.10 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/copyright.sh: no need to find configure.ac -- just let it use the current directory if no arguments are specified. |
| 07:17.27 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/Makefile.am: it's a nifty generic script, but there's still no real pressing need or justification to install it. |
| 07:42.06 | *** join/#brlcad SWPadnos (n=Me@dsl245.esjtvtli.sover.net) | |
| 07:57.44 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/header.sh: should not be using the gpl or gfdl any more |
| 09:05.43 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/header.sh: oop another ref to gpl/gfdl removed. |
| 09:21.43 | *** join/#brlcad clock_ (i=clock@84-72-63-118.dclient.hispeed.ch) | |
| 10:31.37 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/sh/ (template.sh Makefile.am): |
| 10:31.37 | CIA-5 | BRL-CAD: initial addition of this file template script. it generates an empty template |
| 10:31.37 | CIA-5 | BRL-CAD: file with a standard header and footer based on the provided license and file |
| 10:31.37 | CIA-5 | BRL-CAD: type. basically, it's a warpper on the header.sh and footer.sh script along |
| 10:31.37 | CIA-5 | BRL-CAD: with other clean-up steps generally needed. add script to dist. |
| 10:33.37 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/src/mged/chgmodel.c: fix comment, mirror commands, not m command |
| 12:06.39 | *** join/#brlcad digitalfredy (n=digitalf@200.71.62.161) | |
| 12:30.53 | *** join/#brlcad digitalfredy (n=digitalf@200.71.62.161) | |
| 18:36.36 | *** join/#brlcad clock_ (i=clock@84-72-63-118.dclient.hispeed.ch) | |
| 18:58.36 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/src/mged/chgmodel.c: check for valid access before looking up the object |
| 21:43.31 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/ (6 files in 4 dirs): remove the tolerance parameter from nmg_invert_shell. it's not used, contrary to the commented purpose, and just complicates the interface unnecessarily so remove it. |
| 22:22.26 | *** join/#brlcad clock_ (i=clock@84-72-63-118.dclient.hispeed.ch) | |
| 22:29.50 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/src/librt/mirror.c: |
| 22:29.51 | CIA-5 | BRL-CAD: add a new rt_mirror() routine for mirroring an object about a specified axis. |
| 22:29.51 | CIA-5 | BRL-CAD: the routine is derived from mged's f_mirror command hook and is one step closer |
| 22:29.51 | CIA-5 | BRL-CAD: towards getting rid of the switch on primitive types. routine takes a database |
| 22:29.51 | CIA-5 | BRL-CAD: instance, a from and to object name, and the axis and returns a directory |
| 22:29.53 | CIA-5 | BRL-CAD: pointer to the new mirror object. |
| 22:30.34 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/include/raytrace.h: add the new rt_mirror func |
| 22:30.44 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/src/librt/Makefile.am: add new mirror.c file |
| 22:32.02 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/src/mged/chgmodel.c: use the new rt_mirror() function for the guts to f_mirror(). the latter is now just a wrapper on the tcl/mged command interface semantics. part of general effort to move mged's editing facilities into librt. |
| 22:39.13 | CIA-5 | BRL-CAD: 03brlcad * 10brlcad/src/mged/clone.c: one step closer, stub out the matrix transformation on solids |