| 02:15.51 | brlcad | it's not so much getting windows to behave like unix, it's actually a recommended practice on windows for security, to override the default searching to a specific dir |
| 02:16.51 | brlcad | it's just not the default, so there's a little learning curve |
| 02:18.13 | brlcad | the only reason it's not common is because it's not the default, but most of the notable apps do this practice for safe library loading |
| 02:19.15 | brlcad | one of the first links I came across were the bug reports for iTunes initially using the default, but then switching to using SetDllDirectory() |
| 02:19.54 | brlcad | so really the only unixy aspect of this is the name of the dir itself ('lib'), which is irrelevant |
| 02:21.55 | brlcad | starseeker: that all said, it's not worth spending days on figuring out the bootstrap so that aspect can be back-burnered if simply linking it in with delayload didn't work |
| 02:22.17 | brlcad | I presume you got a dll-could-not-be-found error message of some sort? or was it some other failure? |
| 02:23.08 | brlcad | could be that "if windows" hacks have to be removed for it to work, like what bob and I saw yesterday with archer |
| 02:28.30 | brlcad | if we can't get it to work, the next step will probably be deprecating bu_brlcad_root and adding bu_brlcad_bin+bu_brlcad_lib then just making windows return bin for lib |
| 02:29.38 | brlcad | (or, rather, having cmake set bin as the libdir then everything else falling through so libbu doesn't have to have a windows conditional) |
| 02:31.22 | brlcad | if that code has to be hit up again, it'd be a good time to make it not be brlcad-specific api like it is now |
| 02:31.45 | brlcad | but then that's more work :) |
| 02:32.20 | brlcad | gets buzy |
| 02:51.57 | starseeker | brlcad: yeah, libged.dll wasn't found (iirc) |
| 02:52.31 | starseeker | it's quite possible I didn't hook it in quite right - MSVC isn't my strong suite |
| 02:52.46 | starseeker | probably worth taking a quick look at next week |
| 02:55.12 | CIA-48 | BRL-CAD: 03starseeker * r49133 10/brlcad/trunk/CMakeLists.txt: For now, take out the dll relocation - if someone wants to do testing or we put it back in 'live', all that should be needed is a reverse merge on this commit. Until we're ready, leave Windows in a working state. |
| 04:32.58 | brlcad | k, i'll take a looksee next week |
| 05:59.18 | *** join/#brlcad milamber (~devlin@d118-75-244-176.try.wideopenwest.com) |
| 07:23.21 | *** join/#brlcad Technicus (~Technicus@DSLPool-net208-2.wctc.net) |
| 12:17.29 | *** join/#brlcad merzo (~merzo@211-77-200-46.pool.ukrtel.net) |
| 14:09.30 | *** join/#brlcad jordisayol (~jordisayo@unaffiliated/jordisayol) |
| 14:11.46 | *** join/#brlcad Stattrav (~Stattrav@61.12.114.82) |
| 14:11.47 | *** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav) |
| 16:30.12 | starseeker | brlcad: was your scheme for implementing the equalivent of getopt_long (or something better) documented anywhere? |
| 19:08.31 | *** join/#brlcad merzo (~merzo@56-68-132-95.pool.ukrtel.net) |
| 22:45.15 | *** join/#brlcad packrat (~packrator@c-98-209-146-133.hsd1.mi.comcast.net) |