This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Arguments for Obj-C++ in GCC 4.0

here are some arguments if you need them to get obj-c++ out of the door for 4.0


stefankst (18:38:35 Uhr): does anyone know about the state of objc++?
alexm (18:38:48 Uhr): hi Stiivi
alexm (18:39:02 Uhr): stefankst: "dead"?
grey_gandalf (18:39:36 Uhr): stefankst: I hope dead
alexm (18:39:59 Uhr): there was a mail about it in gcc-patches lately...
stefankst (18:40:04 Uhr): who's dead?
grey_gandalf (18:40:08 Uhr): it seems like gnustep awaits objc++ like some "gift from the gods" and when it comes all problems will be solved :)
alexm (18:40:21 Uhr): stefankst: obj-c++ :)
stefankst (18:40:25 Uhr): no, most problems will appear :-)
stefankst (18:40:37 Uhr): but i still have WebKit in the pipe
alexm (18:40:44 Uhr):
alexm (18:43:13 Uhr): (near the end)
irc (18:46:59 Uhr): n2 has left gnustep
froog_ (18:47:37 Uhr): bummer
irc (18:49:01 Uhr): rhix has left gnustep
tehdely (18:49:15 Uhr): jesus
tehdely (18:49:17 Uhr): still no movement on this
tehdely (18:49:25 Uhr): this is what is wrong with cathedrals, my friend
• (18:49:28 Uhr): tehdely pings ESR
alexm (18:49:40 Uhr): heh
alexm (18:49:53 Uhr): the problem is that nobody's working on it
tehdely (18:50:03 Uhr): that's only half of the problem
alexm (18:50:12 Uhr): no, that's the entire problem
tehdely (18:50:13 Uhr): the other half is that those who were working on it have given up because of all the bureaucratic red tape
alexm (18:50:30 Uhr): not "those"
tehdely (18:50:30 Uhr): remember that one fella from apple who was holding it up for weeks
tehdely (18:50:32 Uhr): and wouldn't explain why?
tehdely (18:50:38 Uhr): even to his own coworkers
tehdely (18:50:39 Uhr): geoff something
alexm (18:50:41 Uhr): Zem is the only person who has ever worked on it
alexm (18:50:47 Uhr): and now he's busy with other stuff
tehdely (18:50:57 Uhr): i will bother him through appledirectory
tehdely (18:50:59 Uhr): ;)
alexm (18:51:04 Uhr): that was sorted out eventually, though
tehdely (18:51:21 Uhr): i am an important apple employee, you see. i have a red phone to steve jobs on my desk.
• (18:51:25 Uhr): tehdely barfs up some feces
alexm (18:52:24 Uhr): (from what I can gather through wild unsubstantiated rumors, the apple compiler guys have personal issues getting along)
tehdely (18:53:10 Uhr): yes, i gathered as much
tehdely (18:53:20 Uhr): why they couldn't just all go to lunch and sort things out, i'm not quite sure
tehdely (18:53:27 Uhr): i imagine they work ont he same floor in the same building
irc (18:53:40 Uhr): _chris has joined gnustep
_chris (18:54:13 Uhr): hello
tehdely (18:54:20 Uhr): howdy
_chris (18:54:31 Uhr): i want writeToFile: to write macos x style property lists
_chris (18:54:37 Uhr): i read that i have to set NSWriteOldStylePropertyLists
_chris (18:54:59 Uhr): but in which domain? defaults write <what to insert?> NSWriteOldStylePropertyLists 0
alexm (18:55:17 Uhr): tehdely: iirc, they did :)
tehdely (18:55:48 Uhr): i hope it was delicious
alexm (18:56:06 Uhr): anyway, Zem's busy, so nobody's working on it, and thus, there'll be no progress for a while
alexm (18:56:23 Uhr): tehdely: I meant they did work on the same floor in the same building
tehdely (18:56:30 Uhr): ah
tehdely (18:56:31 Uhr): hehe
alexm (18:56:44 Uhr): frankly, I think people should just plan as if we won't get obj-c++
tehdely (18:57:16 Uhr): i'm just tired of every possible attempt to get it in gcc dying because of lack of interest and bureaucracy
alexm (18:57:20 Uhr): more productive than waiting for something that may or may not happen, but probably won't happen soon either way
froog_ (18:57:23 Uhr): when is 4.0 scheduled for release? and when can we expect to see 4.1? 6 months time? august?
tehdely (18:57:25 Uhr): hell, what would it cost to fund the development of a proper objc++ patch
tehdely (18:57:32 Uhr): we should raise some money
alexm (18:57:47 Uhr): froog_: "early 2005"
froog_ (18:58:01 Uhr): and 4.1?
stefankst (18:58:08 Uhr): uumpf, no objectivec++
froog_ (18:58:34 Uhr): when is it reasonable to expect a 4.1
stefankst (18:58:35 Uhr): alexm: you're probably right. it's better not to think about objc++ anymore
alexm (18:58:51 Uhr): froog_: not even tentatively scheduled
froog_ (18:59:01 Uhr): ok
alexm (18:59:17 Uhr): froog_: but judging from past releases, I'd guess early-mid 2006
_chris (18:59:31 Uhr): hm, my question remains unanswered... :)
ludovic (18:59:41 Uhr): people should just forget about obj-c++ and get over it.
tehdely (19:00:00 Uhr): obj-c++ is preventing us from using so much FS/OSS code it's not even funny
stefankst (19:00:06 Uhr): this means no usable GNUstep WebKit during the next months
tehdely (19:00:10 Uhr): GNUstep is like a stunted child without it
• (19:00:16 Uhr): tehdely starts ebegging
alexm (19:01:07 Uhr): tehdely: it doesn't prevent anything, it just makes it a bit more complicated
ludovic (19:01:12 Uhr): tehdely: it'll take years before we get it and years before those apps get fully ported. in the meantime, much stuff will happen that might make those apps meaningless
stefankst (19:01:36 Uhr): alexm: it prevents me from doing webkit :-(
irc (19:01:53 Uhr): knebel has joined gnustep
alexm (19:02:52 Uhr): stefankst: how? is there any reason wrappers can't work?
knebel (19:03:17 Uhr): hey, all.. I cannot get the GNU live cd to boot. I guess I burned it incorrectly.
stefankst (19:03:20 Uhr): alexm: the wrappers are too messy. bug fixing and finding is a real nigthmare
alexm (19:03:30 Uhr): (granted, wrappers aren't fun, but it seems like they could handle this)
stefankst (19:03:47 Uhr): alexm: and it's impossible to integrate changes from apple without rewriting large parts
alexm (19:04:16 Uhr): _chris: I think NSPropertyList.h has docs for this
froog_ (19:04:20 Uhr): can SWIG help here?
knebel (19:04:52 Uhr): Anyone know how to make a bootable iso with Roxio?
alexm (19:05:14 Uhr): froog_: possible
stefankst (19:05:17 Uhr): froog_: i already thought about it. and it looks like now i have to think about it again
• (19:05:44 Uhr): ludovic smiles at stefankst
alexm (19:05:44 Uhr): if I really wanted to mix c++ and obj-c, I'd go for some auto-generated wrapper approach
ludovic (19:06:11 Uhr): if I really wanted to mix c++ and obj-c, I'd shoot myself
alexm (19:06:17 Uhr): or, maybe, hack up POC or something that can pre-process obj-c++ to c++
grey_gandalf (19:06:45 Uhr): is there no way of getting another web thing ?
stefankst (19:06:45 Uhr): ludovic: i agree :-)
grey_gandalf (19:07:00 Uhr): what are the hurdles in getting the old owmniweb core? license problems ?
froog_ (19:07:25 Uhr): alexm: Without knowing too much about it, that sound like an interesting idea (POC, I mean)
stefankst (19:07:27 Uhr): grey_gandalf: the good thing is that webkit does not depend on webcore. it would be fairly easy to replace webcore with something else. but with what?
stefankst (19:08:19 Uhr): alexm: POC might be and interesting option
knebel (19:08:47 Uhr): Do I need to emulate HDD or Floppy or None?
knebel (19:09:02 Uhr): When burning an ISO for the Live CD.
_chris (19:09:02 Uhr): i set it NSGlobalDomain, but no effect
_chris (19:09:15 Uhr): knebel: which software do you use to burn it?
_chris (19:09:25 Uhr): knebel: but it should not make any difference
knebel (19:09:42 Uhr): Roxio CD creator _chris. :]
knebel (19:09:59 Uhr): Maybe The download was incomplete.
stefankst (19:10:13 Uhr): grey_gandalf: omnicore has a proprietary license.
knebel (19:10:19 Uhr): I've tried nearly Six times and checked with google.
_chris (19:10:23 Uhr): knebel: did you ever burn other bootable cds, like some bsd- oder linux-image?
alexm (19:10:32 Uhr): has anyone actually tested the obj-c++ support in mainline gcc?
_chris (19:10:44 Uhr): knebel: or try to boot a vmware with that image
knebel (19:10:48 Uhr): Nah, I'm GNU to the whole process.
knebel (19:11:05 Uhr): I'm on an windows distro.
stefankst (19:11:31 Uhr): grey_gandalf: something that says that all modifications of the sources automatically belong to omnigroup
stefankst (19:11:43 Uhr): grey_gandalf: ... if i remember correctly
alexm (19:12:02 Uhr): anyway, I think my take on obj-c++ in gcc is:
alexm (19:12:13 Uhr): a. there's a good chance that it won't be in 4.0
knebel (19:12:15 Uhr): I occasionally get an GRUB error.
alexm (19:12:27 Uhr): b. there's an even better chance that if it is in 4.0, there will be serious bugs with it
knebel (19:12:51 Uhr): Mostly I've gotton an error, "no OS found" and once I recieved a Curser...
irc (19:13:02 Uhr): Stiivi has left gnustep
stefankst (19:13:42 Uhr): alexm: i vote for c. obj-c++ is in 4.0 and works like a charm :-))
alexm (19:13:50 Uhr): thus, I'd estimate 12-18 months until there is a gcc release with working (for our purposes) obj-c++ support
alexm (19:14:20 Uhr): (i.e. in 4.1.x, x>0)
knebel (19:14:21 Uhr): I'll try to download it again. :]
_chris (19:14:37 Uhr): do they have something like md5 hashes?
_chris (19:14:43 Uhr): knebel: you could compare those
alexm (19:14:52 Uhr): and, of course, even longer until it will be widely distributed
knebel (19:15:07 Uhr): I wouldn't know how to.
grey_gandalf (19:15:10 Uhr): stefankst: but omnicore is no longer developed, right ?
knebel (19:15:14 Uhr): It's 92 megs.
grey_gandalf (19:15:15 Uhr): they should open it up:)
stefankst (19:15:36 Uhr): grey_gandalf: yep :-)
stefankst (19:16:08 Uhr): grey_gandalf: but on the other hand, it's old and would require much work to make it supporting todays standards
grey_gandalf (19:16:24 Uhr): stefankst: yes, but it would be a clean core
grey_gandalf (19:16:29 Uhr): no c++ crap
grey_gandalf (19:16:30 Uhr): :)
grey_gandalf (19:16:51 Uhr): and it is proven to work on osx :)
stefankst (19:18:17 Uhr): there must be something else!?
irc (19:18:20 Uhr): rIO has joined gnustep
rIO (19:18:24 Uhr): hello
knebel (19:18:48 Uhr): Is there a cross platform installer?
knebel (19:19:05 Uhr): I guess I could look into it, myself. :|
alexm (19:23:46 Uhr): ah, new gcc 4.0 status report
alexm (19:27:07 Uhr): Mark is now removing "fix by 4.0" marks from all bugs that aren't important c/c++ bugs
alexm (19:27:08 Uhr): sounds like the release is getting close

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]