|Home » Learning Curve » Red Hat Diaries
The Leopard You Don't Yet See
Does the name '\342\220\200\342\220\200\342\220\200\342\220\200HFS+ Private Data' ring a bell?
Apple Maccie fanboy sites are going wild right now. Three or four times a day they're publishing new (perhaps illicit) screenshots of the coming Leopard due out in two months. Perhaps these 'leaks' are part of the typical PR rollout, perhaps not.
What's really fun is to read all the fanboy comments. Things like 'it's really looking better now' or this classic.
I'm really glad they delayed the release so they could get things right.
Clue: fanboys don't know shit. They look at pictures. That's it.
They check the pictures, they drool excitedly over them - and on that basis alone decide whether an OPERATING SYSTEM is worthy of release.
Run for cover.
But there is an operating system - or at least an operating system 'upgrade' - slated for release in October. It's just that the details of what's really going on are cloaked in NDA agreements. So even if outsiders know what's going on they can't say.
But if they could say - and if they had any brains - they wouldn't be talking about unified toolbars and Cover Flow™. They'd be talking about things like this instead.
- Input managers. Are they finally gone? They should be.
- /Library. This is a blithering fanboy mess and needs to be cleaned up. Will it be cleaned up?
- Mail. Is it possible any major app from any other major vendor has been so screwed up for so long? There are myriad bugs that have persisted since April 2005 and never been fixed. To the point one has to wonder if they want to - or can - fix them. To wit: lost folders, clipped text, text and scroll bar remnants, 'text tokens' that don't work right, munged menu bars, message windows appearing behind the document window, plain text inserted as rich text, accessing the wrong pasteboard - the list is ginormous. All these bugs are recognised by Apple but none of them are fixed.
- Safari. The 'world's fastest browser' still suffers from cosmetic issues from Day One. Again one wonders what the F they're doing. Other things as well. To wit: it can't adjust the size of frames, you can't eliminate the Google box, it can't remember what URL it's on, and so forth. All these bugs are recognised by Apple but none of them are fixed.
GDE Copyright (C) Rixstep. All rights reserved.
d_ino: 2 d_reclen: 12 d_type: DT_DIR d_namlen: 1 d_name: .
d_ino: 1 d_reclen: 12 d_type: DT_DIR d_namlen: 2 d_name: ..
d_ino: 712008 d_reclen: 24 d_type: DT_REG d_namlen: 15 d_name: .hotfiles.btree
d_ino: 0 d_reclen: 20 d_type: DT_REG d_namlen: 8 d_name: .journal
d_ino: 0 d_reclen: 28 d_type: DT_REG d_namlen: 19 d_name: .journal_info_block
d_ino: 519572 d_reclen: 24 d_type: DT_DIR d_namlen: 15 d_name: .Spotlight-V100
d_ino: 519579 d_reclen: 20 d_type: DT_DIR d_namlen: 8 d_name: .Trashes
d_ino: 21 d_reclen: 16 d_type: DT_DIR d_namlen: 4 d_name: .vol
d_ino: 22 d_reclen: 24 d_type: DT_DIR d_namlen: 12 d_name: Applications
d_ino: 520774 d_reclen: 20 d_type: DT_DIR d_namlen: 9 d_name: automount
d_ino: 2915 d_reclen: 12 d_type: DT_DIR d_namlen: 3 d_name: bin
d_ino: 30110 d_reclen: 16 d_type: DT_DIR d_namlen: 5 d_name: cores
d_ino: 635841 d_reclen: 16 d_type: DT_REG d_namlen: 4 d_name: css2
d_ino: 2959 d_reclen: 12 d_type: DT_DIR d_namlen: 3 d_name: dev
d_ino: 528777 d_reclen: 20 d_type: DT_DIR d_namlen: 9 d_name: Developer
d_ino: 2960 d_reclen: 12 d_type: DT_LNK d_namlen: 3 d_name: etc
d_ino: 2961 d_reclen: 16 d_type: DT_DIR d_namlen: 7 d_name: Library
d_ino: 4699617 d_reclen: 16 d_type: DT_LNK d_namlen: 4 d_name: mach
d_ino: 4699616 d_reclen: 20 d_type: DT_REG d_namlen: 8 d_name: mach.sym
d_ino: 2993 d_reclen: 20 d_type: DT_REG d_namlen: 11 d_name: mach_kernel
d_ino: 30112 d_reclen: 16 d_type: DT_DIR d_namlen: 7 d_name: Network
d_ino: 2994 d_reclen: 16 d_type: DT_DIR d_namlen: 7 d_name: private
d_ino: 3109 d_reclen: 16 d_type: DT_DIR d_namlen: 4 d_name: sbin
d_ino: 3173 d_reclen: 16 d_type: DT_DIR d_namlen: 6 d_name: System
d_ino: 11276 d_reclen: 12 d_type: DT_LNK d_namlen: 3 d_name: tmp
d_ino: 520379 d_reclen: 36 d_type: DT_LNK d_namlen: 27 d_name: User Guides And Information
d_ino: 28448 d_reclen: 16 d_type: DT_DIR d_namlen: 5 d_name: Users
d_ino: 11277 d_reclen: 12 d_type: DT_DIR d_namlen: 3 d_name: usr
d_ino: 11561 d_reclen: 12 d_type: DT_LNK d_namlen: 3 d_name: var
d_ino: 11562 d_reclen: 16 d_type: DT_DIR d_namlen: 7 d_name: Volumes
d_ino: 0 d_reclen: 40 d_type: DT_DIR d_namlen: 29 d_name: ␀␀␀␀HFS+ Private Data
3 cloaked items.
- The protocol hole. What a wonderful way to sneak malware onto a system. The solution worked but it really sucked too as it didn't address the underlying design issue.
- The Safari hole. The media loved this one: users were getting clobbered because Beige Box Technology(tm) was fooling the system on the nature of downloads. What did they do? Fixed their own web apps of course! Everyone else's - Seamonkey, Thunderbird, Firefox, Camino, Eudora, et al, et al - are still wide open. Will they actually get off their fanboy behinds and fix this design issue?
- The MOAB 15 hole. It's been wide open for over eight months now. No fixes. No questions asked and any rogue process can COMPLETELY take over a machine. Yes as root - and YES just like that. Is this going to be fixed before the next millennium?
- The Opener hole. Sorts under /Library above but should be mentioned on its own for the 'solution' they finally arrived at (after several clumsy false starts) is downright ridiculous. Will they finally move all such functionality back into /System/Library where it belongs?
- $PATH hijacking. Both through bash and through legacy NS stuff. Will this be eliminated? Yes if you read Rixstep and use the solutions here you are protected but otherwise you are wide open. The number of attack vectors into this 'OS' is getting ridiculously big.
- launchd. This nonsense has to be stopped before it gets any worse. But the fear is it won't be stopped. And in such case one can guarantee a year or two from now 'OS X' will be a mess, corrupted and humiliated by flagrant attacks from everywhere. One doesn't need to be a betting man to be sure to clean up big on such a wager.
- HFS horseshit. Name it, take your pick, take all. To wit: resource forks, unsupervised data streams, dangerously unreliable file associations, Finder info in volume control blocks, file management logic removed from the frameworks, COLOURS IN A UNIX SYSTEM, kIsInvisible, .hidden, inability of the system shell to cope with hard links (user files are lost all the time because of this nonsense) and so on.
Add to that a dangerously fragile file system, a hopelessly paraplegic 'trash' facility, the overwriting of entire directory trees with single document files, AppleDouble, Windows friendly attachments, creator codes, file types, 'Carbon', 'Classic', 'PEF' executables, AppleBloodyScript (check how they take straightforward open source projects and wrap them in AppleScript so the retards on the platform can better cope) and so forth.
PkgInfo, keyedobjects, ruinous NIBs and chaotic Interface Builder, frazzled Quartz drivers - and of course the incomparable '\342\220\200\342\220\200\342\220\200\342\220\200HFS+ Private Data'.
This rabbit hole goes through to the mountains of Tibet.
- Keeping FOSS up to date. It's too late to hope they're going to play by the rules - their ridiculous 'macosforge' site is an insult - but at least they can keep other people's better code up to date? The module they fixed to guard themselves against Charlie Miller's iPhone exploit is again FOUR VERSIONS OUT OF DATE. Will this next 'OS' ship with up to date FOSS components across the board?
The Leopard you don't yet see might turn out to be the Leopard you don't want to see.