Bumblebee



  • Why would I expect someone who has worked in Linux to know that the /usr folder is important?  Because... because there's only a few folders that you ever use, and that's one of them. 



  •  Note to self: Do not share root password with Blakey.



  • @TheChewanater said:

    I like this better:

    DIRECTORY="/usr/share/whatever"

    rm -rf $DIRECTRY/*

     

    One time when I was about to install a new operating system, for fun I tried the command

    rm -rf /*

    just to see how far it would get. It leaves a surprising lot of stuff on the hard disk because the 'rm' command dies when it deletes things that are critical for it's operation. Not enough stuff left to reboot, but still running. Back then it couldn't finish the job by itself. Now with disk cacheing you might be able to,

     


  • Garbage Person

    @AndyCanfield said:

    One time when I was about to install a new operating system, for fun I tried the command

    rm -rf /*

    just to see how far it would get. It leaves a surprising lot of stuff on the hard disk because the 'rm' command dies when it deletes things that are critical for it's operation. Not enough stuff left to reboot, but still running. Back then it couldn't finish the job by itself. Now with disk cacheing you might be able to,

    rm can and should be loaded entirely into memory while it's executing, as should the kernel and shell. The system will totally keep running, and if you could somehow run an ls, you'd find a totally blank filesystem. More likely the stuff that was leftover was locked, or some shit. Assuming Linux even supports file locking.


  • For the lulz, I just typed sudo rm -rf /* in an xterm.¹ Gnome looks a bit funny, now that all fonts are gone. Only /proc, /sys and a few other directories in /var and /dev have remained. One thing surprised me, though: The desktop background did not simply disappear. Instead, it faded to black. Reality slowly approaches Hollywood.

    ¹ In a throwaway VM. What did you think?



  • @blakeyrat said:

    The worst part now is that nobody in that long-ass thread is saying, "ok, how do we prevent mistakes like this from happening in the future?"

    Incorrect. Some of the posters DID say "this is why you should use QUOTES" around what you or I would call the directory name.

    And stop whining about the somewhat mild abuse you're getting for FLAUNTING your ignorance, and for doing so in your usual abrasive way. After all, that's exactly what you do in other threads here; except that the abuse you dish out to others seems only to have a 'full on' and 'full off' setting, rather like a flamethrower. (Pun intended)


  • Discourse touched me in a no-no place

    @AndyCanfield said:

    rm -rf /*
    http://www.feep.net/~roth/geek-humor/unix/unix-crash-recovery-story for those who haven't seen it. (Well it is from 1986)



  • @Cad Delworth said:

    @blakeyrat said:

    The worst part now is that nobody in that long-ass thread is saying, "ok, how do we prevent mistakes like this from happening in the future?"

    Incorrect. Some of the posters DID say "this is why you should use QUOTES" around what you or I would call the directory name.

    Thats not a solution. It's slightly less prone to human error, I suppose, but its certainly not anything that can be considered a "solution". One of the posters (probably jokingly) said "that's why Linux needs a recycle bin", which actually is a solution, but I'm sure Linux users are too pig-headed to adopt it. Another solution might be using an install/uninstall system that allows automated roll-backs, you know, like most other OSes already have.

    If Linux developers honestly think "just put quotes around it" is a permanent solution to this problem, it says a lot about the quality of Linux software. If Linux developers have gone this long without actually solving this rather simple problem, it says more...

    @Cad Delworth said:

    And stop whining about the somewhat mild abuse you're getting for FLAUNTING your ignorance,

    What ignorance am I FLAUNTING? Yes, I didn't know what goes in /usr. Big fucking woop. It's the poster of the WTF's responsibility to actually communicate to the rest of us what the WTF is.

    My biggest crime is assuming this WTF is worse than it actually was... based on the forum responses, I assumed it was wiping out user data, not just re-installable system stuff. Of course, none of this would have happened if someone had explained the WTF in the first place.

    If you're saying I'm FLAUNTING my ignorance by having used Linux for almost a decade and still not knowing what /usr is... well, I still don't hear any compelling reasons I should have known what it is. Computers are tools. I use the tool to do work. If I have to know implementation details like what specific files go into some specific folder, then the OS has failed. Ironically, in this case, the OS has not failed me (well, it did in hundreds of ways, but that's another post), and yet people are so used to it failing that they just assuming that sooner or later I'd have come across /usr. (Similarly, I don't know what .exe powers IIS because it's never come up.)

    If you're saying I'm FLAUNTING my ignorance because I don't spend my free time researching what three letter folders do in Linux, then a) you need to learn that some people are not like you, and in fact the IT monoculture is extremely harmful, and it's a good thing that there are people in IT who do not think like you, and b) you need to get a life, seriously.

    @Cad Delworth said:

    rather like a flamethrower. (Pun intended)

    Puns are never funny. But explaining the pun makes it somehow even less funny, driving your humor quotient well into the negative values.



  • re: FLAUNTING your ignorance.

    1) You readily admit you do not know...Ignorance is established

    2) You are conspicuously, defiantly, and/or boldly presenting this .....Flaunting is established.

    So you most certainly are flaunting your ignorance...not a "value judgment", just a valid semantic way of referring to your post.



  • @blakeyrat said:

    If Linux developers honestly think "just put quotes around it" is a permanent solution to this problem, it says a lot about the quality of Linux software.

    So, are you flaunting your ignorance, or do you genuinely not believe that the exact same problem affects windows in exactly the same way?




  • @DaveK said:

    @blakeyrat said:

    If Linux developers honestly think "just put quotes around it" is a permanent solution to this problem, it says a lot about the quality of Linux software.

    So, are you flaunting your ignorance, or do you genuinely not believe that the exact same problem affects windows in exactly the same way?


    I don't believe I mentioned Windows at all. Also the article you refer to isn't the same problem.

    But since you brought it up, software written for Windows makes use of installers that automatically track changes made to the system and can undo them. Additionally, Microsoft has added Windows File Protection to critical system files, helping the OS to recover no matter what a particular piece of malware does to the system.

    Is it perfect? Probably not. But it's a fuckload better than "hey just put quotes around it oh and BTW we're not giving you this advice until after you've already fucked it up once."

    And hell, for all I know, those technologies already exist for Linux. Maybe even better ones. Which would make it even more indefensible that the posters in that thread aren't suggesting that this developer actually use them.

    Relying on a human being to not make errors is not in any way a solution. If a destructive action is prone to human error, the correct solution is to get the human out of the picture and automate it.



  •  @blakeyrat said:

    And hell, for all I know, those technologies already exist for Linux. Maybe even better ones. Which would make it even more indefensible that the posters in that thread aren't suggesting that this developer actually use them.

    Relying on a human being to not make errors is not in any way a solution. If a destructive action is prone to human error, the correct solution is to get the human out of the picture and automate it.

    Most projects use GNU Autotools, which the bumblebee developers really should have used.  Yes, it's still possible to roll your own uninstall script and delete an important directory, but the same goes for any OS. 

    However, on Unix-like systems, it would need explicit permission to modify anything in /usr, which people really shouldn't give to home-grown scripts they found online.

     



  • @blakeyrat said:

    Computers are tools. I use the tool to do work.
     

    Linux computers are not tools; they are toys.

    Nobody owns a hot rod just to commute. I run Linux because I love the power it gives me, and the joy of controlling everything. That power comes from "sudo bash". In Linux, when you are root, you can do EVERYTHING. I like that. This computer belongs to ***ME***, not BIll Gates, not Linus Torvalds, not Steve Jobs, ***ME***. 

    Linux does not defend the computer against the "dumb user". That's part of the design philosophy.

    I have fixed bugs in the Linux operating system. The fact that the bug was there is a negative, but far outweighed by the fact that I had the power to fix the bug.

    • OS X is like Switzerland; high security but worth it.
    • Windows is like Homeland Security; hots of hassel and not worth it.
    • Linux is like Thailand; you can do damned near anything here (except insult the King).

    Yeah, it might be nice if "sudo rm -rf /*" asked you "Are you sure?". But only if it's typed from a terminal. And that will still not defend you against "sudo rm -rf ~/*".

    By the way, I have three partitions - "/", swap, and "/data". All user data goes into /data. So I can reinstall the operating system in "/" without losing user data. May take days to get the new system configured properly, wouldn't protect against "sudo rm -rf /*", but if the OS crashes I can reinstall safely. 

     



  • @AndyCanfield said:

    @blakeyrat said:

    Computers are tools. I use the tool to do work.
     

    Linux computers are not tools; they are toys.

    Well, opinions vary. YOU can treat it as a "toy" if YOU want. For ME it was a tool, and a cut-rate one at that. The only reason the MUD wasn't hosted on Windows is because renting dedicated Windows servers (especially back in 2002-ish) is more expensive.

    @AndyCanfield said:

    Linux does not defend the computer against the "dumb user". That's part of the design philosophy.

    Imagine if Boeing decided they weren't going to design their planes to protect against pilot error!

    Now ***I*** have three bits of Linux design philosophy ***I*** think are completely wrong:
    1) Release early and often
    2) Programs should be small and do only one task
    3) The computer shouldn't be protected against typos

    @AndyCanfield said:

    By the way, I have three partitions - "/", swap, and "/data". All user data goes into /data. So I can reinstall the operating system in "/" without losing user data. May take days to get the new system configured properly, wouldn't protect against "sudo rm -rf /*", but if the OS crashes I can reinstall safely.

    Yeah; that's why this bug isn't as severe as I first thought. The OS stuff is easily re-installable, the only truly valuable data on a system is the user's data.

    (Stupidly, most computer security designs still get this one exactly wrong. Both Windows and Linux have permissions set up so that a malware script can easily delete all of the user's data, but can't touch the OS itself... well, good for the OS ***I*** guess, but ***MY*** data was the valuable part! Thankfully, Microsoft has Shadowcopy and a good Backup tool, and OS X has Time Machine, and Mozy and DropBox exist, so it's not as big a problem now as it was a decade ago.)



  • @AndyCanfield said:

    That power comes from "sudo bash". In Linux, when you are root, you can do EVERYTHING. I like that. This computer belongs to ***ME***, not BIll Gates, not Linus Torvalds, not Steve Jobs, ***ME***.

    If you take the time to learn what you are doing, you can fuck up do anything you want in Windows just like you can in Linux.  For the last 20+ years I have done whatever I want on my computer and Bill Gates has never once tried to stop me.

    @AndyCanfield said:

    Linux does not defend the computer against the "dumb user". That's part of the design philosophy.
    That's perfectly fine and wonderful  if ****YOU**** are the only person using the computer.  However, putting an OS that "does not defend the computer against the dumb user" on a computer that is used by more than one person is just plain stupid.  As soon as ****SOMEONE ELSE**** touches the computer, all ****YOUR**** knowledge and/or careful mucking about goes out the window.

     


     

     

    Asterisks Asterisks Asterisks Asterisks
    \



  • @El_Heffe said:

    For the last 20+ years I have done whatever I want on my computer and Bill Gates has never once tried to stop me.
    Oh, maybe it's just me then. A couple of years back I tried to delete a file from \system32\ and Bill landed his helicopter on the roof, abseiled down the outside of my building, smashed in through the window, and grabbed the mouse from my hand.



  • @blakeyrat said:

    most computer security designs still get this one exactly wrong. Both Windows and Linux have permissions set up so that a malware script can easily delete all of the user's data, but can't touch the OS itself... well, good for the OS ***I*** guess, but ***MY*** data was the valuable part!
    The problem is that most (all?) operating systems are desgined around the idea of a multi-user server.  If you have lots of people all connected to the same computer then you want to make sure that the OS is protected so that the computer is always acessable, and if one person's data gets deleted ..... well .... that's what backups are for.   Unfortunately, for a computer with only a single user, that design makes no sense.  The OS and any applications I use can easily be re-installed but those pictures of my neighbor's wife pets are not easily replaced.



  • @El_Heffe said:

    The OS and any applications I use can easily be re-installed

    I guess it depends on your definition of "easily". Two weeks ago I had a machine drop dead. The good news is that all of the "user data" was backed up. A new machine was ordered.

     From the time that the machine was received (next day), until the OS and applications were installed, there was over 30 hours of labor (most of it waiting - while billable at over $100 per hour) along with the need to (re-)purchase upwards of $2K work of software [lergely because it was not possible to track back which specific licenses were on that machine, or the original media had become unavailable.

     Now I grant there are a number of WTF's in here, but the concept that $5K of expense simply to reinstall an OS and application is "easy" is radically absurd.

     Of course, there is an "Easy" way to prove me wrong. If each perons who replies to the remainder of this thread sends me $5K (hey , thats easy), then I will retract my statement.

     Even in a single user environment, it is often still more important (at the machine level) to protect the OS than the user data.



  • @blakeyrat said:

    Now ***I*** have three bits of Linux design philosophy ***I*** think are completely wrong:

    1) Release early and often

    http://en.wikipedia.org/wiki/Windows_Vista


    @blakeyrat said:

    2) Programs should be small and do only one task

    Yep, you're right.  Programs should be huge and do many unrelated tasks.

    @blakeyrat said:

    3) The computer shouldn't be protected against typos

    I'm glad Windows protects against typos, unlike Linux.

     

    C:\> del Foo \Bar

    Are you sure (Y/N)? Y

    C:\> 

     

    Is so much better than...

     

    ~$ rm -r Foo/ Bar 

    rm: remove write-protected directory `Foo'? yes

    ~$


     

     



  • @El_Heffe said:

    @AndyCanfield said:
    Linux does not defend the computer against the "dumb user". That's part of the design philosophy.
    That's perfectly fine and wonderful  if ****YOU**** are the only person using the computer. However, putting an OS that "does not defend the computer against the dumb user" on a computer that is used by more than one person is just plain stupid.

    My Linux notebook is used by only one person - me. Our Linux servers are not used by anyone except "adminu" for maintenance and updates. And if you're the sysadmin you NEVER give anyone the root password. We have people that have used Linux on a notebook for years and never have been given the root password.

    My correction to me: the power comes in knowing the root password. Never tell anyone.
    IMHO same as in Windows - we have people using Windows workstations, and we don't tell them the Administrator password either.

     



  • @blakeyrat said:

    (Linux)Release early and often

    We run Ubuntu Linux. There are two releases every year, in April and October. That doesn't seem too often to me. And the release is optional; you don't have to install the new release if you don't want to. Our servers are STILL running Ubuntu Linux 8.04 LTS from 2008/04.

    Oh yes - upgrades are free.

    @blakeyrat said:

    (Linux)Programs should be small and do only one task

    Command-line programs are small doing only one task so that the user can string them together to do something more complex. If you want a program to do the whole job itself stay with the GUI. I find that a toolbox is more powerful and useful than a one-size-fits-all Swiss army knife.

    @blakeyrat said:

    (Linux)The computer shouldn't be protected against typos

    Actually Linux has protected against dumb users since it was first written, just like Windows NT. You don't give out the root password to dumb users. 

    Yes, a design philosophy of Unix, all the way back to the 1970's, is "you asked for it, you got it". You have the freedom and power to hurt yourself. You want to be protected, never open a terminal window.

     



  • @TheCPUWizard said:

    the concept that $5K of expense simply to reinstall an OS and application is "easy" is radically absurd.
     

    Disk images.  Say it with me.  Disk images.  And a hard drve mounted in a removable caddy.  If you are using a computer for anything remotely important, this is the only sane non-WTF way to do it.

    It just so happens that today I replaced the hard drive in one of my computers.  One minute to swap out the hard drive and less than 15 minutes to load the disk image.  Manually copy over a few misc files and I'm up and running and fully functional in less than 20 minutes from start to finish.  I call that easy. 




  • @El_Heffe said:

    Disk images.  Say it with me.  Disk images.  And a hard drve mounted in a removable caddy. 

     That is generally good advice. Unfortunately in this instance it would not have helped as the replacement system was not hardware identical to the original system. I did attempt a restore of a 2 day old image [they are taken each Sunday during a scheduled "Freeze" period] but much of the software simply would not run properly without being re-installed.

    For scenarios, that do NOT require hardware dependency (I do alot of industrial systems, audio/video processing - all which have dedicated and specialized hardware/software - this was pretty much a worst case scenario) I usualy go one step further an run everyting inside a virtualized environment - but that was not possible here.



  • @TheCPUWizard said:

    . Two weeks ago I had a machine drop dead.
    What killed it? Just wondering if there's a lesson in that.



  • @intertravel said:

    @TheCPUWizard said:
    . Two weeks ago I had a machine drop dead.
    What killed it? Just wondering if there's a lesson in that.

     The lesson is "be very careful when you override safety interlocks!". The root cause was human error by one of the engineers; the final COD was an (estimated) 1.5KV spike back-feeding into the computer proper.

     Just lucky it didn't flame (there were scortch marks everywhere, and many of the chips had "blew their lids"). You cant believe what happens if the fire suppression system goes off....



  • @TheCPUWizard said:

     The lesson is "be very careful when you override safety interlocks!".
    Hey, that was the lesson from Chernobyl, and has anyone learnt it?

    The root cause was human error by one of the engineers; the final COD was an (estimated) 1.5KV spike back-feeding into the computer proper.
    Oh, right. I suspect that's not one I'll run across day-to-day :)


  •  The bug was reported on May 23-24 and fixed on May 24.  Apparently, it didn't go viral until about 2 days ago.



  • @blakeyrat said:

    @DaveK said:

    @blakeyrat said:

    If Linux developers honestly think "just put quotes around it" is a permanent solution to this problem, it says a lot about the quality of Linux software.

    So, are you flaunting your ignorance, or do you genuinely not believe that the exact same problem affects windows in exactly the same way?

    I don't believe I mentioned Windows at all.

    Maybe not, but you sure as hell weren't talking about TOPS-20 or VMS! 

    @blakeyrat said:

    Also the article you refer to isn't the same problem.

    They both looked like the same problem to me: the shell-uses-space-as-a-word-separator-but-it-can-be-a-legitimate-file-or-path-character problem.  It's an old and well-known problem (see Unix Hater's Guide, for example), and it exists on all of *nix and NT/9x architectures.

    @blakeyrat said:

    But since you brought it up, software written for Windows makes use of installers that automatically track changes made to the system and can undo them.

    Yes, and a) the same exists on Linux systems, and b) there are still badly-written windows programs that come with crappy .bat file installer scripts, so what's your point? 

    @blakeyrat said:

    Additionally, Microsoft has added Windows File Protection to critical system files, helping the OS to recover no matter what a particular piece of malware does to the system.

    Is it perfect? Probably not. But it's a fuckload better than "hey just put quotes around it oh and BTW we're not giving you this advice until after you've already fucked it up once."

    From the MSDN docs for CreateProcess:

    If you do pass NULL for lpApplicationName, use quotation marks around the executable path in lpCommandLine, as shown in the example below.

    Same problem.  Same proposed solution.  Different standard in operation when you criticise one and not the other.

    @blakeyrat said:

    And hell, for all I know, those technologies already exist for Linux. Maybe even better ones. Which would make it even more indefensible that the posters in that thread aren't suggesting that this developer actually use them.

    The posters in that thread are a random bunch of internet trolls, not some kind of representative delegation from the open source movement as you seem to assume.  Note also that that is the comment thread on a webcvs-style view of a commit entry, and not any kind of response to the original bug reporters.

    @blakeyrat said:

    Relying on a human being to not make errors is not in any way a solution. If a destructive action is prone to human error, the correct solution is to get the human out of the picture and automate it.

    Well duh!  The problem in this case was that there was a bug in the script written to automate it.



  • @TheCPUWizard said:

    Even in a single user environment, it is often still more important (at the machine level) to protect the OS than the user data.

    "Often" in this case meaning, "if you don't keep track of any of your software licenses." And even then we still can't judge because you didn't actually quote the value of the data stored on the computer-- for all we know it cost you $5k and the data was worth $25k.

    And why are you billing for "waiting"? Waiting on-hold, or waiting for disks to arrive? Because if you're billing for the latter, you're a jerk.



  • @TheChewanater said:

    @blakeyrat said:

    Now ***I*** have three bits of Linux design philosophy ***I*** think are completely wrong:

    1) Release early and often

    http://en.wikipedia.org/wiki/Windows_Vista

    Oh snap!

     



  • @TheChewanater said:

    @blakeyrat said:

    Now I have three bits of Linux design philosophy I think are completely wrong:

    1) Release early and often

    http://en.wikipedia.org/wiki/Windows_Vista

    Did you have a point to make? Or are you just sharing random URLs you've come across?

    @TheChewanater said:

    @blakeyrat said:

    2) Programs should be small and do only one task

    Yep, you're right.  Programs should be huge and do many unrelated tasks.

    Photoshop and Microsoft Office are more popular than any open source product has ever been.

    @TheChewanater said:

    @blakeyrat said:

    3) The computer shouldn't be protected against typos

    I'm glad Windows protects against typos, unlike Linux.

     

    C:> del Foo \Bar

    Are you sure (Y/N)? Y

    C:> 

     

    Is so much better than...

     

    ~$ rm -r Foo/ Bar 

    rm: remove write-protected directory `Foo'? yes

    ~$

    Hey, I don't fucking use Linux. Your compatriot in this very thread told me that Linux doesn't protect against "user stupidity." Again: I AM NOT TELEPATHIC. If someone on this board tells me it doesn't protect against typos, and they don't have any reason to lie to me, I assume that Linux doesn't protect against typos.

    Fucking shit. Does anybody on this thread even know what Linux is? Does it protect against typos or not? Does /usr stand for user or U-something system resources? What the holy shit people! Tell you what, why don't you huddle-up and figure it fucking out before replying to my points next time, 'kay?

    @AndyCanfield said:

    (everything this man says)

    Someone left Slashdot's gate open, again. They're escaping.



  • @blakeyrat said:

    @TheChewanater said:

    @blakeyrat said:

    Now ***I*** have three bits of Linux design philosophy ***I*** think are completely wrong:

    1) Release early and often

    http://en.wikipedia.org/wiki/Windows_Vista

    Did you have a point to make? Or are you just sharing random URLs you've come across?

    Pretty sure I can answer that one.  The point is that Vista was actually a half-finished version of 7 that was thrown out the door early and often solely in order that the department in MS could start to book some incoming revenue against a project that was embarrassingly late and over budget; they sold a half-finished buggy version of the product in order to fund the completion of the actual development work that they should have finished before they started flogging it.




  • @DaveK said:

    @blakeyrat said:

    @DaveK said:

    @blakeyrat said:

    If Linux developers honestly think "just put quotes around it" is a permanent solution to this problem, it says a lot about the quality of Linux software.

    So, are you flaunting your ignorance, or do you genuinely not believe that the exact same problem affects windows in exactly the same way?

    I don't believe I mentioned Windows at all.

    Maybe not, but you sure as hell weren't talking about TOPS-20 or VMS! 

    I wasn't talking about any OS specifically. What is with putting words in my mouth? Please, read my posts, and respond ONLY to what is typed in them. I can't possibly enter a debate against the magical fairies in your brain, or wherever the shit you got that from. Or just don't respond at all. But don't fucking put words into my mouth.

    @DaveK said:

    They both looked like the same problem to me: the shell-uses-space-as-a-word-separator-but-it-can-be-a-legitimate-file-or-path-character problem.  It's an old and well-known problem (see Unix Hater's Guide, for example), and it exists on all of *nix and NT/9x architectures.

    We're not talking about "a typo in a command line command", we're talking about "the process of uninstalling a program." Which, to my understanding, is what this thread is all about.

    Don't confuse the implementation of a task with the task itself.

    @DaveK said:

    Yes, and a) the same exists on Linux systems, and b) there are still badly-written windows programs that come with crappy .bat file installer scripts, so what's your point?

    I would hope that a Windows developer who did something like this would be advised to use an actual installer/uninstaller system and not just ad-hoc their own from batch files.

    @DaveK said:

    From the MSDN docs for CreateProcess:

    If you do pass NULL for lpApplicationName, use quotation marks around the executable path in lpCommandLine, as shown in the example below.

    Same problem.  Same proposed solution.  Different standard in operation when you criticise one and not the other.

    What the fuck? We're talking about uninstalling a program! We're NOT talking about "typing a path into the command line without quotes." What is wrong with you people? How did you so seamlessly change the subject so that we're talking about some stupid little implementation detail of how the uninstaller works, instead of talking about the uninstaller itself?

    @DaveK said:

    Well duh!  The problem in this case was that there was a bug in the script written to automate it.

    Yes, but my point is... writing a script like this is the best way to accomplish the task of uninstalling the program? Seriously? Because it sucks in many fundamental ways. (As should be obvious about now...)

    And if there's a better way, then why the hell aren't they using it?

    That is a valid criticism.



  • @DaveK said:

    Pretty sure I can answer that one.  The point is that Vista was actually a half-finished version of 7 that was thrown out the door early and often solely in order that the department in MS could start to book some incoming revenue against a project that was embarrassingly late and over budget; they sold a half-finished buggy version of the product in order to fund the completion of the actual development work that they should have finished before they started flogging it.

    If that's the argument he intended to make, it's a pretty shitty argument, since anybody who's actually used Vista knows it was a perfectly fine, and finished, product. Yes, there are valid complaints about Vista-- the "bonuses" for the Ultimate Edition were crap, for example, but saying it was unfinished? Ridiculous.

    Or maybe he was just pasting in random URLs he came across. Maybe he's relying his instructions from Planet Weeboo, which broadcasts directly into his brain from a secret UFO factory owned by the Illuminati in Area 51. Who fucking knows.



  • @blakeyrat said:

    And why are you billing for "waiting"? Waiting on-hold, or waiting for disks to arrive? Because if you're billing for the latter, you're a jerk

    And I suppose you or your company will keep a senior person on-site for the better part of 4 days (including paying his salary, travel, hotel, car, etc) for FREE?

    If so, please contact me directly, and I can sent you to DesMoines Iowa for a week, and you won't charge me a cent....

     (ps: it was neither of those, the machine arrived the next day, and the necessary purchases were made on-line - it was the time *waiting* for the system to become operational again.



  • @TheCPUWizard said:

    @blakeyrat said:

    And why are you billing for "waiting"? Waiting on-hold, or waiting for disks to arrive? Because if you're billing for the latter, you're a jerk

    And I suppose you or your company will keep a senior person on-site for the better part of 4 days (including paying his salary, travel, hotel, car, etc) for FREE?

    Yeah, you didn't mention that. Remember, and this is becoming a theme in this thread: I AM NOT TELEPATHIC! If you don't SAY something, I don't know it.



  • @blakeyrat said:

    Photoshop and Microsoft Office are more popular than any open source product has ever been. 

    Wait, the software you use is more popular than the software I use?  Sorry, I guess that means everything I've said is wrong.   Also, MediaWiki is pretty popular.  It only gets, like, 400,000,000 unique visits per day, just counting the official Wikimedia projects

    Back to my original point, are you saying that Microsoft Office does several unrelated tasks?  As I recall, it's made up of many programs, each of which performs one (perhaps complex) task. Not that that has anything to do with the topic, but you're probably going to claim it's not true anyways.

    @blakeyrat said:

    And if there's a better way, then why the hell aren't they using it?

    That is a valid criticism.

     

    Yes, it is a valid criticism.  They should have used Autotools or something similar, we get that.  I don't understand why you need to start complaining about Linux, though. 

     



  • @TheChewanater said:

    Wait, the software you use is more popular than the software I use?  Sorry, I guess that means everything I've said is wrong.

    Well, think about it. If Word was written the Linux Way, the spell-checker would be a program. The word count would be a different program. The outliner would be a different program. Etc. Obviously this is a shitty and useless idea because, guess what, not even open source Linux-based word processors are built that way. Because at some level, Linux developers know it's a dumb idea.

    I won't comment on MediaWiki because I can't for the life of me figure out how it's relevant.

    @TheChewanater said:

    Back to my original point, are you saying that Microsoft Office does several unrelated tasks?  As I recall, it's made up of many programs, each of which performs one (perhaps complex) task. Not that that has anything to do with the topic, but you're probably going to claim it's not true anyways.

    Oh I see, so all you do is redefine the word "task" until it fits.

    @TheChewanater said:

    I don't understand why you need to start complaining about Linux, though.

    It's more the unprofessionalism in open source software development, not Linux specifically. Open source programmers are always just playing around, they don't give a crap about the quality of their work, or whether they've (potentially) hurt the people, or whether they're using the right tool for the job. A bug like this? It's not something to chide someone over, to say, "hey take some pride in your work!" No, instead you point and laugh and nobody's learned anything at all. And, like that article I linked earlier about the CADT development model, you end up in a situation where version 0.8 is always followed by version 0.8 which is always followed by version 0.8.

    If you're going to do it, if you're going to write software you want other people to use, do it well. That's all.



  • @blakeyrat said:

    Yeah, you didn't mention that. Remember, and this is becoming a theme in this thread: I AM NOT TELEPATHIC! If you don't SAY something, I don't know it.

     Please go back and re-read:

     @blakeyrat said:

    until the OS and applications were installed, there was over 30 hours of labor (most of it waiting ..)

    Just in case that is not clear enough....

    1) Put OS Disk in Drive (Wait...)
    2) Answer Initial Questions (Wait...)
    3) Boot into OS (Wait...)
    4) Install 5 Custom Drivers (Wait, reboot, Wait, revoot, Wait, reboot, Wait, reboot, Wait, reboot)
    5) Install SP1 of Os (Wait)
    6) Install PAtches to 3 of the Drivers (Wait,Wait,Wait)
    7) Install Application #1 (Wait)
    8) Apply Patches and Service Packs to Application #1 (Wait)
    9) Install Application #2 (Wait)
    10) Apply Patches and Service Packs to Application #2 (Wait)
    11)Install Application #3 (Wait)
    12) Apply Patches and Service Packs to Application #3 (Wait)

    [Repeat for about 25 distinct applications]

    14) Apply Final Service Packs that are dependent on Multipl Application
    15) Install Custom Software (Wait)
    16) Run Calibration & Verification Tests [these alone tak a full day]

    System dies on Monday, Everything necessary was in-house by Tuesday Afternoon, when this was completed is was Late Friday evening (and my guy, had to delay his flight until Saturday morning....only to have to fly back out the following Monday)

     Gee... that was "EASY"



  • @blakeyrat said:

    @TheChewanater said:
    Wait, the software you use is more popular than the software I use?  Sorry, I guess that means everything I've said is wrong.

    Well, think about it. If Word was written the Linux Way, the spell-checker would be a program. The word count would be a different program. The outliner would be a different program. Etc. Obviously this is a shitty and useless idea because, guess what, not even open source Linux-based word processors are built that way. Because at some level, Linux developers know it's a dumb idea.

    That's not even the "Linux" way - that's the Unix  philosophy, and most software targetting Linux is not developed that way.

    Of course, I can understand why you don't know that.  After all, you're not telepathic, and it's not like this information is well-known, or even availible online so anyone who wants to complain about it can know what they're talking about first.

    @blakeyrat said:

    I won't comment on MediaWiki because I can't for the life of me figure out how it's relevant.

    You said that no open source products were as popular as Photoshop.  Not that I can figure out how that was relevant either.

     



  • @blakeyrat said:

    @DaveK said:

    @blakeyrat said:

    @DaveK said:

    @blakeyrat said:

    If Linux developers honestly think "just put quotes around it" is a permanent solution to this problem, it says a lot about the quality of Linux software.

    So, are you flaunting your ignorance, or do you genuinely not believe that the exact same problem affects windows in exactly the same way?

    I don't believe I mentioned Windows at all.

    Maybe not, but you sure as hell weren't talking about TOPS-20 or VMS! 

    I wasn't talking about any OS specifically. What is with putting words in my mouth? Please, read my posts, and respond ONLY to what is typed in them. I can't possibly enter a debate against the magical fairies in your brain, or wherever the shit you got that from. Or just don't respond at all. But don't fucking put words into my mouth.

    Yes, you were talking about one OS specifically:

    @blakeyrat said:

    If Linux developers honestly think "just put quotes around it" is a permanent solution to this problem, it says a lot about the quality of Linux software.

    It is therefore valid to point out that the situation is the same in the other mainstream OS and so your criticism is biased.

    @blakeyrat said:

    @DaveK said:
    They both looked like the same problem to me: the shell-uses-space-as-a-word-separator-but-it-can-be-a-legitimate-file-or-path-character problem.  It's an old and well-known problem (see Unix Hater's Guide, for example), and it exists on all of *nix and NT/9x architectures.

    We're not talking about "a typo in a command line command", we're talking about "the process of uninstalling a program." Which, to my understanding, is what this thread is all about.

    You need to check your understanding there.  The thread is about a software author making a bug in their uninstall script; the bug that they made in their script was a typo in a command-line command; the same bug could equally have been made by anyone writing an uninstall script on any other platform, was my point. 

    @blakeyrat said:

    Don't confuse the implementation of a task with the task itself.

    Blakey, this is TDWTF, we're all programmers, the entire site is about implementations, and the many and varied ways in which they go wrong.

    @blakeyrat said:

    @DaveK said:
    Yes, and a) the same exists on Linux systems, and b) there are still badly-written windows programs that come with crappy .bat file installer scripts, so what's your point?

    I would hope that a Windows developer who did something like this would be advised to use an actual installer/uninstaller system and not just ad-hoc their own from batch files.

    Yes, they are, and yet yes some of them still do - exactly how the situation is on Linux.  You don't need to put an ad-hoc install/uninstall script in your package when you can use a .rpm or whatever and get everything automatically handled by the system for you, but people still do.

    @blakeyrat said:

    @DaveK said:

    From the MSDN docs for CreateProcess:

    If you do pass NULL for lpApplicationName, use quotation marks around the executable path in lpCommandLine, as shown in the example below.

    Same problem.  Same proposed solution.  Different standard in operation when you criticise one and not the other.

    What the fuck? We're talking about uninstalling a program! We're NOT talking about "typing a path into the command line without quotes."

    Yes, that is exactly what we're talking about; we're talking about the author of an uninstall script typing a path into a command line without quotes and thereby causing a disaster. 

    @blakeyrat said:

    What is wrong with you people? How did you so seamlessly change the subject so that we're talking about some stupid little implementation detail of how the uninstaller works, instead of talking about the uninstaller itself?

    Because that was where the bug arose, and this is TDWTF, not TD-well-planned-and-implemented-code-with-no-bugs.

    @blakeyrat said:

    @DaveK said:
    Well duh!  The problem in this case was that there was a bug in the script written to automate it.

    Yes, but my point is... writing a script like this is the best way to accomplish the task of uninstalling the program? Seriously? Because it sucks in many fundamental ways. (As should be obvious about now...)

    And if there's a better way, then why the hell aren't they using it?

    That is a valid criticism.

    Yes, but it's a criticism that doesn't apply just to Linux, so it's not valid of you to try and use it as a stick to flog your favourite chip-on-your-shoulder with.

     

     

    Apropos of nothing, is it crunch time at your work by any chance? 

     



  • @blakeyrat said:

    [ . . . ] anybody who's actually used Vista knows it was a perfectly fine, and finished, product. [ . . . ]

    If that's what you call fine and finished, you have low standards.

    @blakeyrat said:


    Or maybe he was just pasting in random URLs he came across. Maybe he's relying his instructions from Planet Weeboo, which broadcasts directly into his brain from a secret UFO factory owned by the Illuminati in Area 51. Who fucking knows.

    I do.  I do, because I read people's posts and try to understand them, rather than deliberately try to fail just so that I can have something to be grouchy about.  The fact that it was a response to your "release early and often" comment made it obvious where the OP was coming from.

    cheer up blakey!


    You're in a lousy mood today.  Hope the cartoon cheered you up!

     



  •  I'll give blakeyrat another point. IMHO the software developers of Linux are excellent; the best around. But the QA is terrible.This thread started with a bug in an uninstall script (or was it an install script?). Wasn't the final version tested before being shipped? If it had been tested at all it would have been noticed and set off alarms everywhere.

    Another weakness of open source software is documentation. Sometimes not enough (latest version out now, documentation later this year), sometimes inconsistant (progname --help and man progname and http://progname.sourceforge.net/documentation.html don't match), sometimes too much (blog for how to install Apache is from 2002, but still on the net). How can we teach Open Source Programmers that you NEVER release code until you've updated the documentation?



  • @DaveK said:

    @blakeyrat said:

    [ . . . ] anybody who's actually used Vista knows it was a perfectly fine, and finished, product. [ . . . ]

    If that's what you call fine and finished, you have low standards.

    Ok. What was unfinished? Specifically?

    @DaveK said:

    I do. I do, because I read people's posts and try to understand them, rather than deliberately try to fail just so that I can have something to be grouchy about. The fact that it was a response to your "release early and often" comment made it obvious where the OP was coming from.

    You assume you do. You don't have any more insight into the statement than I do, and for all you know the aliens from Weeboo explanation is the correct one. The fact that your interpretation of his argument is factually incorrect... well, that doesn't help.

    @DaveK said:

    You're in a lousy mood today. Hope the cartoon cheered you up!

    It's an XKCD cartoon, so now I feel suicidal.



  •  @AndyCanfield said:

    I'll give blakeyrat another point. IMHO the software developers of Linux are excellent; the best around. But the QA is terrible.This thread started with a bug in an uninstall script (or was it an install script?). Wasn't the final version tested before being shipped? If it had been tested at all it would have been noticed and set off alarms everywhere.

    The bug occured in version 1.4.31, which, having an odd microversion, is traditionally not a final release.

    This is an excellent example of Linus's Law. 

     



  • @blakeyrat said:

    Puns are never funny. But explaining the pun makes it somehow even less funny, driving your humor quotient well into the negative values.

    I'd expected it to become less than one. I also expected my account to be approved by now, so there. Negative hq funny.



  • Puns are Never Funny

    Yet, Fun can often be rather "punny"...Personally though, I do prefer the humer created by Spoonerisms.



  • @intertravel said:

    @El_Heffe said:
    For the last 20+ years I have done whatever I want on my computer and Bill Gates has never once tried to stop me.
    Oh, maybe it's just me then. A couple of years back I tried to delete a file from \system32\ and Bill landed his helicopter on the roof, abseiled down the outside of my building, smashed in through the window, and grabbed the mouse from my hand.
    You can download a hotfix for that.



  • @TheCPUWizard said:

     Of course, there is an "Easy" way to prove me wrong. If each perons who replies to the remainder of this thread sends me $5K (hey , thats easy), then I will retract my statement.

    If anyone was planning on doing that, I'll retract his statement for half the price.



  • @DaveK said:

    It's an old and well-known problem (see Unix Hater's Guide, for example), and it exists on all of *nix and NT/9x architectures.
    Don't worry, DOS also supports spaces in filenames, except since no DOS programs had any kind of quoting/escaping support, it was very uncommon to find any (and when you did, it was very very hard to get rid of).@DaveK said:
    b) there are still badly-written windows programs that come with crappy .bat file installer scripts, so what's your point? 
    Hah, that reminded me of Xonar D2X firmware update installer - on 64bit Windows, it launched a batch file to copy a driver that the installer placed to SysWOW64\drivers to the 64bit System32\drivers directory - except that I had (at the time 32bit only) TakeCommand installed, and the batch file ran through it, which resulted in it not doing anything (and the updater then failing, since the driver didn't load).


Log in to reply