CVSNT, or "how to piss off your users!"



  • We've been using CVS like... forever here. We've always been using some old version of TortoiseCVS on windows XP. Now that we are migrating to Windows 7 that old version no longer works. So we installed the latest version. Which is using CVSNT from "march-hare".



    And that's where it's all going to hell. It shows a poopup once a day, asking you to pay, or to contribute.

    It contains no links to the source (but it has GPL2 license), or any way to get to the source. GPL violation, yay.


  • Discourse touched me in a no-no place

    @Daid said:

    . It shows a poopup[sic] once a day,...
    Was that deliberate, or Freudian?



  • @Daid said:

    It contains no links to the source (but it has GPL2 license), or any way to get to the source. GPL violation, yay.

    not that it makes things much better, but the source code is under the download button on their website, if you scroll down it says:



    "The public CVS server for source code is not guarenteed to be available 24x7, is provided by a volunteer and may be withdrawn at any time:

    cvs -d :pserver:cvs:@cvs.cvsnt.org:/cvsnt co -r CVSNT_2_0_x cvsnt "



    the font is a bit smaller than the rest of the page and the text is well below the "buy now" button, but it's there.


    The rest of the page tries to explain why the licensing is setup the way it is but I never seem to grasp artificial stuff like that but the way i see it that is the source for the previous community edition which they try to charge you 500 bucks for something you apparently can build yourself



  •  TRWTF is using CVS in the first place.  At least migrate to SVN if you can't persuade the guy in charge to use a modern VCS.



  •  never change a properly working system :)



  • @tchize said:

     never change a properly working system :)


    Emphasis on "properly working"; if your users are posting it as an example on TDWTF, it is likely insufficient in this regard.



  • @Daid said:

    We've been using CVS like... forever here. We've always been using some old version of TortoiseCVS on windows XP. Now that we are migrating to Windows 7 that old version no longer works. So we installed the latest version. Which is using CVSNT from "march-hare".



    And that's where it's all going to hell. It shows a poopup once a day, asking you to pay, or to contribute.

    It contains no links to the source (but it has GPL2 license), or any way to get to the source. GPL violation, yay.

    Download the "full" package of WinCVS, it includes an old but Win7-compliant CVSNT: http://prdownloads.sourceforge.net/cvsgui/WinCvs2_0_2-4.zip?download.

    Or use Hg, we recently upgraded to it from CVS and smooth sailing all the way.


  • Trolleybus Mechanic

    @t-bone said:

    @Daid said:
    It contains no links to the source (but it has GPL2 license), or any way to get to the source. GPL violation, yay.

    not that it makes things much better, but the source code is under the download button on their website, if you scroll down it says:

    Don't risk your company's well being on a janky free edition, but also don't blow $500 on a novelty "open sores" product neither.

    Instead, have them feel confident purchasing the $250 version I've spent the last few minutes years putting up a hotlink to their csv developing. You won't believe your eyes or your wallets when you see how nigh non-existant the learning curve from your old, mouldy product to my brand-spanking new one is. It's tens of thousands of potential-man hours saved (from not having to potentially hire potential-trainers).

     



  • @piskvorr said:

    @tchize said:

     never change a properly working system :)

    Emphasis on "properly working"; if your users are posting it as an example on TDWTF, it is likely insufficient in this regard.
     

     

    CVS works properly, CVSNT does not  ;) Because some computer have temporary troubles accessing CVS, you need to migrate all your production server to SVN, with the risk of shutding down production lines? You don't know the context. More over there are other tools than CVSNT ;)



  • @Iago said:

     TRWTF is using CVS in the first place.  At least migrate to SVN if you can't persuade the guy in charge to use a modern VCS.

    Development is, but the 120 other people not in development are still using CVS to archive anything from documents to important emails. We also archive our releases in CVS. It's not something we can replace easly, to give you an idea, the CVS server contains 180GB of data. Spread over 15 repositories.




    And SVN is great, until you discover that you use the CVS tagging system a LOT, and SVN's "tag by making a directory copy" isn't a suitable replacement. It requires changing a lot of workflows and directory structures.



  • Sounds like you've abused your SC for too long. 180GB of stuff that mostly isn't source code doesn't sound good. Time for your financiers to man up and spend some money on decent tools.



  • @jasmine2501 said:

    Sounds like you've abused your SC for too long. 180GB of stuff that mostly isn't source code doesn't sound good. Time for your financiers to man up and spend some money on decent tools.

    Why? What's wrong with this?

    Are you saying it's bed because CVS has trouble handling that many files? Or are you saying it's bad because source control systems should only be used for source code?



  • @t-bone said:

    "The public CVS server for source code is not guarenteed to be available 24x7, is provided by a volunteer and may be withdrawn at any time:
    cvs -d :pserver:cvs:@cvs.cvsnt.org:/cvsnt co -r CVSNT_2_0_x cvsnt "

    [danix@betelgeuse tortoise]$ cvs -d :pserver:cvs:@cvs.cvsnt.org:/cvsnt co -r CVSNT_2_0_x cvsnt

    /cvsnt: no such repository

    [danix@betelgeuse tortoise]$

    0wned! Looks like there's no more repository. :(



  • @The_Assimilator said:

    Or use Hg, we recently upgraded to it from CVS and smooth sailing all the way.

    +1 for Hg. Being able to do experimental branches without fear, and local commits, are like water in the desert. Joel Spolskey hit the nail on the head during his Kiln tour; people using centralized version control systems usually stop version controlling. Since centralize version control is "commit and subject", most developers won't commit large features for a long time, sometimes weeks on end. This is only compounded by the fact that branching and merging is often so painful.

    Then again, being able to even change directory structures is a godsent when you're talking about CVS.



  • @tchize said:

    Because some computer have temporary troubles accessing CVS, you need to migrate all your production server to SVN, with the risk of shutding down production lines?

    Hmm... are you, by chance, working for one particular German company making huge specialized hardware and Java software to go along it?



  • @The_Assimilator said:

    @Daid said:
    We've been using CVS like... forever here. We've always been using some old version of TortoiseCVS on windows XP. Now that we are migrating to Windows 7 that old version no longer works. So we installed the latest version. Which is using CVSNT from "march-hare".



    And that's where it's all going to hell. It shows a poopup once a day, asking you to pay, or to contribute.

    It contains no links to the source (but it has GPL2 license), or any way to get to the source. GPL violation, yay.

    Download the "full" package of WinCVS, it includes an old but Win7-compliant CVSNT: http://prdownloads.sourceforge.net/cvsgui/WinCvs2_0_2-4.zip?download.

    I have both a Windows 7 32 bit system and a Windows 7 64 bit system. The CVSNT package that comes with the latest WinCVS installs fine on the 32 bit system, but the installation hangs on the 64 bit system.
    Here's the workaround:
    Install TortoiseCVS; this will also install a newer version of CVSNT (v2.5.05.x), which installs fine on Windows 7 64 bit. Then you can uninstall all the TortoiseCVS components in the control panel and install just the WinCVS gui.



  • @PJH said:

    @Daid said:
    . It shows a poopup[sic] once a day,...
    Was that deliberate, or Freudian?
     

    does it really matter? it's hillarious, i'm never going to spell it other way from now.



  • Too lazy to pick and post the quotes....but...

    1) By no means does CVS come close to working for any company even halfway atempting to practice ALM [Application Lifecycle Management]. With a few hundred man-hours SVN can be configured and integreated to something that might make the minimum bar.

    2) Revision Control is much more than controlling source code. In a comprehensive approach would should be able to restore your entire state to match any point in time. [This is actually a fun one to test. Make a backup of a random development machine, put it waway for 6 months, and just keep the date handy. When the six months are up, recreate the environment to that you can build the exact same binaries, have the exact same tests, have the exact same documentation. Then compare it for mismatches against the backup. In a recent study that took this approach 3 (count them 3) out of 100 companies who rated themselves "extremely diligent" regarding revision control passed the test]



  • Hydrogyrum leads to madness, so I'm not sure you should be recommending it as a remedy to any particular issues ;-)  (I have to wonder if the folks that picked that name were aware of that. Especially when you use the chemical symbol rather than the planetary one.)

    That said, I still don't quite understand one of the aspects of "distributed" version control, in that if something happens to your local machine, anything that you didn't "push" is lost.  Also, considering that most "centralized" VCS let you build a repository on your local machine, you can pretty much emulate a DVCS by having one local repository, and one remote, and commiting your local changes to the remote one.  (Of course, from what I understand of DVCS is that it integrates these two things in a nice way so that your local version history is somehow reflected in the "pushed" version?)

    I'm also not sure I agree with the animosity toward storing "non-code" documents in VCS.  After all, a file is a file is a file, so why should it matter if the document is computer-readable or human-readable?

    But, yes, unless you have an overall process into which VCS is an integral part, just having VCS doesn't really buy you that much.



  • @TheCPUWizard said:

    [This is actually a fun one to test. Make a backup of a random development machine, put it waway for 6 months, and just keep the date handy. When the six months are up, recreate the environment to that you can build the exact same binaries, have the exact same tests, have the exact same documentation. Then compare it for mismatches against the backup. In a recent study that took this approach 3 (count them 3) out of 100 companies who rated themselves "extremely diligent" regarding revision control passed the test]

    Unless you keep your entire toolchain in revision control (which in many cases would be a licence violation) or never update (or use a toolchain that's no longer updated), you're basically never going to pass that "test".

    What purpose does it serve anyway? You archive the binaries for each release and document which version of the toolchain they were built with, in case newer revisions rely on new toolchain features. I can't think of a situation where being able to rebuild byte-for-byte identical binaries is critical.



  • @mallard said:

    @TheCPUWizard said:
    [This is actually a fun one to test. Make a backup of a random development machine, put it waway for 6 months, and just keep the date handy. When the six months are up, recreate the environment to that you can build the exact same binaries, have the exact same tests, have the exact same documentation. Then compare it for mismatches against the backup. In a recent study that took this approach 3 (count them 3) out of 100 companies who rated themselves "extremely diligent" regarding revision control passed the test]

    Unless you keep your entire toolchain in revision control (which in many cases would be a licence violation) or never update (or use a toolchain that's no longer updated), you're basically never going to pass that "test".

    What purpose does it serve anyway? You archive the binaries for each release and document which version of the toolchain they were built with, in case newer revisions rely on new toolchain features. I can't think of a situation where being able to rebuild byte-for-byte identical binaries is critical.

    1) I have NEVER seen any "tool" that prohibits backup of the original media for archival purposes. If I ever did, rest assured I would not purchase it.

     2) Yes, I keep all tools, utilities, service packs, hot fixes, under revision control.

    3) Byte-4-Byte binaries are a requirement in many areas (Military, Medical, Space, etc..), even getting a waiver for things like different timestamps is a formal process. If you are subjed to a full inspection, you will be handed a "bare metal" machine told to configure it, load the appropriate source (for any version of the inspectors choosing) and replicate the executables and all other generated material.



  • I work for March Hare Software full time as part of the CVSNT development team. It's great to hear that you've found CVSNT useful and have 120 people using it. The announcement about the changes to the project is here: http://www.evscm.org/modules/Downloads/

    CVSNT is NoT CVS, CVSNT&EVS are the only open source versioning systems with failsafe audit, single point merge with merge tracking, user defined and atomic changesets and many other features. To my knowledge neither hg nor svn has these features. As another person pointed out - features like these are required to implement comprehensive change management/ALM. CVSNT is a fully modern change management engine, including support for versioning binary files (efficiently), access control, single-sign-on, promotion/release levels, reserved and unreserved versioning, centralised control of methodology etc etc.

    If you have any questions please contact March Hare Software via the web site.



  • @aaaaa said:

    I work for March Hare Software full time as part of the CVSNT development team. It's great to hear that you've found CVSNT useful and have 120 people using it. The announcement about the changes to the project is here: http://www.evscm.org/modules/Downloads/

    CVSNT is NoT CVS, CVSNT&EVS are the only open source versioning systems with failsafe audit, single point merge with merge tracking, user defined and atomic changesets and many other features. To my knowledge neither hg nor svn has these features. As another person pointed out - features like these are required to implement comprehensive change management/ALM. CVSNT is a fully modern change management engine, including support for versioning binary files (efficiently), access control, single-sign-on, promotion/release levels, reserved and unreserved versioning, centralised control of methodology etc etc.

    If you have any questions please contact March Hare Software via the web site.

    Seriously? You picked a thread about your software with the title "how to piss off your users" on TheDailyWTF to promote your software? Wow. Just wow.



  • @dohpaz42 said:

    Seriously? You picked a thread about your software with the title "how to piss off your users" on TheDailyWTF to promote your software? Wow. Just wow.

    What's funny isn't that here's here, but that he didn't even read the thread!

    I'd actually love it is owners of software I bitch about would find the thread and come in and try to defend himself. This guy didn't defend or apologize for the software, he just pasted some feel-good copy from the website.



  • Of course I read the thread, I specifically looked for issues in the thread and tried as best to respond to them as I could.

    My main aim was to let the people who use our software know that they can contact us directly, we have always been and are still open and listening and actively developing useful software.


Log in to reply