Author Topic: Linux 21.0.1.0 and Perforce History glitches  (Read 8526 times)

Dan

  • SlickEdit Team Member
  • Senior Community Member
  • *
  • Posts: 2897
  • Hero Points: 153
Re: Linux 21.0.1.0 and Perforce History glitches
« Reply #15 on: December 15, 2016, 09:16:17 PM »
OK.  Let me see if I can reproduce that. I hit something else too.

Thanks for being a multi-decade fan.  Are you a Perforce fan too?

mrothman

  • Senior Community Member
  • Posts: 122
  • Hero Points: 1
Re: Linux 21.0.1.0 and Perforce History glitches
« Reply #16 on: December 15, 2016, 09:36:58 PM »
OK, I think I have (sort of) a better answer. I went ahead and changed my launcher (similar to a Windows shortcut) to launch 21 instead of 18.  Immediately the problem went away, even if I had the P4DIFF variable set.  I am not a Linux or xfce or shell guru, but I suspect that the launcher doesn't run ("source") the startup file which has the environment variables.  Or putting it another way, SlickEdit isn't running from within a shell in that case.  (Yes, I can see the variables in SlickEdit's process buffer, but I think that's just because SlickEdit starts that buffer using a shell, so it DOES run the .cshrc startup file).  (In case I haven't said this, I shut down and start SlickEdit after each of these tweaks).  Here's the final kicker: if I run SlickEdit 18 directly from the interactive terminal, it has the same problem!!!

So this entire thing is nothing new after all - the problem was just a side effect of running SlickEdit in a way that I don't normally run it (directly from a terminal window rather than from a launcher button).  So I hope I have diagnosed a real problem for you guys (which happens if someone runs from a terminal and has the P4DIFF environment variable set to SlickEdit's diff tool), but it doesn't affect me at all, since I don't normally run that way!

mrothman

  • Senior Community Member
  • Posts: 122
  • Hero Points: 1
Re: Linux 21.0.1.0 and Perforce History glitches
« Reply #17 on: December 15, 2016, 09:44:54 PM »
Just saw your question about Perforce.  SlickEdit was my own choice, many years ago (it was "Visual" SlickEdit - or maybe even earlier?  Was there an earlier?), and still haven't had a good reason to switch.  Your macro language was a very clever attractor - after you've built a fair number of things you depend on, it's very hard to walk away!  But seriously, all my jobs have involved multi-language development and several have been multi-OS, and SlickEdit just kept being the best for maintaining projects across those boundaries.

Perforce is the system used by my company, and I have no objections to it - for the most part it does what you'd want and does it well.  In my time I've used RCS, CVS, Clearcase, Perforce and several others I can no longer remember: all had strengths and weaknesses.  I'm sorry to NOT have used some of the more currently popular systems like Git, Mecurial, etc.  Just to be hip.

mrothman

  • Senior Community Member
  • Posts: 122
  • Hero Points: 1
Re: Linux 21.0.1.0 and Perforce History glitches
« Reply #18 on: December 15, 2016, 09:56:40 PM »
I couldn't resist, tried one more thing.  I looked up how to set an environment variable inside an xfce launcher.  I then did that and used it to set the P4DIFF variable to SlickEdit's diff.  Sure enough, the problem came back.

So it's clearly something to do with having that variable set within the environment that then runs the application. Good luck.  As I said, at least for me, there's no urgency since I always run from the launcher anyway.

Dan

  • SlickEdit Team Member
  • Senior Community Member
  • *
  • Posts: 2897
  • Hero Points: 153
Re: Linux 21.0.1.0 and Perforce History glitches
« Reply #19 on: December 15, 2016, 09:57:33 PM »
It was Visual SlickEdit until around version 8.0 I think.  Before Visual SlickEdit 1.0 in late 1993,  there was a textmode version (SlickEdit) for many years (starting in 1988).

I've written 95% of our version control support, and I always find Perforce hard to get setup. I've played with the vast majority of the systems out there, and a few hardly anybody has heard of (remember Microsoft Deltas?), and I always find Perforce's setup tedious.  Maybe it's just because I'm not great at "reading instructions" and it requires a little more attention.  Once it's setup, it's fine.  To be fair, I've never setup ClearCase, or Microsoft Team.  So perhaps it's just the size.

Dan

  • SlickEdit Team Member
  • Senior Community Member
  • *
  • Posts: 2897
  • Hero Points: 153
Re: Linux 21.0.1.0 and Perforce History glitches
« Reply #20 on: December 15, 2016, 09:58:14 PM »
I couldn't resist, tried one more thing.  I looked up how to set an environment variable inside an xfce launcher.  I then did that and used it to set the P4DIFF variable to SlickEdit's diff.  Sure enough, the problem came back.

So it's clearly something to do with having that variable set within the environment that then runs the application. Good luck.  As I said, at least for me, there's no urgency since I always run from the launcher anyway.

Thanks for double checking that.

Dan

  • SlickEdit Team Member
  • Senior Community Member
  • *
  • Posts: 2897
  • Hero Points: 153
Re: Linux 21.0.1.0 and Perforce History glitches
« Reply #21 on: December 16, 2016, 02:12:33 PM »
Have you had any trouble submitting files?

mrothman

  • Senior Community Member
  • Posts: 122
  • Hero Points: 1
Re: Linux 21.0.1.0 and Perforce History glitches
« Reply #22 on: December 16, 2016, 04:07:51 PM »
I haven't submitted any, and it's hard for me to do so because our system has perforce integration that will immediately sweep that into the company build system.  The main work I'm doing isn't ready to have that happen.  I may be able to test on another branch in a few days.

Dan

  • SlickEdit Team Member
  • Senior Community Member
  • *
  • Posts: 2897
  • Hero Points: 153
Re: Linux 21.0.1.0 and Perforce History glitches
« Reply #23 on: December 16, 2016, 04:12:50 PM »
I found an issue yesterday.  I don't understand why it seems to be Linux only, but it does.  The good news is it should be hotfixable.