#485 Mercurial Repository error

KevinKelley Fri 20 Mar 2009

Not used to Mercurial; installed TortoiseHg, no setup but entering my name, then pointed it at the repository. contrib came in fine, just a few files; but fan-1.0 failed after about a minute of data with this:

requesting all changes
adding changesets
adding manifests
adding file changes
*** transaction abort!
*** rollback completed
*** global name 'IncompleteRead' is not defined
[command interrupted]

Should I be doing something differently, or is something wrong?

f00biebletch Fri 20 Mar 2009

What version of tortoise are you using? I am on 0.6 and if I enter

http://hg.fandev.org/repos/fan-1.0

for the source path, everything is fine.

KevinKelley Fri 20 Mar 2009

Version is 0.7; just downloaded today from "latest stable".

I just noticed "Hg" post from a couple weeks ago about bootstrapping through v1002 so tried that; same error. I've tried both tortoiseHg and cmd line Hg; same error, but that gives a python walkback starting off "unknown exception".

Maybe I didn't configure something? Weird.

f00biebletch Fri 20 Mar 2009

I looked at 0.7, it looks a little scary; use 0.6, it has a standard windows installer and works like a champ: http://fileforum.betanews.com/detail/TortoiseHg/1210542371/1/

KevinKelley Fri 20 Mar 2009

I'll try that now. I've sent a bug report to the Selenic guys; will uninstall 0.7 and see if 0.6 works better. Appreciate the quick feedback!

KevinKelley Fri 20 Mar 2009

Well, if any of this nuisance is my fault I apologize. But I've been using Subversion for long time -- quick count, I've got 23 projects checked out with TortoiseSVN, besides using it for years to hold many many megs of all my code going back 25 years and whatever else I want never to lose, and never problem one.

And first try with Hg I get this hassle. Anyhow, after backing down to TortoiseHg v0.6, I got an error again; so went to command line Hg and tried and finally success. So for the moment I'm chalking it up to Hg weirdness and getting back to work.

Sorry to bother y'all, and thanks for the help.

brian Fri 20 Mar 2009

And first try with Hg I get this hassle. Anyhow, after backing down to TortoiseHg v0.6, I got an error again; so went to command line Hg and tried and finally success. So for the moment I'm chalking it up to Hg weirdness and getting back to wor

Sorry for your trouble. Although keep in mind that Mercurial itself is released software and seems pretty rock solid. TortoiseHG is really a separate project which is still very much a development project (it just happens to bundle core hg). I notice all sorts of weird bugs and annoyances with TortoiseHG, but I've used 0.4, 0.5, and 0.6 and never seen any issues like that. But I've found its developers to be very quick to respond and eager to help solve issues.

KevinKelley Fri 20 Mar 2009

My frustration is/was with my computer, which is loaded with so much by now that I'm sure it's really my own fault at bottom. I hadn't loaded Hg yet, since Svn has been so solid for me, but I've known I'd eventually want it for something, so now that's done and I'm a step ahead for the trouble. And complaining about free software is incredibly senseless, especially when it's the choice because it's better than most commercial stuff.

I do appreciate the quick and useful help I got, and had no right to expect. And if I can help in any way to figure out what went wrong, to make it better yet, I'll be glad to.

Cheers,

Oh yeah. This might be a good time for me to add, I've been digging hard in this for not long but enough to be sure of this: this is an extremely well-engineered, extremely high-quality product. As I said I spend a lot of time working with or in open-source code, all across the spectrum, and this Fan project, from top to bottom, is as pretty and sweet as anything I've ever seen.

There are a few other open-source products out there at this level, but not many, and I've seen none that hadn't had years of widely supported development to get to this stage.

KevinKelley Sat 21 Mar 2009

I got this reply to my bug report, in case it might apply I'm passing it on here. Fwiw, I was getting pretty consistent failures, at a guess about a minute into the process.

That's a glitch in Mercurial's error handling (Peter Arrenbrecht posted a patch for this to mercurial-devel last week, so it might be fixed in the latest version), but from what I can tell, it's triggered if the server cuts you off. How long does it take before you get the error? Is that time constant?

Login or Signup to reply.