[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Re: Re: Re: Crash in repo browser

From: Simon Large <simon.tortoisesvn_at_googlemail.com>
Date: Wed, 28 Jul 2010 19:44:16 +0100

On 28 July 2010 17:27, Stefan Fuhrmann <stefanfuhrmann_at_alice-dsl.de> wrote:
> Hi Simon,
>
> sorry that it took me so long to react.
> I've basically been sick for the last 2 weeks.

Hope you're feeling better now.

>
> On 2010-07-21, Simon Large wrote:
>> On 18 July 2010 22:50, Simon Large <simon.tortoisesvn_at_googlemail.com> wrote:
>> I have now deleted the SimonsCrashHack key and I can confirm that the
>> latest changes have fixed the crash.
>>
>> The issues listed below were unrelated and still stand of course.
>
> Great news! Thanks for your patience.
>
>> > Other things worth noting, unrelated to the crash problem:
>> >
>> > While the editor is open the repo browser window does not repaint if
>> > you swap back to it. In fact it takes about a second after the editor
>> > exits before it redraws itself.
>
> The reason is / was that the "edit file" command
> got executed in the repobrowser's message loop -
> blocking all other events until the editor was
> closed. I changed that in r19948.

Great!

>> > Once when I tried to open version.txt for edit I got a message in the
>> > checkout progress to say that it was a file not a directory. I only
>> > saw that once and cannot reproduce it.
>
> This could happen if there is a hole in my temporary
> working copy code. Possibly some file access race.
> I will have a look at this next weekend.
>
>> > Other attempts to open version.txt were OK, but when the editor closes
>> > it always shows as modified. I note that the working base has
>> > svn:eol-style set to native and the working copy has no such property.
>> > As a result, the working copy appears to have LF line-endings and the
>> > working base has CR-LF.
>
> Hm. That sounds like a problem with the SVN libs.
> However, I cannot reproduce it.

Strange, not sure why my system should be any different. The chain of
events was:

* Open repo browser on TSVN trunk.
* Right click and edit version.txt
* Exit the editor without saving
* File is marked as modified in the commit dialog
* Diff reveals that the WC copy has no svn:eol-style property
* Second diff window indicates that the WC file has LF line-endings

Another thing I notice when I explore to the temp folder - there are
lots of them left over from my previous experiments. Normally TProc
deletes its temporary objects when they are more than a day old, but
maybe it only does that correctly for files and not folders.

>> > I am testing on a single core CPU in case it makes any difference.
>
> It makes a huge difference. We obviously now see the
> opposite of what happened in the past: multi-threaded
> applications seldom crashed on single core machines
> but often on multi-core ones due to missing
> synchronization. Most people today develop on SMP
> machines where the background threads have a
> reasonable chance to finish their jobs just quickly
> enough. So, controlled tear-down seems to be the
> most common issue today.

Simon

-- 
:       ___
:  oo  // \\      "De Chelonian Mobile"
: (_,\/ \_/ \     TortoiseSVN
:   \ \_/_\_/>    The coolest Interface to (Sub)Version Control
:   /_/   \_\     http://tortoisesvn.net
------------------------------------------------------
http://tortoisesvn.tigris.org/ds/viewMessage.do?dsForumId=757&dsMessageId=2639844
To unsubscribe from this discussion, e-mail: [dev-unsubscribe_at_tortoisesvn.tigris.org].
Received on 2010-07-28 20:44:28 CEST

This is an archived mail posted to the TortoiseSVN Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.