You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by "Peter N. Lundblad" <pe...@famlundblad.se> on 2004/10/31 16:23:19 UTC
r11557
Hi,
In 1.1.x/STATUS philip asks how to check if this works. There is a recipe
in issue 1905 that could be used. You just have to merge the addition of a
file with spaces in its name and check the copyfrom URL.
JUst so you know if you want to vote.
Regards,
//Peter
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: r11557
Posted by John Peacock <jp...@rowman.com>.
Julian Foad wrote:
>
> Tee-hee! Feeling red in the face yet, John?
Yeah, responded too early in the morning...
John
--
John Peacock
Director of Information Research and Technology
Rowman & Littlefield Publishing Group
4501 Forbes Boulevard
Suite H
Lanham, MD 20706
301-459-3366 x.5010
fax 301-429-5748
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: r11557
Posted by Julian Foad <ju...@btopenworld.com>.
John Peacock wrote:
> Folker Schamel wrote:
>> BTW, it would be also nice to change the handling of parallel running
>> commits:
>> Instead of "the winner takes it all" it may be more fair to take
>> a proportional number of lines from each commit and finalize them.
Brilliant! Very well spoofed, Folker!
> Umm, no. Commits are atomic; either all lines are committed or no lines
> are committed. There is no possible middle ground which is consistent.
> The worst thing that happens if a second commit cannot proceed because
> of an overlapping prior commit, is that the second committer gets an
> "out of date" warning and has to update (and possibly handle conflicts)
> before committing. The system cannot possibly do that automatically.
>
> John
Tee-hee! Feeling red in the face yet, John?
- Julian
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: r11557
Posted by John Peacock <jp...@rowman.com>.
Folker Schamel wrote:
> BTW, it would be also nice to change the handling of parallel running
> commits:
> Instead of "the winner takes it all" it may be more fair to take
> a proportional number of lines from each commit and finalize them.
Umm, no. Commits are atomic; either all lines are committed or no lines are
committed. There is no possible middle ground which is consistent. The worst
thing that happens if a second commit cannot proceed because of an overlapping
prior commit, is that the second committer gets an "out of date" warning and has
to update (and possibly handle conflicts) before committing. The system cannot
possibly do that automatically.
John
--
John Peacock
Director of Information Research and Technology
Rowman & Littlefield Publishing Group
4720 Boston Way
Lanham, MD 20706
301-459-3366 x.5010
fax 301-429-5747
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: r11557
Posted by Folker Schamel <sc...@spinor.com>.
> Let me just make it clear, crystal clear, we should hunt down, kill
> and destroy all bugs by our test framework. Mr. Lundblad's biggest
> mistake was that when he has cornered that nasty bug somewhere in
> libsvn_client, near diff.c, he outsourced testing of this bug to the
> foreing shell scripts, and potentially let this bug at large. I won't
> never do that. I will stop at absolutely nothing to hunt down,
> capture or kill the bugs wherever they are, whatever it takes.
BTW, how does svn handle hanging and bulging commits? Are they finalized?
In case of modern computers, is it possible to print out commits?
I want to be sure that my commit is not manipulated afterwards or lost.
BTW, it would be also nice to change the handling of parallel running commits:
Instead of "the winner takes it all" it may be more fair to take
a proportional number of lines from each commit and finalize them.
Folker
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: r11557
Posted by Jani Averbach <ja...@jaa.iki.fi>.
On 2004-11-01 13:49-0500, Mark Benedetto King wrote:
> On Mon, Nov 01, 2004 at 10:14:26AM -0600, kfogel@collab.net wrote:
> > Jani Averbach <ja...@jaa.iki.fi> writes:
> > > Let me just make it clear, crystal clear, we should hunt down, kill
> > > and destroy all bugs by our test framework. Mr. Lundblad's biggest
> > > mistake was that when he has cornered that nasty bug somewhere in
> > > libsvn_client, near diff.c, he outsourced testing of this bug to the
> > > foreing shell scripts, and potentially let this bug at large. I won't
> > > never do that. I will stop at absolutely nothing to hunt down,
> > > capture or kill the bugs wherever they are, whatever it takes.
> > >
> > > Sorry, couldn't resist.
> >
> > Oh. My. Goodness.
> >
> > You have us laughing out loud here in the Chicago office, Jani! :-)
> >
>
> Jani wins the "right post and the right time" award. But does
> it pass the global regression-tests?
I will never let any global regression test to veto the correctness
of Subversion.
BR, Jani
--
Jani Averbach
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: r11557
Posted by Mark Benedetto King <mb...@lowlatency.com>.
On Mon, Nov 01, 2004 at 10:14:26AM -0600, kfogel@collab.net wrote:
> Jani Averbach <ja...@jaa.iki.fi> writes:
> > Let me just make it clear, crystal clear, we should hunt down, kill
> > and destroy all bugs by our test framework. Mr. Lundblad's biggest
> > mistake was that when he has cornered that nasty bug somewhere in
> > libsvn_client, near diff.c, he outsourced testing of this bug to the
> > foreing shell scripts, and potentially let this bug at large. I won't
> > never do that. I will stop at absolutely nothing to hunt down,
> > capture or kill the bugs wherever they are, whatever it takes.
> >
> > Sorry, couldn't resist.
>
> Oh. My. Goodness.
>
> You have us laughing out loud here in the Chicago office, Jani! :-)
>
Jani wins the "right post and the right time" award. But does
it pass the global regression-tests?
--ben
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: r11557
Posted by kf...@collab.net.
Jani Averbach <ja...@jaa.iki.fi> writes:
> Let me just make it clear, crystal clear, we should hunt down, kill
> and destroy all bugs by our test framework. Mr. Lundblad's biggest
> mistake was that when he has cornered that nasty bug somewhere in
> libsvn_client, near diff.c, he outsourced testing of this bug to the
> foreing shell scripts, and potentially let this bug at large. I won't
> never do that. I will stop at absolutely nothing to hunt down,
> capture or kill the bugs wherever they are, whatever it takes.
>
> Sorry, couldn't resist.
Oh. My. Goodness.
You have us laughing out loud here in the Chicago office, Jani! :-)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: r11557
Posted by Jani Averbach <ja...@jaa.iki.fi>.
On 2004-10-31 20:01+0100, Peter N. Lundblad wrote:
> On Sun, 31 Oct 2004, Philip Martin wrote:
>
> > "Peter N. Lundblad" <pe...@famlundblad.se> writes:
> >
> > > In 1.1.x/STATUS philip asks how to check if this works. There is a recipe
> > > in issue 1905 that could be used. You just have to merge the addition of a
> > > file with spaces in its name and check the copyfrom URL.
> > >
> > Oh, I know how to do it, I was just being awkward to make a point. If
> > someone cares enough about a bug to fix it then writing a regression
> > test ensures that it stays fixed.
> >
> Also, it is a more general question. We have many bug fixes that go
> in without test cases being written.
And it is bad.
Let me just make it clear, crystal clear, we should hunt down, kill
and destroy all bugs by our test framework. Mr. Lundblad's biggest
mistake was that when he has cornered that nasty bug somewhere in
libsvn_client, near diff.c, he outsourced testing of this bug to the
foreing shell scripts, and potentially let this bug at large. I won't
never do that. I will stop at absolutely nothing to hunt down,
capture or kill the bugs wherever they are, whatever it takes.
Sorry, couldn't resist.
Br, Jani
--
Jani Averbach
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: r11557
Posted by "Peter N. Lundblad" <pe...@famlundblad.se>.
On Sun, 31 Oct 2004, Philip Martin wrote:
> "Peter N. Lundblad" <pe...@famlundblad.se> writes:
>
> > In 1.1.x/STATUS philip asks how to check if this works. There is a recipe
> > in issue 1905 that could be used. You just have to merge the addition of a
> > file with spaces in its name and check the copyfrom URL.
> >
> Oh, I know how to do it, I was just being awkward to make a point. If
> someone cares enough about a bug to fix it then writing a regression
> test ensures that it stays fixed.
>
This is a good point and all that, and I really could get better at
writing regression tests. I don't think, however, that this belongs in the
STATUS file for the branch, but on the list. Also, it is a more general
question. We have many bug fixes that go in without test cases being
written.
Regards,
//Peter
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: r11557
Posted by Philip Martin <ph...@codematters.co.uk>.
"Peter N. Lundblad" <pe...@famlundblad.se> writes:
> In 1.1.x/STATUS philip asks how to check if this works. There is a recipe
> in issue 1905 that could be used. You just have to merge the addition of a
> file with spaces in its name and check the copyfrom URL.
>
> JUst so you know if you want to vote.
Oh, I know how to do it, I was just being awkward to make a point. If
someone cares enough about a bug to fix it then writing a regression
test ensures that it stays fixed.
In this particular case I assume the person who committed the change
tested it as described in 1905 and that it works. Who is responsible
for testing it on the branch? Is the release manager expected to
merge the STATUS changes, run the regression tests and then look for
manual regression tests as well? Does the release manager have to run
the manual regression tests before the merge to observe the failure,
thus ensuring that they are being done correctly?
Even just staying fixed should not be taken for granted. In the past
some of my libsvn_wc modifications have changed, i.e. "broken", things
that were not regression tested.
--
Philip Martin
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: r11557
Posted by "Peter N. Lundblad" <pe...@famlundblad.se>.
On Sun, 31 Oct 2004, Jani Averbach wrote:
> On 2004-10-31 17:23+0100, Peter N. Lundblad wrote:
> > In 1.1.x/STATUS philip asks how to check if this works. There is a recipe
> > in issue 1905 that could be used. You just have to merge the addition of a
> > file with spaces in its name and check the copyfrom URL.
> >
> > JUst so you know if you want to vote.
> >
>
> I am doing a regression test for that.
>
Thank you:-)
//Peter
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: r11557
Posted by Jani Averbach <ja...@jaa.iki.fi>.
On 2004-10-31 17:23+0100, Peter N. Lundblad wrote:
> Hi,
>
> In 1.1.x/STATUS philip asks how to check if this works. There is a recipe
> in issue 1905 that could be used. You just have to merge the addition of a
> file with spaces in its name and check the copyfrom URL.
>
> JUst so you know if you want to vote.
>
I am doing a regression test for that.
Br, Jani
--
Jani Averbach
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org