Has anything changed with CVS recently? NB is messing up some source files!

[Thomas Wolf]

I’m working off the trunk and my colleague committed some changes to CVS (he uses Eclipse, I use Netbeans 8.01 with all the latest updates). When I did the usual Project->CVS->Update->Update on my project, the file he had changed didn’t compile due to errors. Here’s what Netbeans shows me as his changes:




The problem is that the actual changes (as shown in Eclipse) are:



Notice that the changes Netbeans is showing at lines 202-203 weren’t actually real? It seems to have simply pasted the lines from 210-211 to 202-203?!

Does anyone know of a recent change to the CVS-related modules that might be causing this? Or is this some cache-corruption issue (although this seems far fetched)? Is there perhaps a bug open for this already?

Thanks for any info,
Tom

[Thomas Wolf]

As a follow-on to my own question: if I simply do a Team->CVS->Checkout…, then the content of the file is actually correct! So the question is what does {File|Project}->CVS->Update->Update do that is different from a pure checkout? That’s where the bug is, I imagine.

thanks again for any help. in the meantime, I will get the developer’s build and see if it’s a problem there.
Tom

<Screen Shot 2014-10-22 at 11.43.51 AM.png>

[Ondrej Vrabec]

I'm pretty sure we haven't change anything in CVS for a couple of
months. So i doubt it's caused by any recent changes.
Is the content of the current file actually corrupted? Or can you see
just a wrong diff in the Search History view?

O.

Post by Thomas Wolf
As a follow-on to my own question: if I simply do a Team->CVS->Checkout…, then the content of the file is actually correct! So the question is what does {File|Project}->CVS->Update->Update do that is different from a pure checkout? That’s where the bug is, I imagine.
thanks again for any help. in the meantime, I will get the developer’s build and see if it’s a problem there.
Tom

<Screen Shot 2014-10-22 at 11.43.51 AM.png>

[Thomas Wolf]

Hi Ondrej,
The actual file that was retrieved is corrupt (that’s why the compile failed). I went into the file using ‘vi’ to make sure.

I have used Netbeans/CVS on this project since 2007 - and I’ve never seen this before. Since more than 20 modules changed the last time NB offered to update itself, II brought up the possibility of a recent change.

Hope this helps,
Tom

p.s. I also moved to Oracle JDK 1.8u25 recently. But I don’t think that makes a difference.

I'm pretty sure we haven't change anything in CVS for a couple of months. So i doubt it's caused by any recent changes.
Is the content of the current file actually corrupted? Or can you see just a wrong diff in the Search History view?
O.

Post by Thomas Wolf
As a follow-on to my own question: if I simply do a Team->CVS->Checkout…, then the content of the file is actually correct! So the question is what does {File|Project}->CVS->Update->Update do that is different from a pure checkout? That’s where the bug is, I imagine.
thanks again for any help. in the meantime, I will get the developer’s build and see if it’s a problem there.
Tom

<Screen Shot 2014-10-22 at 11.43.51 AM.png>

[Ondrej Vrabec]

I did only this change that might be related:
http://hg.netbeans.org/core-main/rev/422854cdc273

Could you send me the two versions of the file (1.42 + 1.43) to my email
so i can test it?

Ondra

Post by Thomas Wolf
Hi Ondrej,
The actual file that was retrieved is corrupt (that’s why the compile failed). I went into the file using ‘vi’ to make sure.
I have used Netbeans/CVS on this project since 2007 - and I’ve never seen this before. Since more than 20 modules changed the last time NB offered to update itself, II brought up the possibility of a recent change.
Hope this helps,
Tom
p.s. I also moved to Oracle JDK 1.8u25 recently. But I don’t think that makes a difference.

I'm pretty sure we haven't change anything in CVS for a couple of months. So i doubt it's caused by any recent changes.
Is the content of the current file actually corrupted? Or can you see just a wrong diff in the Search History view?
O.

Post by Thomas Wolf
As a follow-on to my own question: if I simply do a Team->CVS->Checkout…, then the content of the file is actually correct! So the question is what does {File|Project}->CVS->Update->Update do that is different from a pure checkout? That’s where the bug is, I imagine.
thanks again for any help. in the meantime, I will get the developer’s build and see if it’s a problem there.
Tom

<Screen Shot 2014-10-22 at 11.43.51 AM.png>

[Ondrej Vrabec]

Hmm,
http://hg.netbeans.org/core-main/rev/422854cdc273 may probably be
related only on Windows and in case you had had local modifications in
the file before the update. Is it your case?

Ondra

Post by Ondrej Vrabec
http://hg.netbeans.org/core-main/rev/422854cdc273
Could you send me the two versions of the file (1.42 + 1.43) to my
email so i can test it?
Ondra

Post by Thomas Wolf
Hi Ondrej,
The actual file that was retrieved is corrupt (that’s why the compile
failed). I went into the file using ‘vi’ to make sure.
I have used Netbeans/CVS on this project since 2007 - and I’ve never
seen this before. Since more than 20 modules changed the last time
NB offered to update itself, II brought up the possibility of a
recent change.
Hope this helps,
Tom
p.s. I also moved to Oracle JDK 1.8u25 recently. But I don’t think
that makes a difference.

On Oct 22, 2014, at 2:21 PM, Ondrej Vrabec
I'm pretty sure we haven't change anything in CVS for a couple of
months. So i doubt it's caused by any recent changes.
Is the content of the current file actually corrupted? Or can you
see just a wrong diff in the Search History view?
O.

Post by Thomas Wolf
As a follow-on to my own question: if I simply do a
Team->CVS->Checkout…, then the content of the file is actually
correct! So the question is what does
{File|Project}->CVS->Update->Update do that is different from a
pure checkout? That’s where the bug is, I imagine.
thanks again for any help. in the meantime, I will get the
developer’s build and see if it’s a problem there.
Tom

I’m working off the trunk and my colleague committed some changes
to CVS (he uses Eclipse, I use Netbeans 8.01 with all the latest
updates). When I did the usual Project->CVS->Update->Update on my
project, the file he had changed didn’t compile due to errors.
<Screen Shot 2014-10-22 at 11.43.51 AM.png>

[Thomas Wolf]

Hi Ondra (I’ll send you 1.42 and 1.43 separately - have to get back to the office for that first).
I am on Mac OS X 10.10 - i.e. not Windows. The corrupted file was not modified as far as I know - but since I didn’t have it open as a tab at the time and didn’t remember to do my usual "CVS->Show Changes” first, I can’t be sure. Is there a way I can check that now?

Tom

Hmm,
http://hg.netbeans.org/core-main/rev/422854cdc273 may probably be related only on Windows and in case you had had local modifications in the file before the update. Is it your case?
Ondra

Post by Ondrej Vrabec
http://hg.netbeans.org/core-main/rev/422854cdc273
Could you send me the two versions of the file (1.42 + 1.43) to my email so i can test it?
Ondra

Post by Thomas Wolf
Hi Ondrej,
The actual file that was retrieved is corrupt (that’s why the compile failed). I went into the file using ‘vi’ to make sure.
I have used Netbeans/CVS on this project since 2007 - and I’ve never seen this before. Since more than 20 modules changed the last time NB offered to update itself, II brought up the possibility of a recent change.
Hope this helps,
Tom
p.s. I also moved to Oracle JDK 1.8u25 recently. But I don’t think that makes a difference.

I'm pretty sure we haven't change anything in CVS for a couple of months. So i doubt it's caused by any recent changes.
Is the content of the current file actually corrupted? Or can you see just a wrong diff in the Search History view?
O.

Post by Thomas Wolf
As a follow-on to my own question: if I simply do a Team->CVS->Checkout…, then the content of the file is actually correct! So the question is what does {File|Project}->CVS->Update->Update do that is different from a pure checkout? That’s where the bug is, I imagine.
thanks again for any help. in the meantime, I will get the developer’s build and see if it’s a problem there.
Tom

<Screen Shot 2014-10-22 at 11.43.51 AM.png>