I've just seen on the DL page, there is a new CC-version on the fast-changing market.
As it's called 4.20, my question is:
Is it the follow-up of 4.19 or the new release of the 4.6x-strand?
I've read here that the public release of the 4.6x-strand would possibly be named 4.20.

And of course: ! IT'S DEVELOPMENT STUFF, USE AT OWN RISK !
I know that, just being curious.

---

Gruesse vom Saenger

For questions about Boinc look in the BOINC-Wiki

> And of course: ! IT'S DEVELOPMENT STUFF, USE AT OWN RISK !
> I know that, just being curious.

Yeah, the whole world wants to know!

There is not even a mention in the developers mailing list.

Maybe the information is stored in the database. That means we won't see it for another week or two ...

---

4.20 is what 4.67 would have been. Alpha software all the way. ;)

---

v4.20 is v4.66 + latest bug-fixes, and is the latest alpha-build.

v4.2x is the release-branch, and should now only get bug-fixes till the next public release-version is out, this is planned to happen within a fortnight, but as always some unforseen bugs can delay things...

Till v4.2x finished alpha-testing, as always you can try it out but don't complain if there's a new build every day or something...


Since they needs the version-numbers, they're also resetting the development-branch back to v4.50. As always, this branch will get new features, but no new build should be released before v4.2x is out.

Hi All,
4.20 seems to be working great. I have it on one of my machines. I like the new features, especially suspending WU's to let ones closer to time finish. A very nice touch. Was a large download On dial up though. Over 11MB.


Have A Great Day And A Better Tomorrow!

Regards,

Rocky

---

www.boincsynergy.com

The Dev page offers 4.19 (official release), 4.20 developmental release (dated 8 feb) and 4.62 developmental release (dated January 24).

So I'm guessing that 4.66 was removed and turned into 4.20. Since 4.66 was pulled, it was replaced on the page with the old 4.62. Is this correct?

tony

---

> The Dev page offers 4.19 (official release), 4.20 developmental release (dated
> 8 feb) and 4.62 developmental release (dated January 24).
>
> So I'm guessing that 4.66 was removed and turned into 4.20. Since 4.66 was
> pulled, it was replaced on the page with the old 4.62. Is this correct?
>
> tony
>

v4.62 is still showing up since this is the latest non-windows-development-release, but for some reason they're also showing this old windows-build.

They're having some problems to build the non-windows-versions, and have therefore not released updated v4.20-builds yet.

Boinc V4.20 successfully installed. No lost WU. Everything seems fine.

(note. 4.20 is developmental build and is risky to run. It's not a general release. It can cause puter problems. Don't install unless you're willing to risk loss of computer OS and/or Hardware. Don't ask me where to get it. It might do bad things to your puter if you use it. Have I made enough of a disclaimer?)

tony

---

4.20 Now working on WinXPSP2, P4 1.8Ghz, and WinXPhome, Athlon XP 2200.

see below disclaimer

tony

---

I'm sorry, but I just don't understand why they are going backwards in release numbers! (I've never seen that done by anybody, be it alpha/beta/release in almost 30 years of programming!)

After 4.5x-4.64 have been "out and about" and discussed on many forums, and the versions are known to many who stick with "released" versions.

Making 4.66 be 4.20 just makes no sense to me!

Does the client version need to match the server SW version or something?

Does this mean that there could be completely different SW released in the future with is version 4.5x and bears no semblance to that which was in alpha as 4.5x?

---

Basically we were going to be throwing away quite a few potential version numbers.

Minor versions have to be two characters otherwise there is potential for bad things to happen, plus the major versions have to be in sync between the client and server, otherwise the server will ignore the client, and the client will reset all the projects.

So if we kept going, we would have only been allowed 25 or so updates to the client before changing the logic of the client and server completely around, or forcing in mass all the projects to update.

For those who are running CPDN, it would have been a very painful experience.

Since 4.5x and 4.6x were never officially released to the public, we could reclaim those version numbers and hold off any potential compatibility bugs that might result because of a version number scheme change.

I would love to switch over to the windows version number scheme where versions are selected based on feature changes, and the build number is incremented for each new build, but I haven’t been able to sell it to David successfully yet. I’m still working on it.

---

----- Rom
BOINC Development Team, U.C. Berkeley
My Blog

> Basically we were going to be throwing away quite a few potential version
> numbers.
>
> Minor versions have to be two characters otherwise there is potential for bad
> things to happen, plus the major versions have to be in sync between the
> client and server, otherwise the server will ignore the client, and the client
> will reset all the projects.
>
> So if we kept going, we would have only been allowed 25 or so updates to the
> client before changing the logic of the client and server completely around,
> or forcing in mass all the projects to update.
>
> For those who are running CPDN, it would have been a very painful experience.
>
> Since 4.5x and 4.6x were never officially released to the public, we could
> reclaim those version numbers and hold off any potential compatibility bugs
> that might result because of a version number scheme change.
>
> I would love to switch over to the windows version number scheme where
> versions are selected based on feature changes, and the build number is
> incremented for each new build, but I haven’t been able to sell it to David
> successfully yet. I’m still working on it.
>


Pure and utter BS words from someone who clearly has never been resposible for the release of software to the general public... Face it Rom, it's a bozo move!

The design of the SW dictates that you need to be able to "reuse" version numbers! Geeze - get a job at MS where Gates dictates reality!

There's no way in HE!! that anybody with experience would release a version 4.2x which is more advanced than version 4.6x of the same sw claiming that they "needed to save version numbers" so server SW could keep up! (this is a paraphrase, but you get the point!)

Do you expect to have 25 more release before you go to version 5.x? What happens when you release a "real" 4.6x as you now have ~80 subversions before you need to deal with 5.x and yest need to deal with the 4.6x which was posted weeks back?

4.5x->4.6x has been used by many - alpha testers and those who found out about it by the forums on ALL the boinc projects!

This 4.6x->4.20 is complete crap and will viewed as crap by any professional programmer. It's the worst move I've ever seen on any softeare!

Hey, I just got a CD of the update to Windows NT labeled "version 3.9" which fixes the problems with Windows NT 4.0!

---

> I would love to switch over to the windows version number scheme where
> versions are selected based on feature changes, and the build number is
> incremented for each new build, but I haven’t been able to sell it to David
> successfully yet. I’m still working on it.
>

It's not a "windows version number scheme", it's called having people in CVS/RCS/etc that have a frigging clue!

You seem to be a windows guy, and might not understand that problems like this were solved long before windows. It's called "version control" and can be implemented for any source, if you have experience!

---

> It's not a "windows version number scheme", it's called having people in
> CVS/RCS/etc that have a frigging clue!
>
> You seem to be a windows guy, and might not understand that problems like this
> were solved long before windows. It's called "version control" and can be
> implemented for any source, if you have experience!

I didn't say it was the best solution, but it is a solution.

Source control systems generally keep track of individual file versions, in many circles it is still up to the people running the project/product to determine what version a release is declared. Source control systems do not determine this.

All I was trying to get across, is that major and minor version numbers should be based on features, and the build number/revision number should be used for tracking purposes.

Heck, this was a case of damned if I do, and damned if I don't.

The versioning system problems aren't going to go away, they have just been delayed is all.

BTW, comparing our version rules with that of Windows is flawed, namely because our versions are limited by the rules built into the system, theirs are determined by feature sets.

---

----- Rom
BOINC Development Team, U.C. Berkeley
My Blog

i just installed Boinc 4.20, but i have got a few problems:

1.) On Windows 2000 (German) there is no Group Users. So i had to create it.
2.) On Windows Millenium nothing is displayed in the GUI. Reading the messagefiles i can the that the client is running and crunching.
3.) I can not shutdown Windows Millenium anymore. i have to close Boinc manually. When i try to shutdown Windows Millenium it displays a message than an application is running and because of this it can not shutdown.

Here are some messages from stdoutgui.txt:
Windows Socket Error '10061'
Windows Socket Error '10061'
Windows Socket Error '10061'
Ermittlung des BOINC-Systemstatus. Bitte etwas Geduld ...
Ermittlung der BOINC-Rechner-Informationen. Bitte etwas Geduld ...

CAN'T FIND RESULT 06no99ab.16579.31536.409648.187_6
CAN'T FIND RESULT 06no99ab.16579.31536.409648.187_6
CAN'T FIND RESULT 06no99ab.16579.31536.409648.187_6
CAN'T FIND RESULT 14mr04ab.14329.12256.284650.22_2
CAN'T FIND RESULT 14mr04ab.14329.12256.284650.22_2
CAN'T FIND RESULT 14mr04ab.14329.12256.284650.22_2
Ermittlung des BOINC-Systemstatus. Bitte etwas Geduld ...
Ermittlung der BOINC-Rechner-Informationen. Bitte etwas Geduld ..

Here are some messages from stderrgui.txt:

***UNHANDLED EXCEPTION****
Reason: Access Violation (0xc0000005) at address 0x00439937 read attempt to address 0xFFFFFFFF

1: 02/08/05 22:20:55
1: SymGetLineFromAddr(): GetLastError = 126

I'm not happy with 4.20. I didn't like the whole 4.5x-4.6x (basically BOINC manager series) either. There's one or two *good* things about it (suspend and abort wu), but everything else is bad. 1- It's 10x larger than the 4.1x installer. 2- It has this godawful XP-style "wasted space" design (the box at the left that you can't resize or get rid of). 3- You can no longer sort your WU's based on whatever column you choose!!! 4- You have to uninstall old versions of BOINC first. 5- Did I mention that it's 10x larger??? 6- No more progress bars for wu's?? That was one of the better features of the old CC.

I guess I'm stuck with 4.19. At least it's been working fine.

---

> I didn't say it was the best solution, but it is a solution.

No, it's NOT a solution.. It's a complete cluster F! THERE IS NO WAY THAT ANY PROGRAMMER WOULD RELEASE A VERSION 4.2x WHICH IS BEYOND VERSION 4.5X (available to the masses, and not just alpha testers) OF THE SAME SW!

It's, like I said, the ultimate stupid move in releasing SW that I've seen in 30 years!
>
> Source control systems generally keep track of individual file versions, in
> many circles it is still up to the people running the project/product to
> determine what version a release is declared. Source control systems do not
> determine this.

So, "people" determined that 4.2x should be based on a widely available 4.5x and 4.6x of the code... Sounds to me that the "people" didn't grasp the concept that 50 comes after 20!

Source control (if set up and used properly) allows someone to build "4.19" of a project, even if there's a thread for version "4.60".

Basically Rom, it's clear that what you have setup for source control isn't working!

It's also clear that the "powers that be" don't understand that when releasing software, version "4.20" should not come after version "4.64"! Do some need basic math skills to understand this?

>
> All I was trying to get across, is that major and minor version numbers should
> be based on features, and the build number/revision number should be used for
> tracking purposes.

No version/subversion/sub-subversions need to be defined based on "what's the newest!" Seems there's multiple threads in the code where 4.2x "might" have all the stuff from 4.6x, but then again 4.5x-4.6x don't really exist as the version numbers could get re-used to keep the server happy!
>
> Heck, this was a case of damned if I do, and damned if I don't.
>

No it's a case of "damned if you do something REALLY stupid!" You NEVER want to go backwards in version/subversion numbers for a new release!

THERE IS NO REASON THAT BOINC SHOULD RELEASE A VERSION/SUBVERSON LOWER THAN WHAT ANYBODY WHO"S NOT IN THE ALPHA MIGHT BE RUNNING! This is the sign of a very inexperienced programmer.


> The versioning system problems aren't going to go away, they have just been
> delayed is all.

It's not the VC SW.. It's how well it's set up and administrated... I can use a hammer to drive home a screw. Basically, the tool was used incorrectly, but might look to some as a misunderstanding of how the tool can be best used!

>
> BTW, comparing our version rules with that of Windows is flawed, namely
> because our versions are limited by the rules built into the system, theirs
> are determined by feature sets.
>

Rom, it's clear you don't understand the tool! A Hammer can drive home a screw, but if you understand tools, you's say "hey.. Screwdriver"

What I just can't understand is why anybody who's been around software for more than 2 days can try to justify why a version 4.2x should come after a version 4.6x!

Anybody that tries should not be in the business! If they try, it just doesn't make sense!

(hey I just got winNT 3.8 which corrects problems in WinNT v 3.9! I can now transport to china!)

Rom - Tell dave that he's not technical enough to make edicts like this!

---

AZ,

What gives you the right to criticize software that hasn't been released to the public. the last I checked 4.19 is the release. So your comments to this forum are a waste of breath and space. You are insulting and if your so damned smart make donation to Berkeley or join the team and actively participate in the decision making process! Instead of insults and brow beating.

Geese Woody...


Disgusted

---

Light Finger on the button sorry... I am only Human :=)

---

> > And of course: ! IT'S DEVELOPMENT STUFF, USE AT OWN RISK !
> > I know that, just being curious.
>
> Yeah, the whole world wants to know!
>
> There is not even a mention in the developers mailing list.
>
> Maybe the information is stored in the database. That means we won't see it
> for another week or two ...
>

Hi Paul it was discussed you must have missed that issue.. Ill try to find it for you. I knew it was comming so it was mentioned...


EDIT: I started to go through the list and my eyes got glossed over
This issue from yesterday mentions it;

boinc_alpha Digest, Vol 10, Issue 15




Timmy

---