You have a corrupted app_info.xml file. If any tiny thing is missing or wrong in the file it can't parse (read) it right. It then causes all work units that have been downloaded before it was corrupt to error out. This is why you lost them. Shut down BOINC and edit or replace your app_info.xml file then restart.

BTW, if you are also attached to SETI Beta remember to copy the application section and paste it in and change the version to read 517. If not it will error out again.

---

Jim

Some people plan their life out and look back at the wealth they've had.
Others live life day by day and look back at the wealth of experiences and enjoyment they've had.

10/26/2006 8:42:01 AM|SETI@home|Found app_info.xml; using anonymous platform
10/26/2006 8:42:01 AM|SETI@home|Unparsed line in app_info.xml: <file_ref>
10/26/2006 8:42:01 AM|SETI@home|Unparsed line in app_info.xml: <file_name>setiathome-5.15-DeNitro-em64t.exe</file_name>
10/26/2006 8:42:01 AM|SETI@home|Unparsed line in app_info.xml: <main_program/>

Whether you are attached to BOINC or not, you do have an error in your app_info.xml file.
I believe the line I have highlighted above tells what is wrong.
You have a bad setof xml tags. The section should read:
<main_program>
some lines of code
</main_program>

The closing tag should be </main_program> not <main_program/>

---

Jim

Some people plan their life out and look back at the wealth they've had.
Others live life day by day and look back at the wealth of experiences and enjoyment they've had.

Evidently some versions look just loosely at the data and if it "looks" like good xml it will accept it, or it may have an adaptive system to "guestimate" what the file is trying to say while other parsers look at the "strict" code and if there's anything at all wrong it won't accept it. Much like some web browsers will go ahead and show a page with "bad html" code and others will choke and not display it.
Glad you got it up and running.

---

Jim

Some people plan their life out and look back at the wealth they've had.
Others live life day by day and look back at the wealth of experiences and enjoyment they've had.

I changed all the mention of executable as well, As all had entries at the wrong end, I found this to be true in Both DeNitro's and the Chicken's files had It in error, So people I'd check this and change all the app info files to fix this minor problem, As after I changed the files Benchmark no longer crashed under 5.7.64 as the 64bit version seems to be sensitive to this, Where as the 32bit Boinc 5.4.11 and earlier didn't care I don't think. I also fixed the main program entry too, Now No problems.

<app_info>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<executable>
<name>setiathome-5.15-DeNitro-em64t.exe</name>
</executable>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>515</version_num>
<file_ref>
<main_program>
<file_name>setiathome-5.15-DeNitro-em64t.exe</file_name>
</main_program>
</file_ref>
</app_version>
</app_info>

<app_info>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<executable>
<name>setiathome-5.15-kwsn-sse2.exe</name>
</executable>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>515</version_num>
<file_ref>
<main_program>
<file_name>setiathome-5.15-kwsn-sse2.exe</file_name>
</main_program>
</file_ref>
</app_version>
</app_info>

Batman. Are you running boinc out of the program file folder or program file x64 folder.

I used your app_info and can't get it to work.

I'm using Denitros test 2 64bit with the crunch3rs 64 boinc. Same as you. :-(

---

Whether you are attached to BOINC or not, you do have an error in your app_info.xml file.
I believe the line I have highlighted above tells what is wrong.
You have a bad setof xml tags. The section should read:
<main_program>
some lines of code
</main_program>

The closing tag should be </main_program> not <main_program/>

Hmmm, you sure about that? It looks more like an empty element tag to me. If you moved the slash to before the argument to make it standard closing tag syntax, then there wouldn't be a corresponding opening tag.

FWIW, I ran Simon's default app_info file through the validator here:

Syntax Validator

and it checked out OK.

Alinator

Whether you are attached to BOINC or not, you do have an error in your app_info.xml file.
I believe the line I have highlighted above tells what is wrong.
You have a bad setof xml tags. The section should read:
<main_program>
some lines of code
</main_program>

The closing tag should be </main_program> not <main_program/>

Hmmm, you sure about that? It looks more like an empty element tag to me. If you moved the slash to before the argument to make it standard closing tag syntax, then there wouldn't be a corresponding opening tag.

FWIW, I ran Simon's default app_info file through the validator here:

Syntax Validator

and it checked out OK.

Alinator

It must not be too good as It uses IE and both examples that I posted here tested out as Good on that website.

Besides Boinc did give Me an unparsed error a day or so back, But none since once I made these up.

LOL, roger that! MS isn't known for their strict adherance to any standard that was wasn't of their own origin (sometimes even their own). :-)

However, in this case I'm just trying to understand what's going on here since the <main_program/> is valid xml syntax for an empty element. I haven't had any trouble with it on the 32 bit CC's, so I'm wondering if the authors of the 64 bit clients used a different parser in them.

Alinator

Whether you are attached to BOINC or not, you do have an error in your app_info.xml file.
I believe the line I have highlighted above tells what is wrong.
You have a bad setof xml tags. The section should read:
<main_program>
some lines of code
</main_program>

The closing tag should be </main_program> not <main_program/>

Hmmm, you sure about that? It looks more like an empty element tag to me. If you moved the slash to before the argument to make it standard closing tag syntax, then there wouldn't be a corresponding opening tag.

FWIW, I ran Simon's default app_info file through the validator here:

Syntax Validator

and it checked out OK.

Alinator

It must not be too good as It uses IE and both examples that I posted here tested out as Good on that website.

Besides Boinc did give Me an unparsed error a day or so back, But none since once I made these up.

LOL, roger that! MS isn't known for their strict adherance to any standard that was wasn't of their own origin (sometimes even their own). :-)

However, in this case I'm just trying to understand what's going on here since the <main_program/> is valid xml syntax for an empty element. I haven't had any trouble with it on the 32 bit CC's, so I'm wondering if the authors of the 64 bit clients used a different parser in them.

Alinator

Well, I'm not any expert in xml, but I do know that <tag>data</tag> is a proper xml syntax. If you will look in the files in question you will see the opening tag a few lines above the closing tag, but it's there. It may be that the syntax with a trailing slash would work with some parsers and not others, just as IE and Firefox read html tags differently. However if your assumption that it's an "empty element" tag, then where's the data that is supposed to be there? Empty? If so that is why it gave the error, because it was looking for data and not finding any. It is possible one parser saw the opening tag and just assumed the next was a closing tag although malformed, and the other would see the opening tag then see an empty element tag and really get confused! This use of an "empty tag" is new to me. Every xml file I've ever seen has the <tag>data<</tag> syntax or for an empty element, <tag></tag>.

---

Jim

Some people plan their life out and look back at the wealth they've had.
Others live life day by day and look back at the wealth of experiences and enjoyment they've had.

Well, I'm not any expert in xml, but I do know that <tag>data</tag> is a proper xml syntax. If you will look in the files in question you will see the opening tag a few lines above the closing tag, but it's there. It may be that the syntax with a trailing slash would work with some parsers and not others, just as IE and Firefox read html tags differently. However if your assumption that it's an "empty element" tag, then where's the data that is supposed to be there? Empty? If so that is why it gave the error, because it was looking for data and not finding any. It is possible one parser saw the opening tag and just assumed the next was a closing tag although malformed, and the other would see the opening tag then see an empty element tag and really get confused! This use of an "empty tag" is new to me. Every xml file I've ever seen has the <tag>data<</tag> syntax or for an empty element, <tag></tag>.

Here's a link to the definition of an empty element:

http://www.w3.org/TR/REC-xml/#sec-starttags

Perhaps the ending note about the empty element tag is the key here? Still, it seems strange it works fine with the 32 bit clients, but now is causing an issue with the 64 bit ones.

As you say though, nothing wrong with using the explicit syntax and in this case seems to be the correct workaround for now.

Alinator

...
Still, it seems strange it works fine with the 32 bit clients, but now is causing an issue with the 64 bit ones.

Alinator

BOINC is changing to a new XML parser, the 64 bit client was built from 5.7.0 sources...

                                                        Joe

...
Still, it seems strange it works fine with the 32 bit clients, but now is causing an issue with the 64 bit ones.

Alinator

BOINC is changing to a new XML parser, the 64 bit client was built from 5.7.0 sources...

                                                        Joe

Ahh, well that explains it. ;-)

Guess we have to call it a bug now, since it isn't parsing the empty tag correctly. :-)

Alinator

...
Still, it seems strange it works fine with the 32 bit clients, but now is causing an issue with the 64 bit ones.

Alinator

BOINC is changing to a new XML parser, the 64 bit client was built from 5.7.0 sources...

                                                        Joe

Ahh, well that explains it. ;-)

Guess we have to call it a bug now, since it isn't parsing the empty tag correctly. :-)

Alinator

Ok. Thanks for the link. Very informative, but it's not a bug. If the contents of the app_info.xml file tags (I think it was <main_application>information about main application</main_application>) were used as the "empty" tag, then you couldn't pass the data it is needing. It is rather a bug in the use of the "empty" <tag/> form to end a set of tags that contain data. As I see it, the 32 bit parser doesn't care about the strict form where the parser in the 64 bit program does.

---

Jim

Some people plan their life out and look back at the wealth they've had.
Others live life day by day and look back at the wealth of experiences and enjoyment they've had.

...
Still, it seems strange it works fine with the 32 bit clients, but now is causing an issue with the 64 bit ones.

Alinator

BOINC is changing to a new XML parser, the 64 bit client was built from 5.7.0 sources...

                                                        Joe

Ahh, well that explains it. ;-)

Guess we have to call it a bug now, since it isn't parsing the empty tag correctly. :-)

Alinator

Ok. Thanks for the link. Very informative, but it's not a bug. If the contents of the app_info.xml file tags (I think it was <main_application>information about main application</main_application>) were used as the "empty" tag, then you couldn't pass the data it is needing. It is rather a bug in the use of the "empty" <tag/> form to end a set of tags that contain data. As I see it, the 32 bit parser doesn't care about the strict form where the parser in the 64 bit program does.

Well, I'm not sure exactly what the deal is here.

The whole point of the empty element tag is there is no content involved with it, so I was assuming the tag name itself was the parameter which was getting passed from the parser to the CC.

Did you try just using the verbose syntax on the same line (rather than nesting the exe file name with it)? That would indicate whether it was a parser bug, or inappropriate usage of the the empty tag as you suggest.

In looking at the default example on the BOINC main site, these two tags have always been empty tags.

http://boinc.berkeley.edu/anonymous_platform.php

Alinator

Well, I'm not sure exactly what the deal is here.

The whole point of the empty element tag is there is no content involved with it, so I was assuming the tag name itself was the parameter which was getting passed from the parser to the CC.

Did you try just using the verbose syntax on the same line (rather than nesting the exe file name with it)? That would indicate whether it was a parser bug, or inappropriate usage of the the empty tag as you suggest.

In looking at the default example on the BOINC main site, these two tags have always been empty tags.

http://boinc.berkeley.edu/anonymous_platform.php

Alinator

Ok, You win! I was mistaken by thinking that there was an opening tag above it. After rereading the post in question (with my glasses on this time! Haha!) I realize I was wrong and it "is" an empty element. And it appears that the use as per the page you refered to is correct. So there I agree with you in that there's something fishy. I see Batman got his working which is the main point of this thread, however it has now brought up an issue with the parser used. Could this tag be something that has been added to the basic specification? Or is a valid tag in one specification but not in another? (assuming there are different versions of the specifications.)

---

Jim

Some people plan their life out and look back at the wealth they've had.
Others live life day by day and look back at the wealth of experiences and enjoyment they've had.

NO matter what, hopefully someone will notice this and look into It, But It is a workaround in case someone else needs It.

Well, I guess this is an instance of helping someone by mistake! haha! Just glad my "mistake" led to a fix and got you crunching again.

---

Jim

Some people plan their life out and look back at the wealth they've had.
Others live life day by day and look back at the wealth of experiences and enjoyment they've had.

LOL, well I don't know if it's a matter of winning or not, I'm just trying to understand what's going on. ;-)

I read through a bunch of the XML documentation and my take was the empty element tag is something that's been there all along, so with the switch in parsers my vote is it's a bug. That's why I asked about the alternative structure.

Alinator

The error and the otter........

---

"Time is simply the mechanism that keeps everything from happening all at once."