TF fetching/reporting toolkit for Linux
 

I've been toying with the idea of writing a Python base MISFIT tool set, but need to know if there is a need. [B]Haters of automation or Python, or the internet, or Linux need not reply....[/B]

Here is my basic thinking.
A library of four Python console programs consisting of
1. Fetch TF work via GPU72
2. Reporting TF work to GIMPS
3. Fetch TF work from GIMPS
4. MISFIT control

I realize Chalsall has #2 as a Pearl utility but if I build this I'll want to deploy a unified kit.

MISFIT control would be the final module that binds them all together with additional features like emailing, stalled processor detection, etc.

If you no one wants MISFIT control, and prefers to chron everthing on their own that's fine too.

The question is.... Is there a potential user base for this or do this utilities already exist for Linux?

Let me know.

I don't use Linux with powerful gpus, but it seems to me that that would be a welcome addition to the community tool kit. As far as I know there are no good ways to fetch work in Linux without writing something homegrown in perl. Obviously challsall's spider works pretty well for submitting - I use that myself. Improvements probably could be made though, with the server no longer accepting segments out of order.

That's one vote.... We just need about 9 more...

[QUOTE=swl551;371986]That's one vote.... We just need about 9 more...[/QUOTE]

I simply don't have the time to implement this myself. I started quite some time ago, and I have some initial work developed in Perl which I would be happy to share. I ended up wasting about two weeks of work trying to get "fork" to work in Perl under Windows... :bangheadonwall: :no:

I do think it would be very welcomed and used by our Linux workers, and I would be happy to assist from the Server API side of things -- including (if you're willing) an "estimated completion" and "updated" data exchange.

+1 Would definitely be useful personally.

[QUOTE=chalsall;371988]I simply don't have the time to implement this myself. I started quite some time ago, and I have some initial work developed in Perl which I would be happy to share. I ended up wasting about two weeks of work trying to get "fork" to work in Perl under Windows... :bangheadonwall: :no:

I do think it would be very welcomed and used by our Linux workers, and I would be happy to assist from the Server API side of things -- including (if you're willing) an "estimated completion" and "updated" data exchange.[/QUOTE]
Chris I plan to use the standard MISFIT gimps pages George setup for me and to interact with GPU72 the same way MISFIT does. Given those existing pieces the implementation is really a port. The key change is using a language that is well established in Linux that I can easily work with.

In-fact all my development work will be done on a Linux machine with no consideration for Windows (WinBlows - clarifying which OS for Chris) :smile:

However this will be boring for me if there is no user base as the interactions with the community is the reward. (usually).

None-the-less I will probably need advice as I get into this as I've logged exactly 12.5587 hours with Linux. Hey, at least I'm trying right!

[QUOTE=swl551;371991]In-fact all my development work will be done on a Linux machine with no consideration for Windows (WinBlows - clarifying which OS for Chris) :smile:[/QUOTE]

Cool. My head-banging occurred because I was trying to make my "fetching spider" be "cross-platform". This was before you brought MISFIT forward. WinBlows simply doesn't do fork()'ing under Perl well -- the call is there, but you can't get signals back reliably, nor can you kill() a child process which has launched another program.

[QUOTE=swl551;371991]However this will be boring for me if there is no user base as the interactions with the community is the reward. (usually).[/QUOTE]

I can't speak for others, but I, personally, would use it.

[QUOTE=swl551;371991]None-the-less I will probably need advice as I get into this as I've logged exactly 12.5587 hours with Linux. Hey, at least I'm trying right![/QUOTE]

Always happy to help. So you know, my "native tongues" are C and Perl. I can do Python if I have to, but I seriously take offence to a language which tells the programmer how the indentation should be, and refuses to run if the white-space is not the way [B][I][U]it[/U][/I][/B] likes it. That's what brackets are for.

I'm the human. You're the software. Do what I tell you to do (even if I might be wrong).

[QUOTE=chalsall;371995]Cool. My head-banging occurred because I was trying to make my "fetching spider" be "cross-platform". This was before you brought MISFIT forward. WinBlows simply doesn't do fork()'ing under Perl well -- the call is there, but you can't get signals back reliably, nor can you kill() a child process which has launched another program.



I can't speak for others, but I, personally, would use it.



Always happy to help. So you know, my "native tongues" are C and Perl. I can do Python if I have to, but I seriously take offence to a language which tells the programmer how the indentation should be, and refuses to run if the white-space is not the way [B][I][U]it[/U][/I][/B] likes it. That's what brackets are for.

I'm the human. You're the software. Do what I tell you to do (even if I might be wrong).[/QUOTE]
I think this comic sums it up...
[url]http://imgs.xkcd.com/comics/python.png[/url]

I always thought the forced whitespace was python's best feature, drawing a line against lazy coders! The not-lazy coders would be tabbing correctly anyway. :)

EDIT: Besides, in Python the tabs are just invisible and more versatile brackets. :)

[QUOTE=Aramis Wyler;372013]I always thought the forced whitespace was python's best feature, drawing a line against lazy coders! The not-lazy coders would be tabbing correctly anyway. :)

EDIT: Besides, in Python the tabs are just invisible and more versatile brackets. :)[/QUOTE]

I understand your point. But...

In my world, VIM is my "IDE". If I want to remove a bit of code (C, Perl, Javascript, et al) by quickly adding a "if (0) { ... }" section around a block, I can. If I want to add a bit of code I can simply type it in. If a section of code is more readable in one line rather than two or three, I should able to do so (e.g. "if ([some condition]) {[action]}").

Hell, if a programmer wants to write their entire program on a single line, they should be able to do so. This is often done (using tools) with Javascript to lessen the download size of production code.

[QUOTE=chalsall;372019][I]Reason: Corrected whitespace -- didn't compile initially....[/I][/QUOTE]
Bwaaa haaa haaa! This is genial. :rofl: