OFFICIAL "SERVER PROBLEMS" THREAD

[snme2pm1]

Originally Posted by Madpoo

First: The colors of the left-hand menus. "Garish" was the word that came to my mind, but maybe folks like the high contrast.

Yeah, pretty much my thoughts when that happened some months ago.
A less ghastly rainbow effect would not have raised my eyebrow.

[retina]

Originally Posted by Madpoo

http://www.mersenne.org/default.mock.php

Attached is what I see. That big grey space is consuming the screen.

And, OMG, the side menus actually work without JS. Good job.
You see, we don't need JS to make a useful website

[kladner]

I'm curious about one layout detail. (See attached.) Perhaps this is browser-related? (I habitually run Firefox, currently V. 28.0.)

[tha]

Originally Posted by kladner

I'm curious about one layout detail. (See attached.) Perhaps this is browser-related? (I habitually run Firefox, currently V. 28.0.)

I run v32 and it sure looks different.

[kladner]

Originally Posted by tha

I run v32 and it sure looks different.

Upgrading resolved that issue. It also reminded me WHY I was still at 28.0. Higher version include "helpful" and totally unwanted layout changes. This sort of gratuitous change with an update, unannounced and unchangeable, is one of my biggest complaints about Firefox.

EDIT: Rolled back to 28.0. Still working to restore my desired layout.

[brilong]

Originally Posted by brilong

I've also experienced multiple problems over the last few weeks. At first I thought it was the submit_spider Perl script I'm using. Occasionally it gives output like this (missing the number before GHz):

20140530_144705 INFO:  M69277711 submitted; 13.8069 GHz Days credit.
Use of uninitialized value $GHzDays in concatenation (.) or string at /home/horde/GIMPS/gpu0/submit_spider line 154, <IN> line 12.
20140530_144728 INFO:  M69277711 submitted;  GHz Days credit.

It turns out the server is causing this issue (notice the 23 second gap). I've been able to resubmit the same entry again and it usually works fine, but it's a manual process.

I've also had issues submitting new factors. Even when I try to submit them on the Manual webpage, I get an error:

CGI Timeout

The specified CGI application exceeded the allowed time for processing. The server has deleted the process.

It appears submit_spider needs to be improved to catch these errors, sleep a while and try resubmitting. It does NOT label the results file "not_submitted" with the bad entry in this particular case.

Hi guys, I'm experiencing this problem again even with the new Server setup. I use GPU72's submit_spider to submit results. It does not properly catch the fact that the server is assigning zero credit.

The two exponents I was working on are M68647807 and M68648053 both from 2^72 to 2^73. When I submit them, I get the following from submit_spider:

20140919_133623 INFO: M68648053 submitted; GHz Days credit.

There should be a number before GHz Days. Last time I looked into this, it was a server-side timeout that submit_spider was not catching.

Any ideas how I can further diagnose this problem? I was able to submit the exact same results into the manual submission form and it worked instantaneously:

Found 5 lines to process.
processing: TF factor 8823644269361374859767 for M68647807 (272-273)
CPU credit is 12.5662 GHz-days.
processing: TF factor 8102569144038786442433 for M68648053 (272-273)
CPU credit is 10.8524 GHz-days.
Done processing:
* Parsed 4 lines.
* Found 0 datestamps.
GHz-days
Qty	Work	Submitted	Accepted	Average
2	Trial Factoring: factor	14.781	23.419	7.391
2	- all -	14.781		

Found factors lengths:
Bits	Count
73 	2
Digits	Count
22 	2
K size	Count
1013	10,000,000,000,000	2

Status codes:
Code	Meaning	Count
0	no error	2

[Mark Rose]

Originally Posted by Madpoo

So, I've been monkeying around with the site layout in my spare time.

My only complaint would be the huge fonts (went from 11.9 pt to 16.4 pt). If the font size is to stay, I'd switch to serif font like Georgia, which is easier to read at large sizes. Otherwise I'm indifferent.

The new menus are easier to use on my phone, but it's not obvious they're a menu. As to how to best achieve that, I don't know. I'm not a UX guy :)

[Uncwilly]

Originally Posted by Madpoo

So, I've been monkeying around with the site layout in my spare time.

We thought it might be informative to get some feedback on a few design changes, see what sticks.

So far it's just the home page layout. Everything is functional but once you navigate away from this special home page, you'll see the old layout.

http://www.mersenne.org/default.mock.php

Looks like it all works in Opera.


Something for George. Somewhere on the front page or on the "getting Started" page, there should be discussion on GPUs participation.

[Madpoo]

Originally Posted by tha

I run v32 and it sure looks different.

Wow, that's some old Firefox. :)

I haven't looked at older versions of FF myself but I'm *guessing* the reason is that older versions of Firefox (and Chrome) had their own custom ways of handling "box-sizing" in CSS. FF/Chrome each had their own custom thing. FF needs "-moz-box-sizing" and Chrome used, I think, "-webkit-box-sizing".

The reason for adding that to the default styling is so that when adding padding/margins to a div section, it adds those to the *inside* of the div, not to the outside, which makes it SO much easier to lay out your sections without worrying about how many pixels are popping out from padding, margins, borders, etc.

My guess is that older browsers that don't handle just plain "box-sizing: border-box" for that setting are getting the borders added outside the divs and then all of a sudden my side-by-side sections don't fit next to each other in the constrained width. I ran into that a lot when trying to shoehorn it in without using border-box.

I (mistakenly) assumed nobody in their right mind would still be running old FF/Chrome so I didn't bother including those alternate styles... guess I need to add those back in for backwards compatibility.

For all the (boring?) details, see: http://www.paulirish.com/2012/box-si...order-box-ftw/

And sure enough, FF 28 was the last version that still didn't have support for the almost-standard "box-sizing" by itself.

Chrome has supported it since way back at v10, and even IE has supported it since IE 8. Safari since v5.1, etc.

So...I'll add this to the CSS and it should fix it for old FF, <= v28.
-moz-box-sizing: border-box; /* old Firefox, etc */

[Madpoo]

Originally Posted by Mark Rose

My only complaint would be the huge fonts (went from 11.9 pt to 16.4 pt). If the font size is to stay, I'd switch to serif font like Georgia, which is easier to read at large sizes. Otherwise I'm indifferent.

The new menus are easier to use on my phone, but it's not obvious they're a menu. As to how to best achieve that, I don't know. I'm not a UX guy :)

One of my "updates" involved stripping out all the old <font size="x"> tags and replacing with CSS. I don't know if anyone else has ever tried matching old school font sizes to px, pt, em or % font sizes but it seems like somewhat of a dark art.

I settled on font size 3=1em and size 4=1.35em which is actually a touch larger than the 1.13em I'd seen elsewhere.

I also marked up the "articles" on the home page with some proper h2/h3 elements and tried to tweak the font-size on them to kind of match what the old page has. H2 for instance is 1.5em and H3 used as the "title" of each article section is 1.35em and it replaced the old font size=4 they all used to have.

I considered using % instead of em for the size up/down but just didn't bother at this point.

The default font size for the normal text *should* all be 12pt, and a boring "font-family: Tahoma, Arial, Helvetica, sans-serif;"

If you're seeing a larger font size for the regular text, like 16.4, that seems weird. :) If you're able to check using the Chrome inspect element tools, I'd be curious to see where it thinks it's inheriting a larger font size from.

[kladner]

Thanks! Sorry for being such a stick-in-the-mud.