1) Message boards : Number crunching : The new look bugs (Message 18205)
Posted 16 Oct 2007 by Mikus
OK herein lies the rub I have now made width = auto which is, of course, what I've always wanted to do but if you look here:
You will see why I haven't(only an issue in FF as far as I can see). If there is an element contained within a table which has no breaks then FF can't do word wrap and so you get that ugly POS.
I will however leave width at auto.

curiouser and curiouser

(1) With the change to 'auto', all the webpages regarding my account (including "Results for user") look good in my FF. [No more overlapping or truncated text.]

(2) The Switzerland page you referenced -- pages 2, 3, and 4 of that set always look good in my FF. It is only page 1 of that set that looks ugly. [And the very first time I looked at that page 1, it *also* looked good. But when I returned to page 1, it now looks ugly (no matter what I try).]

Thanks for making this change to auto.
2) Message boards : Number crunching : Please increase our daily machine quota (Message 18159)
Posted 14 Oct 2007 by Mikus
I can live with 2/day/CPU. But that is not what I am being sent.

I run off-line. Meaning that I when I connect, I should fetch enough work to keep my system busy until the next connect_occasion. When I connected today, LHC sent me one workunit, and told me I would have to wait one hour before asking for more LHC work. Well, in one hour I will *not* be connected. [My next connection might not be until tomorrow.]

LHC appears to have a fetch_backoff policy as well as a work_limit policy. I have a multi-core computer. Was it the intent of the LHC project that the effect of me doing only an occasional connect resulted in just one LHC workunit being sent to my computer today ?
3) Message boards : Number crunching : The new look bugs (Message 18157)
Posted 14 Oct 2007 by Mikus

the width thing -

I have weak eyes, so I have changed __all__ my desktop font specifications to a larger font size than the default. I'm using FF, and on my system's 'Results for user' page the columns being shown are too narrow for their contents -- the column headings have words that extend into the next column, while some of the column data texts (e.g., the time) are truncated on the right.

If I stretch my browser window to the maximum that my screen allows, the column widths on the 'Results for user' page stay "fixed" (i.e., they do not increase).
4) Message boards : Number crunching : 64-bit Linux (Message 16966)
Posted 4 Jun 2007 by Mikus
I've been using the 64-bit Linux client from boinc.berkeley.edu. Recently, having read that the master URL for LHC had changed, I detached from LHC and then re-attached.

Since then (if I don't intervene) the LHC server does not download the LHC executable for me -- instead, the server says: platform 'x86_64-pc-linux-gnu' not found.

The current server software provides an <alt_platform> capability. When the project does not support the primary platform requested by the client, but an alternate has been indicated, the server would supply executables and work according to that "alt_platform". As long as LHC does not provide a "native" 64-bit Linux executable, could you please set up your server so that it can supply 'i686-pc-linux-gnu' stuff when 'x86_64-pc-linux-gnu' is requested ?


p.s. Before I did the detach of LHC, I was running with an app_info.xml file for LHC. It specified the 32-bit LHC executable (and caused the requested platform to be 'anonymous'). After the re-atttach, the first time I put in that file the executable did not get downloaded ("URL missing"). Have not gotten around to trying it again. I would prefer that the platform compatibility matter be resolved at the LHC project server, instead of me at my client having to manually set up an app_info.xml file for the LHC project.

©2023 CERN