
<!DOCTYPE rss PUBLIC  "-//Netscape Communications//DTD RSS 0.91//EN"
"http://my.netscape.com/publish/formats/rss-0.91.dtd">
<rss version="0.91">
<channel> 
<title></title> 
<link>http://www-int.stsci.edu/~bsimon/weblog.cgi</link> 
<description>The work log of Bernie Simon</description> 
<language>en</language> 
<item>
<title>Today's Work</title>
<link>http://www-int.stsci.edu/~bsimon/weblog.cgi/post/201040418000.html</link>
<description>
&lt;p&gt;
Karen Keidel sent me  a new version of the back of the book tablees
for the phonebook. So I started working on the phonebook again. I
converted the new tables to LaTeX format, except for the ERC members
list. But when I ran LaTeX on the new file, it barfed. The problem is
that the &lt;tt&gt;\head&lt;/tt&gt; macro expects a table to have four columns. It will 
have to be modified to cope with tables with different numbers of
columns. I also have to compare the office/division phone list in the
new version of the file to the older version of the file.&lt;/p&gt;

&lt;p&gt;The Institute has announced support for Linux and Mac OS X. So unix on 
a PC has now arrived at the Space Telescope.&lt;/p&gt;

&lt;p&gt;I will be on vacation the next two week, expect me back on January
7th.
&lt;/p&gt;
</description>
</item>

<item>
<title>Today's Work</title>
<link>http://www-int.stsci.edu/~bsimon/weblog.cgi/post/191040331600.html</link>
<description>
&lt;p&gt;
I spent all day trying to puzzle out the problem with the tree widget
on the org browser in STLDAPManager. It seems to collapse the tree at
random times. I spent equal amounts of time clicking on the interface
to try to figure out what its problem is and staring at the code. I
came to the conclusion the best way to solve the problem is to hide
it, by setting single=0 in the tree widget. I'm testing out this fix
to see if it's satisfactory.
&lt;/p&gt;
</description>
</item>

<item>
<title>Today's Work</title>
<link>http://www-int.stsci.edu/~bsimon/weblog.cgi/post/181040245200.html</link>
<description>
&lt;p&gt;
Today I solved Gary Truslow's problem with internalUser. Once the fix
is moved to test, he should be able to treat machines on the stsci.org 
subnet as external to the institute.
&lt;/p&gt;
</description>
</item>

<item>
<title>Today's Work</title>
<link>http://www-int.stsci.edu/~bsimon/weblog.cgi/post/171040158800.html</link>
<description>
&lt;p&gt;
I gave a report on what I've been doing at the LDAP status meeting and
I updated the Wiki with the same information. Jim reported a problem
with changePassword, but I found it to be a non-issue. Steve said the
printable page fix worked inconsistently. The problem was caused by
the RAM cache and adding printable to the list of REQUEST variables
seems to fix the problem. I also did some work on a replacement for
the templating code in pagecheck.
&lt;/p&gt;
</description>
</item>

<item>
<title>Today's Work</title>
<link>http://www-int.stsci.edu/~bsimon/weblog.cgi/post/161040072400.html</link>
<description>
&lt;p&gt;
I spent the day testing the changes I made to the LDAP directory
product on Friday. The only problem I uncovered was that the configure 
form was not properly reading the list of LDAP servers. I fixed that
problem. While testing I discovered that you can hose the LDAP manager 
if you feed it a bogus list of LDAP servers. Since it needs to connect 
to LDAP before it can be reconfigured, having an invalid list is an
unrecoverable error. I put a simple check in the code that sets the
servers so that it rejects server names that are obviously incorrect
(no dot in the name) and throws an error if all are rejected. This
isn't wholly satisfactory, but will do for now. I finished up by doing 
a little cosmetic editing on the add and configure forms. Harry tells
me that there are a set of unit tests. Tomorrow I will ask his help
in configuring them and run them.
&lt;/p&gt;
</description>
</item>

<item>
<title>Today's Work</title>
<link>http://www-int.stsci.edu/~bsimon/weblog.cgi/post/131039813200.html</link>
<description>
&lt;p&gt;
The LDAP server started crashing every several minutes. This obviously 
caused a number of complaints from users. Zope seemed to handle the
problem particularly badly and when I looked, I found the
reason. There was no code in our LDAP manager to fail over to the
backup server if the primary server went down. So I spent the day
modifying the STLDAPManager product so that it uses a backup
server. I wrote the code and it passes the preliminary test (it
doesn't crash when you do obvious things.) The more thorough test will 
have to wait until Monday.
&lt;/p&gt;
</description>
</item>

<item>
<title>Today's Work</title>
<link>http://www-int.stsci.edu/~bsimon/weblog.cgi/post/121039726800.html</link>
<description>
&lt;p&gt;
Jim found a rather obscure bug in changePassword. It crashes when
called from the root folder and it is passed a bad user name. Fixing
the bug took all day, mostly because I was editing the wrong file and
wondering why my fix wasn't working.
&lt;/p&gt;
</description>
</item>

<item>
<title>Today's Work</title>
<link>http://www-int.stsci.edu/~bsimon/weblog.cgi/post/111039640400.html</link>
<description>
&lt;p&gt;
I revised the script that generates the list of discrepancies to
include empty fields, ran the script, and sent the results off to
Steve Dignan. Then Jim Grice had a problem with CVS and I helped him
straighten it out. After that I started working on a replacement for
sitecheck based on the current pagecheck code.
&lt;/p&gt;
</description>
</item>

<item>
<title>Today's Work</title>
<link>http://www-int.stsci.edu/~bsimon/weblog.cgi/post/101039554000.html</link>
<description>
&lt;p&gt;
I finished up coding and debugging the printable pagelink. Each page
now has a link on the bottom which generates a printable version of
that page. I aslo generated a list of discrepancies between the
leave table and ldap.
&lt;/p&gt;
</description>
</item>

<item>
<title>Today's Work</title>
<link>http://www-int.stsci.edu/~bsimon/weblog.cgi/post/091039467600.html</link>
<description>
&lt;p&gt;
I switched from working on LDAP to working some Zope calls. This
morning I revised the change password page so that it requests
the user name. I also modified the documentation. (How about that!)
After that was squared away, I modified the standard header and footer 
to support a printable page option. If printable=1 is added to the
url, the header and navigation on a page is excluded from it. I'm
still working on the last part of the problem, some python code to
generate the url for a printable version of the current page
(/stsci/printable_page).
&lt;/p&gt;
</description>
</item>

</channel>
</rss>
