MarmotBlog Comments http://netapps.muohio.edu/marmot/blog Stuff about Marmot! Mon, 23 Nov 2009 06:06:09 +0000 http://wordpress.org/?v=1.5 by: Tim http://netapps.muohio.edu/marmot/blog/archives/2004/03/24/branches-directories-and-a-lesson-in-merging/#comment-20 Wed, 24 Mar 2004 13:35:20 -0500 http://netapps.muohio.edu/marmot/blog/archives/2004/03/24/branches-directories-and-a-lesson-in-merging/#comment-20 I have update -d in my .cvsrc. I would expect that to be enough for the merge to have worked, but I have never actually done it. I have update -d in my .cvsrc. I would expect that to be enough for the merge to have worked, but I have never actually done it.

]]> by: Tim http://netapps.muohio.edu/marmot/blog/archives/2004/03/17/little-stuff/#comment-19 Thu, 18 Mar 2004 09:35:51 -0500 http://netapps.muohio.edu/marmot/blog/archives/2004/03/17/little-stuff/#comment-19 I give up. I thought I had everything working, then the xsltproc on naslnx05 decided it wanted to collapse empty anchor tags, making the entire page light up on :hover. I went back to html output, and that problem went away, but WinIE still does goofy things with the cap at the top of the white box. Maybe I'll try the ALA tricks for a stretchable CSS box layout... Can't be <em>that</em> hard, can it? I give up. I thought I had everything working, then the xsltproc on naslnx05 decided it wanted to collapse empty anchor tags, making the entire page light up on :hover. I went back to html output, and that problem went away, but WinIE still does goofy things with the cap at the top of the white box. Maybe I’ll try the ALA tricks for a stretchable CSS box layout… Can’t be that hard, can it?

]]> by: Tim http://netapps.muohio.edu/marmot/blog/archives/2004/03/03/cvs-quirks/#comment-18 Thu, 4 Mar 2004 09:07:59 -0500 http://netapps.muohio.edu/marmot/blog/archives/2004/03/03/cvs-quirks/#comment-18 Subversion is just CVS with bandaids. Arch rips out that archaic RCS layer and does things <a href="http://wiki.gnuarch.org/moin.cgi/SubVersionAndCvsComparison">better</a> Subversion is just CVS with bandaids. Arch rips out that archaic RCS layer and does things better

]]> by: Clayton http://netapps.muohio.edu/marmot/blog/archives/2004/03/03/cvs-quirks/#comment-17 Thu, 4 Mar 2004 09:03:08 -0500 http://netapps.muohio.edu/marmot/blog/archives/2004/03/03/cvs-quirks/#comment-17 arch? Dirty GNU hippie! <a href="http://subversion.tigris.org/" title="Subversion home page">subversion</a>'s where it's at! arch? Dirty GNU hippie! subversion’s where it’s at!

]]> by: Tim http://netapps.muohio.edu/marmot/blog/archives/2004/02/18/running-post-merge-changelog/#comment-16 Thu, 19 Feb 2004 23:58:43 -0500 http://netapps.muohio.edu/marmot/blog/archives/2004/02/18/running-post-merge-changelog/#comment-16 MU_Logger MU_Simple_Logger MU_Logger is currently an empty class that extends MU_Simple_Logger, a very basic logger class with no persistence between page requests. The deprecation logging has been rewritten to use an MU_Logger object. mu_logline() will write a single line to this global log object mu_stdlib.php _mu_deprecated has been deprecated. (It still works, I just wanted to say that.) We now have _mu_error() which should be used to throw Marmot-level errors. There are currently two types: 'warn', which displays a warning in development stages (just like _mu_deprecate does now), and 'fatal', which logs an error, displays the error, dumps the logger, and exits. mu_autoprepend.php Includes MU_Logger.php and creates a new object in $GLOBALS['_logger'] <b>Update:</b> Since I'm an idiot, I originally made these changes on naslnx07 (Don't ask why, I was playing with KDE apps remotely via X at home), which was on the dev_1-1_branch. I have finally re-applied those changes to the trunk and reverted the changes on the branch. All is believed to be well. MU_Logger
MU_Simple_Logger
MU_Logger is currently an empty class that extends MU_Simple_Logger, a very basic logger class with no persistence between page requests. The deprecation logging has been rewritten to use an MU_Logger object.

mu_logline() will write a single line to this global log object

mu_stdlib.php
_mu_deprecated has been deprecated. (It still works, I just wanted to say that.) We now have _mu_error() which should be used to throw Marmot-level errors. There are currently two types: ‘warn’, which displays a warning in development stages (just like _mu_deprecate does now), and ‘fatal’, which logs an error, displays the error, dumps the logger, and exits.

mu_autoprepend.php
Includes MU_Logger.php and creates a new object in $GLOBALS[’_logger’]

Update:
Since I’m an idiot, I originally made these changes on naslnx07 (Don’t ask why, I was playing with KDE apps remotely via X at home), which was on the dev_1-1_branch. I have finally re-applied those changes to the trunk and reverted the changes on the branch. All is believed to be well.

]]> by: Tim http://netapps.muohio.edu/marmot/blog/archives/2004/02/18/12-feature-requests-or-last-minute-11-additions/#comment-15 Thu, 19 Feb 2004 13:34:28 -0500 http://netapps.muohio.edu/marmot/blog/archives/2004/02/18/12-feature-requests-or-last-minute-11-additions/#comment-15 Modify _mu_deprecated() to always require two parameters and fix anywhere that calls it with one (all my fault). Modify it such that a level of 'warn' will just log, and a level of 'terminate' will log and dump some kind of friendly output that includes anything logged so far. (Just build a new MU_Presentation() and display, while forcing the display of debug messages?) <b>Done, more intelligently</b> _mu_deprecated still takes one or two arguments, and the new _mu_error() takes two arguments, the first of which can be 'warn' or 'fatal'. Modify _mu_deprecated() to always require two parameters and fix anywhere that calls it with one (all my fault). Modify it such that a level of ‘warn’ will just log, and a level of ‘terminate’ will log and dump some kind of friendly output that includes anything logged so far. (Just build a new MU_Presentation() and display, while forcing the display of debug messages?)

Done, more intelligently
_mu_deprecated still takes one or two arguments, and the new _mu_error() takes two arguments, the first of which can be ‘warn’ or ‘fatal’.

]]> by: dirk http://netapps.muohio.edu/marmot/blog/archives/2004/02/18/running-post-merge-changelog/#comment-14 Thu, 19 Feb 2004 10:13:39 -0500 http://netapps.muohio.edu/marmot/blog/archives/2004/02/18/running-post-merge-changelog/#comment-14 MU_Form class Improved handling for file input types on forms. The value of a file input object is now the corresponding array from the $_FILES global. It will be returned from either the value or values method of the form. Improved (actually enabled) requiring a value for input items of type file. PHP errors regarding size and partial uploads are also handled as part of requiring the item. MU_Form class
Improved handling for file input types on forms. The value of a file input object is now the corresponding array from the $_FILES global. It will be returned from either the value or values method of the form.

Improved (actually enabled) requiring a value for input items of type file. PHP errors regarding size and partial uploads are also handled as part of requiring the item.

]]> by: dirk http://netapps.muohio.edu/marmot/blog/archives/2004/02/18/things-to-discuss/#comment-13 Thu, 19 Feb 2004 10:13:03 -0500 http://netapps.muohio.edu/marmot/blog/archives/2004/02/18/things-to-discuss/#comment-13 Actually, I think the "default" method in Marmot ignores any browser issues (or should). Rather than relying on what the browser may happen to pick as a default, Marmot sets the "default" as a hidden form item. Since PHP doesn't cumulatively build values, any clicked button overwrites the hidden element of the same name, leaving us with the value of the button the user clicked. If the user presses enter and the browser doesn't submit a button name, the default value is used. Actually, I think the “default” method in Marmot ignores any browser issues (or should). Rather than relying on what the browser may happen to pick as a default, Marmot sets the “default” as a hidden form item. Since PHP doesn’t cumulatively build values, any clicked button overwrites the hidden element of the same name, leaving us with the value of the button the user clicked.

If the user presses enter and the browser doesn’t submit a button name, the default value is used.

]]> by: Tim http://netapps.muohio.edu/marmot/blog/archives/2004/02/18/12-feature-requests-or-last-minute-11-additions/#comment-12 Wed, 18 Feb 2004 23:55:08 -0500 http://netapps.muohio.edu/marmot/blog/archives/2004/02/18/12-feature-requests-or-last-minute-11-additions/#comment-12 Fix Query Builder to gracefully handle when no output fields are chosen. (Currently calls trigger_error() and bails, leaving subsequent reloads to use the same cached settings and error out.) Fix Query Builder to gracefully handle when no output fields are chosen. (Currently calls trigger_error() and bails, leaving subsequent reloads to use the same cached settings and error out.)

]]> by: dirk http://netapps.muohio.edu/marmot/blog/archives/2004/02/18/12-feature-requests-or-last-minute-11-additions/#comment-11 Wed, 18 Feb 2004 16:26:45 -0500 http://netapps.muohio.edu/marmot/blog/archives/2004/02/18/12-feature-requests-or-last-minute-11-additions/#comment-11 I just used my popup window and discovered a problem. The content_only theme is based on miami_02, so the font color, etc has those characteristics. I was using elegant and it just looked wrong. Should we consider having alternate window styles within each theme rather than separate themes for such things? I just used my popup window and discovered a problem. The content_only theme is based on miami_02, so the font color, etc has those characteristics. I was using elegant and it just looked wrong.

Should we consider having alternate window styles within each theme rather than separate themes for such things?

]]>