On May 25, 2010, at 11:42 AM, Hannes Magnusson wrote: > On Sat, May 22, 2010 at 19:23, Philip Olson <phi...@php.net> wrote: >> philip Sat, 22 May 2010 17:23:13 +0000 >> >> Revision: http://svn.php.net/viewvc?view=revision&revision=299628 >> >> Log: >> Implemented new version.trunk. entities >> >> Changed paths: >> U phpdoc/en/trunk/reference/image/functions/imagecolorset.xml >> U phpdoc/en/trunk/reference/json/functions/json-decode.xml >> >> Modified: phpdoc/en/trunk/reference/image/functions/imagecolorset.xml >> =================================================================== >> --- phpdoc/en/trunk/reference/image/functions/imagecolorset.xml 2010-05-22 >> 17:09:43 UTC (rev 299627) >> +++ phpdoc/en/trunk/reference/image/functions/imagecolorset.xml 2010-05-22 >> 17:23:13 UTC (rev 299628) >> @@ -64,8 +64,11 @@ >> <term><parameter>alpha</parameter></term> >> <listitem> >> <para> >> - Value of alpha component. This parameter is only available in SVN >> (trunk). >> + Value of alpha component. >> </para> >> + <para> >> + &version.trunk.after.53; >> + </para> > > We don't document version requirements inline, its only in the changelog.
Yes, but it felt like the best choice here considering these features don't exist anywhere, and may never exist. Am a little paranoid with this unique situation (and note: eventually these entities will be removed). >> + <refsect1 role="changelog"> >> + &reftitle.changelog; >> + <informaltable> >> + <tgroup cols="2"> >> + <thead> >> + <row> >> + <entry>&Version;</entry> >> + <entry>&Description;</entry> >> + </row> >> + </thead> >> + <tbody> >> + <row> >> + <entry>&version.trunk.changelog;</entry> >> + <entry> >> + The <parameter>alpha</parameter> parameter was added. > > > "Future" doesn't sound that good. > > Imaging the following table: > > Future | $foo was added > 5.2.1 | $foobar was added > > Whattaheck is future? It at least has to be a link to the about version page. This is why the other entity exists, which explains it and links to the about.phpversions page. Adding a long entry as the changelog version feels wrong, but are you are proposing version.trunk.after.53 instead live within the changelog description entry? That could work. I'm still paranoid, but it is indeed closer to how we do changelogs currently. Regards, Philip
