Vim window logic, slimv

This commit is contained in:
Ian Keane 2020-02-24 20:27:04 -05:00
parent babcc9e44b
commit 515847d07e
791 changed files with 51552 additions and 86 deletions

View file

@ -0,0 +1,700 @@
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<title>CL-FAD - A portable pathname library for Common Lisp</title>
<style type="text/css">
pre { padding:5px; background-color:#e0e0e0 }
h3, h4, h5 { text-decoration: underline; }
a { text-decoration: none; padding: 1px 2px 1px 2px; }
a:visited { text-decoration: none; padding: 1px 2px 1px 2px; }
a:hover { text-decoration: none; padding: 1px 1px 1px 1px; border: 1px solid #000000; }
a:focus { text-decoration: none; padding: 1px 2px 1px 2px; border: none; }
a.none { text-decoration: none; padding: 0; }
a.none:visited { text-decoration: none; padding: 0; }
a.none:hover { text-decoration: none; border: none; padding: 0; }
a.none:focus { text-decoration: none; border: none; padding: 0; }
a.noborder { text-decoration: none; padding: 0; }
a.noborder:visited { text-decoration: none; padding: 0; }
a.noborder:hover { text-decoration: none; border: none; padding: 0; }
a.noborder:focus { text-decoration: none; border: none; padding: 0; }
pre.none { padding:5px; background-color:#ffffff }
</style>
</head>
<body bgcolor=white>
<h2>CL-FAD - A portable pathname library for Common Lisp</h2>
<blockquote>
<br>&nbsp;<br><h3><a name=abstract class=none>Abstract</a></h3>
CL-FAD (for "<font color=red>F</font>iles <font color=red>a</font>nd
<font color=red>D</font>irectories") is a thin layer atop Common
Lisp's standard pathname functions. It is intended to provide some
unification between current CL implementations on Windows, OS X,
Linux, and Unix. Most of the code was written by Peter Seibel for his book <a href="http://www.gigamonkeys.com/book/"><em>Practical Common Lisp</em></a>.
<p>
CL-FAD comes with a <a
href="http://www.opensource.org/licenses/bsd-license.php">BSD-style
license</a> so you can basically do with it whatever you want.
<p>
<font color=red>Download shortcut:</font> <a href="http://weitz.de/files/cl-fad.tar.gz">http://weitz.de/files/cl-fad.tar.gz</a>.
</blockquote>
<br>&nbsp;<br><h3><a class=none name="contents">Contents</a></h3>
<ol>
<li><a href="#download">Download and installation</a>
<li><a href="#implementations">Supported Lisp implementations</a>
<li><a href="#dictionary">The CL-FAD dictionary</a>
<ol>
<li><a href="#querying">Querying files, directories and pathnames</a>
<ol>
<li><a href="#directory-exists-p"><code>directory-exists-p</code> [function]</a>
<li><a href="#directory-pathname-p"><code>directory-pathname-p</code> [function]</a>
<li><a href="#file-exists-p"><code>file-exists-p</code> [function]</a>
<li><a href="#pathname-absolute-p"><code>pathname-absolute-p</code> [function]</a>
<li><a href="#pathname-equal"><code>pathname-equal</code> [function]</a>
<li><a href="#pathname-relative-p"><code>pathname-relative-p</code> [function]</a>
<li><a href="#pathname-root-p"><code>pathname-root-p</code> [function]</a>
</ol> </li>
<li><a href="#manipulating">Manipulating pathnames</a>
<ol>
<li><a href="#canonical-pathname"><code>canonical-pathname</code> [function]</a>
<li><a href="#merge-pathnames-as-directory"><code>merge-pathnames-as-directory</code> [function]</a>
<li><a href="#merge-pathnames-as-file"><code>merge-pathnames-as-file</code> [function]</a>
<li><a href="#pathname-as-directory"><code>pathname-as-directory</code> [function]</a>
<li><a href="#pathname-as-file"><code>pathname-as-file</code> [function]</a>
<li><a href="#pathname-directory-pathname"><code>pathname-directory-pathname</code> [function]</a>
<li><a href="#pathname-parent-directory"><code>pathname-parent-directory</code> [function]</a>
</ol> </li>
<li><a href="#traversing">Traversing directories</a>
<ol>
<li><a href="#list-directory"><code>list-directory</code> [function]</a>
<li><a href="#walk-directory"><code>walk-directory</code> [function]</a>
</ol> </li>
<li><a href="#temporary-files">Temporary Files</a>
<ol>
<li><a href="#open-temporary"><code>open-temporary</code> [function]</a>
<li><a href="#with-output-to-temporary-file"><code>with-output-to-temporary-file</code> [macro]</a>
<li><a href="#with-open-temporary-file"><code>with-open-temporary-file</code> [macro]</a>
<li><a href="#star-default-template-star"><code>*default-template*</code> [variable]</a>
<li><a href="#cannot-create-temporary-file"><code>cannot-create-temporary-file</code> [condition]</a>
<li><a href="#invalid-temporary-pathname-template"><code>invalid-temporary-pathname-template</code> [condition]</a>
<li><a href="#missing-temp-environment-variable"><code>missing-temp-environment-variable</code> [condition]</a>
<li><a href="#lp-host-temporary-files"><code>temporary-files</code> [logical pathname host]</a>
</ol>
</li>
<li><a href="#modifying">Modifying the file system</a>
<ol>
<li><a href="#copy-file"><code>copy-file</code> [function]</a>
<li><a href="#copy-stream"><code>copy-stream</code> [function]</a>
<li><a href="#delete-directory-and-files"><code>delete-directory-and-files</code> [function]</a>
</ol> </li>
<li><a href="#package-path"><code>path</code> [package]</a>
</ol>
<li><a href="#ack">Acknowledgements</a>
</ol>
<br>&nbsp;<br><h3><a class=none name="download">Download and installation</a></h3>
CL-FAD together with this documentation can be downloaded from <a
href="https://github.com/edicl/cl-fad/releases/latest">https://github.com/edicl/cl-fad/releases/latest</a>. The
current version is 0.7.6.
<p>
CL-FAD comes with simple system definitions for <a
href="http://www.cliki.net/mk-defsystem">MK:DEFSYSTEM</a> and <a
href="http://www.cliki.net/asdf">asdf</a> so you can either adapt it
to your needs or just unpack the archive and from within the CL-FAD
directory start your Lisp image and evaluate the form
<code>(mk:compile-system&nbsp;&quot;cl-fad&quot;)</code> - or <code>(asdf:oos&nbsp;'asdf:load-op&nbsp;:cl-fad)</code> for asdf - which should compile and load the whole
system.
Installation via <a
href="http://www.cliki.net/asdf-install">asdf-install</a> should as well
be possible. Plus, there are ports
for <a href="http://www.gentoo.org/proj/en/common-lisp/index.xml">Gentoo Linux</a> thanks to Matthew Kennedy
and for <a href="http://packages.debian.org/cgi-bin/search_packages.pl?keywords=cl-fad&amp;searchon=names&amp;subword=1&amp;version=all&amp;release=all">Debian Linux</a> thanks to Ren&eacute; van Bevern.
<p>
If for some reason you can't or don't want to use MK:DEFSYSTEM or asdf you
can just <code>LOAD</code> the file <code>load.lisp</code>.
<p>
The latest version of the source code lives in the github repository <a href="https://github.com/edicl/cl-fad">edicl/cl-fad</a>.
If you want to send patches,
please <a href="http://weitz.de/patches.html">read this first</a>.
Please submit your changes
as <a href="https://github.com/edicl/cl-fad/pulls">GitHub pull
request"</a>.
<br>&nbsp;<br><h3><a class=none name="implementations">Supported Lisp implementations</a></h3>
<p>
The following Common Lisp implementations are currently supported:
<ul>
<li><a href="http://armedbear.org/abcl.html">Armed Bear Common Lisp</a>
<li><a href="http://www.cons.org/cmucl/">CMUCL</a>
<li><a href="http://www.cormanlisp.com/">Corman Common Lisp</a>
<li><a href="http://ecls.sf.net/">ECL</a>
<li><a href="http://www.franz.com/products/allegrocl/">Franz AllegroCL</a>
<li><a href="http://clisp.cons.org/">GNU CLISP</a>
<li><a href="http://www.lispworks.com/">LispWorks</a>
<li><a href="http://www.digitool.com/">Macintosh Common Lisp</a>
<li><a href="http://openmcl.clozure.com/">OpenMCL</a>
<li><a href="http://www.scieneer.com/scl/">Scieneer Common Lisp</a>
<li><a href="http://sbcl.sourceforge.net/">Steel Bank Common Lisp</a>
</ul>
I'll gladly accepts patches to make CL-FAD work on other platforms.
<br>&nbsp;<br><h3><a class=none name="dictionary">The CL-FAD dictionary</a></h3>
<h4><a class=none name="querying">Querying files, directories and pathnames</a></h4>
<p><br>[Function]
<br><a class=none name="directory-exists-p"><b>directory-exists-p</b> <i> pathspec </i> =&gt; <i> generalized-boolean</i></a>
<blockquote><br>
Checks whether the file named by the <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_p.htm#pathname_designator">pathname designator</a> <code><i>pathspec</i></code>
exists and if it is a directory. Returns its <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_t.htm#truename">truename</a> if this is the
case, <code>NIL</code> otherwise. The truename is returned in <em>directory form</em> as if
by <a href="#pathname-as-directory"><code>PATHNAME-AS-DIRECTORY</code></a>.
</blockquote>
<p><br>[Function]
<br><a class=none name="directory-pathname-p"><b>directory-pathname-p</b> <i> pathspec </i> =&gt; <i> generalized-boolean</i></a>
<blockquote><br>
Returns <code>NIL</code> if <code><i>pathspec</i></code> (a <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_p.htm#pathname_designator">pathname designator</a>) does not designate
a directory, <code><i>pathspec</i></code> otherwise. It is irrelevant whether the file or
directory designated by <code><i>pathspec</i></code> does actually exist.
</blockquote>
<p><br>[Function]
<br><a class=none name="file-exists-p"><b>file-exists-p</b> <i> pathspec </i> =&gt; <i> generalized-boolean</i></a>
<blockquote><br>
Checks whether the file named by the <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_p.htm#pathname_designator">pathname designator</a> <code><i>pathspec</i></code>
exists and returns its <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_t.htm#truename">truename</a> if this is the case, <code>NIL</code> otherwise.
The truename is returned in "canonical" form, i.e. the truename of a
directory is returned in <em>directory form</em> as if by <a href="#pathname-as-directory"><code>PATHNAME-AS-DIRECTORY</code></a>.
</blockquote>
<p><br>[Function]<br><a class=none name='pathname-absolute-p'><b>pathname-absolute-p</b> <i>a</i> =&gt; <i>result</i></a>
<blockquote>
<p>Returns true if <code><i>a</i></code> is an absolute pathname. This simply
tests if <code><i>a</i></code>&#039;s directory list starts with <code>:ABSOLUTE</code></p>
</blockquote>
<p><br>[Function]<br><a class=none name='pathname-equal'><b>pathname-equal</b> <i>a b</i> =&gt; <i>result</i></a>
<blockquote>
<p>Returns <em>true</em> if <code><i>a</i></code> and <code><i>b</i></code>
represent the same pathname. This function does not access the
filesystem, it only looks at the components of the two pathnames to
test if they are the same (though by passing both <code><i>a</i></code>
and <code><i>b</i></code> to probe-file one can make this function test for
file &#039;sameness&#039;.</p>
<p>Equality is defined as:</p>
<ul>
<li>strings that are <code>string=</code>
<li>symbols (including <code>nil</code> and keywords) which are <code>eql</code>
<li>lists of the same length with equal (as per these rules) elements.
</ul>
<p>If any of these tree conditions is false for any of the components in
<code><i>a</i></code> and <code><i>b</i></code> then <code><i>a</i></code>
and <code><i>b</i></code> are different, otherwise they are the same.</p>
<p><em>NB:</em> This function does not convert name strings to pathnames. So
&quot;foo.txt&quot; and #P&quot;foo.txt&quot; are different pathnames.</p>
</blockquote>
<p><br>[Function]<br><a class=none name='pathname-relative-p'><b>pathname-relative-p</b> <i>a</i> =&gt; <i>result</i></a>
<blockquote>
<p>Returns true if <code><i>a</i></code> is a relative pathname. This simply
tests if <code><i>a</i></code>&#039;s directory starts
with <code>:RELATIVE</code>.</p>
</blockquote>
<p><br>[Function]<br><a class=none name='pathname-root-p'><b>pathname-root-p</b> <i>a</i> =&gt; <i>result</i></a>
<blockquote>
<p>Returns <em>true</em> if <code><i>pathname</i></code> is the root
directory (in other words, a directory which is its own parent).</p>
</blockquote>
<h4><a class=none name="manipulating">Manipulating pathnames</a></h4>
<p><br>[Function]<br><a class=none name='canonical-pathname'><b>canonical-pathname</b> <i>pathname</i> =&gt; <i>result</i></a>
<blockquote>
<p>Remove redundant information from PATHNAME.</p>
<p>This simply walks down <code>PATHNAME</code>&#039;s
pathname-directory and drops &quot;.&quot; directories, removes :back
and its preceding element.</p>
<p>NB: This function does not access the filesystem, it only looks at the
values in the pathname and works on their known (or assumed)
meanings.</p>
<p>NB: Since this function does not access the filesystem it will only
remove <code>:BACK</code> elements from the path (not <code>:UP</code>
elements). Since some lisps, ccl/sbcl/clisp convert &quot;..&quot; in
pathnames to <code>:UP</code>, and not <code>:BACK</code>, the actual
utility of the function is limited.</p>
</blockquote>
<p><br>[Function]<br><a class=none name='merge-pathnames-as-directory'><b>merge-pathnames-as-directory</b> <i><tt>&amp;rest</tt> pathnames</i> =&gt; <i>result</i></a>
<blockquote>
<p>Given a list of (probably relative) pathnames, this returns a single
directory pathname containing the logical concatenation of them all.</p>
<p>The returned value is the current directory if one were to cd into
each of <code><i>pathnames</i></code> in order. For this reason an
absolute pathname will, effectively, cancel the affect of any previous
relative pathnames.</p>
<p>The returned value&#039;s defaults are taken from the first element of
<code><i>pathnames</i></code> (host, version and device).</p>
<p><em>NB:</em> Since this function only looks at directory names the name and
type of the elements of <code><i>pathnames</i></code> are ignored. Make sure to properly
use either trailing #\/s, or <a href="#pathname-as-directory">pathname-as-directory</a>, to get the
expected results.</p>
<p>Examples:</p>
<pre>
(merge-pathnames-as-directory #P&quot;foo/&quot; #P&quot;bar/&quot;) == #P&quot;foo/bar/&quot;
(merge-pathnames-as-directory #P&quot;foo/&quot; #P&quot;./bar/&quot;) == #P&quot;foo/./bar/&quot;
(merge-pathnames-as-directory #P&quot;foo/&quot; #P&quot;/bar/&quot;) == #P&quot;/bar/&quot;
(merge-pathnames-as-directory #P&quot;foo/&quot; #P&quot;/bar/&quot; #P&#039;quux/file.txt) == #P&quot;/bar/quux/&quot;
</pre>
</blockquote>
<p><br>[Function]<br><a class=none name='merge-pathnames-as-file'><b>merge-pathnames-as-file</b> <i><tt>&amp;rest</tt> pathnames</i> =&gt; <i>result</i></a>
<blockquote>
<p>Given a list of, probably relative, pathnames returns a single
filename pathname containing the logical concatenation of them all.</p>
<p>The returned value&#039;s defaults are taken from the first element of
<code><i>pathnames</i></code> (host, version and device). The returned
values&#039;s name, type and version are taken from the last element
of <code><i>pathnames</i></code>. The intervening elements are used only for
their pathname-directory values.</p>
Examples:
<pre>
(merge-pathnames-as-file #P&quot;foo/&quot; #P&quot;bar.txt&quot;) == #P&quot;foo/bar.txt&quot;
(merge-pathnames-as-file #P&quot;foo/&quot; #P&quot;./bar.txt&quot;) == #P&quot;foo/./bar.txt&quot;
(merge-pathnames-as-file #P&quot;foo/&quot; #P&quot;/bar/README&quot;) == #P&quot;/bar/README&quot;
(merge-pathnames-as-file #P&quot;/foo/&quot; #P&quot;/bar/&quot; #P&#039;quux/file.txt) == #P&quot;/bar/quux/file.txt&quot;
</pre>
</blockquote>
<p><br>[Function]
<br><a class=none name="pathname-as-directory"><b>pathname-as-directory</b> <i> pathspec </i> =&gt; <i> pathname</i></a>
<blockquote><br>
Converts the <em>non-wild</em> <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_p.htm#pathname_designator">pathname designator</a> <code><i>pathspec</i></code> to <em>directory form</em>, i.e. it returns a <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_p.htm#pathname">pathname</a> which would return a <em>true</em> value if fed to <a href="#directory-pathname-p"><code>DIRECTORY-PATHNAME-P</code></a>.
</blockquote>
<p><br>[Function]
<br><a class=none name="pathname-as-file"><b>pathname-as-file</b> <i> pathspec </i> =&gt; <i> pathname</i></a>
<blockquote><br>
Converts the <em>non-wild</em> <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_p.htm#pathname_designator">pathname designator</a> <code><i>pathspec</i></code> to <em>file form</em>, i.e. it returns a <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_p.htm#pathname">pathname</a> which would return a <code>NIL</code> value if fed to <a href="#directory-pathname-p"><code>DIRECTORY-PATHNAME-P</code></a>.
</blockquote>
<p><br>[Function]<br><a class=none name='pathname-directory-pathname'><b>pathname-directory-pathname</b> <i>pathname</i> =&gt; <i>result</i></a>
<blockquote>
<p>Returns a complete pathname representing the directory of
<code><i>pathname</i></code>. If <code><i>pathname</i></code> is
already a directory pathname
(<code>name</code> <code>nil</code>, <code>type</code>
<code>nil</code>) returns a pathname equal (as
per <a href="#pathname-equal">pathname-equal</a>) to it.</p>
</blockquote>
<p><br>[Function]<br><a class=none name='pathname-parent-directory'><b>pathname-parent-directory</b> <i>pathname</i> =&gt; <i>result</i></a>
<blockquote>
<p>Returns a pathname which would, by name at least,
contain <code><i>pathname</i></code> as one of its direct
children. Symlinks can make the parent/child relationship a like
opaque, but generally speaking the value returned by this function is
a directory name which contains <code><i>pathname</i></code>.</p>
<p>The root directory, #P&quot;/&quot;, is its own parent. The parent
directory of a filename is the parent of the filename&#039;s
dirname.</p>
</blockquote>
<h4><a class=none name="traversing">Traversing directories</a></h4>
<p><br>[Function]
<br><a class=none name="list-directory"><b>list-directory</b> <i> dirname <tt>&amp;key</tt> follow-symlinks</i> =&gt; <i> list</i></a>
<blockquote><br>
<p>
Returns a <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_f.htm#fresh">fresh</a> list of pathnames corresponding to
all files within the directory named by the non-wild <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_p.htm#pathname_designator">pathname designator</a> <code><i>dirname</i></code>. The pathnames of sub-directories are returned in
<em>directory form</em> - see <a href="#pathname-as-directory"><code>PATHNAME-AS-DIRECTORY</code></a>.
</p>
<p>
If <code><i>follow-symlinks</i></code> is true (which is the
default), then the returned list contains truenames (symlinks will
be resolved) which essentially means that it might also return files
from <b>outside</b> the directory. This works on all platforms.
</p>
<p>
When <code><i>follow-symlinks</i></code> is <code>NIL</code>, it should return the actual directory
contents, which might include symlinks. (This is currently implemented only on SBCL and CCL.)
</p>
</blockquote>
<p><br>[Function]
<br><a class=none name="walk-directory"><b>walk-directory</b> <i> dirname fn <tt>&amp;key</tt> directories if-does-not-exist test follow-symlinks</i> =&gt; |</a>
<blockquote><br>
<p>
Recursively applies the function designated by the <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_f.htm#function_designator">function
designator</a> <code><i>fn</i></code> to all files within the directory named
by the non-wild <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_p.htm#pathname_designator">pathname
designator</a> <code><i>dirname</i></code> and all of its sub-directories. <code><i>fn</i></code>
will only be applied to files for which the function <code><i>test</i></code>
returns a <em>true</em> value. (The default value for <code><i>test</i></code>
always returns <em>true</em>.) If <code><i>directories</i></code> is not <code>NIL</code>,
<code><i>fn</i></code> and <code><i>test</i></code> are applied to directories
as well. If <code><i>directories</i></code> is <code>:DEPTH-FIRST</code>, <code><i>fn</i></code>
will be applied to the directory's contents first. If <code><i>directories</i></code>
is <code>:BREADTH-FIRST</code> and <code><i>test</i></code> returns <code>NIL</code>, the
directory's content will be skipped. <code><i>if-does-not-exist</i></code> must
be one of <code>:ERROR</code> or <code>:IGNORE</code> where <code>:ERROR</code>
(the default) means that an error will be signaled if the directory <code><i>dirname</i></code>
does not exist.
</p>
<p>
If <code><i>follow-symlinks</i></code> is true (which is
the default), then your callback will receive truenames. Otherwise
you should get the actual directory contents, which might include
symlinks. This might not be supported on all platforms. See
<a href="#list-directory"><code>LIST-DIRECTORY</code></a>.
</p>
</blockquote>
<h4><a class=none name="temporary-files">Temporary Files</a></h4>
<h5>Synopsis</h5>
<p>
Create a temporary file and return its name:
<pre>CL-USER&gt; (temporary-file:<code xmlns=""><a href="#with-output-to-temporary-file">with-output-to-temporary-file</a></code> (foo)
(print "hello" foo))
#P"/var/folders/Yu/YuNMNBNPGoqs9G-1Wmj1dk+++TI/-Tmp-/temp-yjck024x"</pre>
</p>
<p>
Create a temporary file, read and write it, have it be deleted
automatically:
<pre>CL-USER&gt; (temporary-file:<code xmlns=""><a href="#with-open-temporary-file">with-open-temporary-file</a></code> (foo :direction :io)
(print "hello" foo)
(file-position foo 0)
(read foo))
"hello"</pre>
</p>
<h5><a class="none" name="default-temporary-directory">Default temporary file directory</a></h5>
By default, temporary files are created in a system specific
directory that defaults based on operating system conventions. On
Unix and Unix-like systems, the directory <tt>/tmp/</tt> is used
by default. It can be overridden by setting the <tt>TMPDIR</tt>
environment variable. On Windows, the value of the environment
variable <tt>TEMP</tt> is used. If it is not set, temporary file
creation will fail.
<h5><a class="none" name="defining-temporary-directory">Defining the temporary file directory</a></h5>
<p>
The Lisp application can set the default directory in which
temporary files are created by the way of the
<code xmlns=""><a href="#temporary-files">temporary-files</a></code> logical pathname host:
<pre>(setf (<a xmlns="" href="http://www.lispworks.com/documentation/HyperSpec/Body/f_logica.htm">logical-pathname-translations</a> "<code xmlns=""><a href="#temporary-files">temporary-files</a></code>") '(("*.*.*" "/var/tmp/")))</pre>
This would set the directory for temporary files to
<tt>/var/tmp/</tt>. For more information about logical
pathnames, please refer to <a href="http://www.cs.cmu.edu/afs/cs/project/ai-repository/ai/html/cltl/clm/node208.html#SECTION002715000000000000000">Common
Lisp the Language, 2nd Edition</a> and the <a href="http://clhs.lisp.se/Body/19_.htm">Common Lisp
HyperSpec</a>.
</p>
<p>
Physical path names have restrictions regarding the permitted
character in file names. If these restrictions conflict with
your desired naming scheme, you can pass a physical pathname as
TEMPLATE parameter to the temporary file generation function.
</p>
<p>
Here are a few examples:
<pre>CL-USER&gt; (<a xmlns="" href="http://www.lispworks.com/documentation/HyperSpec/Body/f_logica.htm">logical-pathname-translations</a> "temporary-files")
(("*.*.*" #P"/var/folders/Yu/YuNMNBNPGoqs9G-1Wmj1dk+++TI/-Tmp-/"))
CL-USER&gt; (temporary-file:<code xmlns=""><a href="#with-open-temporary-file">with-open-temporary-file</a></code> (foo)
(<a xmlns="" href="http://www.lispworks.com/documentation/HyperSpec/Body/f_pn.htm">pathname</a> foo))
#P"/var/folders/Yu/YuNMNBNPGoqs9G-1Wmj1dk+++TI/-Tmp-/temp-6rdqdkd1"</pre>
This used the temporary directory established in the TMPDIR
environment variable, by the way of the definition of the
temporary-files logical host definition.
<pre>CL-USER&gt; (temporary-file:<code xmlns=""><a href="#with-open-temporary-file">with-open-temporary-file</a></code> (foo :template "/tmp/file.with.dots.in.name.%.txt")
(<a xmlns="" href="http://www.lispworks.com/documentation/HyperSpec/Body/f_pn.htm">pathname</a> foo))
#P"/tmp/file.with.dots.in.name.2EF04KUJ.txt"</pre>
Here, a physical pathname was used for the
<code xmlns=""><i>:template</i></code> keyword argument so that a
filename containing multiple dots could be generated.
<pre>CL-USER&gt; (temporary-file:<code xmlns=""><a href="#with-open-temporary-file">with-open-temporary-file</a></code> (foo :template "temporary-files:blah-%.txt")
(<a xmlns="" href="http://www.lispworks.com/documentation/HyperSpec/Body/f_pn.htm">pathname</a> foo))
#P"/var/folders/Yu/YuNMNBNPGoqs9G-1Wmj1dk+++TI/-Tmp-/blah-72mj450d.txt"</pre>
This used the temporary-files logical pathname host, but changed
the filename slightly.
<pre>CL-USER&gt; *default-pathname-defaults*
#P"/Users/hans/"
CL-USER&gt; (temporary-file:<code xmlns=""><a href="#with-open-temporary-file">with-open-temporary-file</a></code> (foo :template "blah-%.txt")
(<a xmlns="" href="http://www.lispworks.com/documentation/HyperSpec/Body/f_pn.htm">pathname</a> foo))
#P"/Users/hans/blah-5OEJELG2.txt"</pre>
Here, a relative pathname was used in the template, which
caused the file to be generated in the directory established
by <a xmlns="" href="http://www.lispworks.com/documentation/HyperSpec/Body/v_defaul.htm">*default-pathname-defaults*</a>.
</p>
<p>
Alternatively, the <code xmlns=""><a href="#*default-template*">*default-template*</a></code>
special variable can be set to define a custom default template
for generating names.
</p>
<h5 xmlns=""><a class="none" name="security">Security</a></h5>
The TEMPORARY-FILE library does not directly address security
issues. The application that uses it needs to take additional
measures if it is important that files created by one process
cannot be accessed by other, unrelated processes. This can be
done by using the system dependent security mechanisms like
default file permissions or access control lists.
<h5>Dictionary</h5>
<p xmlns="">[Function]<br><a class="none" name="open-temporary"><b>open-temporary</b> <i><clix:lambda-list xmlns:clix="http://bknr.net/clixdoc"><tt>&amp;rest</tt> open-arguments <tt>&amp;key</tt> template generate-random-string max-tries <tt>&amp;allow-other-keys</tt></clix:lambda-list></i>
=&gt;
<i>stream</i></a><blockquote><clix:description xmlns:clix="http://bknr.net/clixdoc">
<p xmlns="http://www.w3.org/1999/xhtml">
Create a file with a randomly generated name and return the
opened stream. The resulting pathname is generated from
<code xmlns=""><i>template</i></code>, which is a string
representing a pathname template. A percent sign (%) in
that string is replaced by a randomly generated string to
make the filename unique. The default for
<code xmlns=""><i>template</i></code> places temporary files in the
<code xmlns=""><a href="#temporary-files">temporary-files</a></code> logical pathname host,
which is automatically set up in a system specific manner.
The file name generated from <code xmlns=""><i>template</i></code>
is merged with <a xmlns="" href="http://www.lispworks.com/documentation/HyperSpec/Body/v_defaul.htm">*default-pathname-defaults*</a>,
so random pathnames relative to that directory can be
generated by not specifying a directory in
<code xmlns=""><i>template</i></code>.
</p>
<p xmlns="http://www.w3.org/1999/xhtml">
<code xmlns=""><i>generate-random-string</i></code> can be passed to
override the default function that generates the random name
component. It should return a random string consisting of
characters that are permitted in a pathname (logical or
physical, depending on <code xmlns=""><i>template</i></code>).
</p>
<p xmlns="http://www.w3.org/1999/xhtml">
The name of the temporary file can be accessed calling the
<a xmlns="" href="http://www.lispworks.com/documentation/HyperSpec/Body/f_pn.htm">pathname</a>
function on <code xmlns=""><i>stream</i></code>. For convenience,
the temporary file is opened on the physical pathname,
i.e. if the <code xmlns=""><i>template</i></code> designate a
logical pathname the translation to a physical pathname is
performed before opening the stream.
</p>
<p xmlns="http://www.w3.org/1999/xhtml">
In order to create a unique file name,
<code xmlns=""><a href="#open-temporary">open-temporary</a></code> may loop internally up
to <code xmlns=""><i>max-tries</i></code> times before giving up and
signalling a
<code xmlns=""><a href="#cannot-create-temporary-file">cannot-create-temporary-file</a></code> condition.
</p>
<p xmlns="http://www.w3.org/1999/xhtml">
Any unrecognized keyword arguments are passed to the call to
<a xmlns="" href="http://www.lispworks.com/documentation/HyperSpec/Body/f_open.htm">open</a>.
</p>
</clix:description></blockquote></p>
<p xmlns="">[Macro]<br><a class="none" name="with-output-to-temporary-file"><b>with-output-to-temporary-file</b> <i><clix:lambda-list xmlns:clix="http://bknr.net/clixdoc">(stream <tt>&amp;rest</tt> args) <tt>&amp;body</tt> body</clix:lambda-list></i>
=&gt;
<i>pathname</i></a><blockquote><clix:description xmlns:clix="http://bknr.net/clixdoc">
Create a temporary file using
<code xmlns=""><a href="#open-temporary">open-temporary</a></code> with
<code xmlns=""><i>args</i></code> and run <code xmlns=""><i>body</i></code>
with <code xmlns=""><i>stream</i></code> bound to the temporary file
stream. Returns the pathname of the file that has been
created. See <code xmlns=""><a href="#open-temporary">open-temporary</a></code> for
permitted options.
</clix:description></blockquote></p>
<p xmlns="">[Macro]<br><a class="none" name="with-open-temporary-file"><b>with-open-temporary-file</b> <i><clix:lambda-list xmlns:clix="http://bknr.net/clixdoc">(stream <tt>&amp;rest</tt> args <tt>&amp;key</tt> keep <tt>&amp;allow-other-keys</tt>) <tt>&amp;body</tt> body</clix:lambda-list></i>
=&gt;
<i>values</i></a><blockquote><clix:description xmlns:clix="http://bknr.net/clixdoc">
Create a temporary file using
<code xmlns=""><a href="#open-temporary">open-temporary</a></code> with
<code xmlns=""><i>args</i></code> and run <code xmlns=""><i>body</i></code>
with <code xmlns=""><i>stream</i></code> bound to the temporary file
stream. Returns the values returned by
<code xmlns=""><i>body</i></code>. By default, the file is deleted
when <code xmlns=""><i>body</i></code> is exited. If a true value is
passed in <code xmlns=""><i>keep</i></code>, the file is not deleted
when the body is exited. See
<code xmlns=""><a href="#open-temporary">open-temporary</a></code> for more permitted
options.
</clix:description></blockquote></p>
<p xmlns="">
[Special variable]<br><a class="none" name="*default-template*"><b>*default-template*</b></a><blockquote><clix:description xmlns:clix="http://bknr.net/clixdoc">
This variable can be set to a string representing the desired
default template for temporary file name generation. See
<code xmlns=""><a href="#open-temporary">open-temporary</a></code> for a description of the
template string format.
</clix:description></blockquote></p>
<p xmlns="">
[Condition type]<br><a class="none" name="cannot-create-temporary-file"><b>cannot-create-temporary-file</b></a><blockquote><clix:description xmlns:clix="http://bknr.net/clixdoc">
Signalled when an attempt to create unique temporary file name
failed after the established number of retries.
</clix:description></blockquote></p>
<p xmlns="">
[Condition type]<br><a class="none" name="invalid-temporary-pathname-template"><b>invalid-temporary-pathname-template</b></a><blockquote><clix:description xmlns:clix="http://bknr.net/clixdoc">
Signalled when the <code xmlns=""><i>template</i></code> argument to
<code xmlns=""><a href="#open-temporary">open-temporary</a></code> does not contain a valid
template string. The template string must contain a percent
sign, which is replaced by the generated random string to
yield the filename.
</clix:description></blockquote></p>
<p xmlns="">
[Condition type]<br><a class="none" name="missing-temp-environment-variable"><b>missing-temp-environment-variable</b></a><blockquote><clix:description xmlns:clix="http://bknr.net/clixdoc">
(Windows only) Signalled when the TEMP environment variable is
not set.
</clix:description></blockquote></p>
<p xmlns="">
[Logical Pathname Host]<br><a class="none" name="lp-host-temporary-files"><b>temporary-files</b></a><blockquote><clix:description xmlns:clix="http://bknr.net/clixdoc">
This logical pathname host defines where temporary files are
stored by default. It is initialized in a suitable system
specific fashion: On Unix and Unix-like systems, the directory
specified in the TMPDIR environment variable is used. If that
variable is not set, /tmp is used as the default. On Windows,
the directory specified in the TEMP environment variable is
used. If it is not set, a
<code xmlns=""><a href="#missing-temp-environment-variable">missing-temp-environment-variable</a></code> error
is signalled.
</clix:description></blockquote></p>
<h4><a class=none name="modifying">Modifying the file system</a></h4>
<p><br>[Function]
<br><a class=none name="copy-file"><b>copy-file</b> <i> from to <tt>&amp;key</tt> overwrite</i> =&gt; |</a>
<blockquote><br>
Copies the file designated by the non-wild <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_p.htm#pathname_designator">pathname designator</a> <code><i>from</i></code> to the
file designated by the non-wild <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_p.htm#pathname_designator">pathname designator</a> <code><i>to</i></code>. If <code><i>overwrite</i></code> is <em>true</em> (the default is <code>NIL</code>)
overwrites the file designated by <code><i>to</i></code> if it exists.
</blockquote>
<p><br>[Function]
<br><a class=none name="copy-stream"><b>copy-stream</b> <i> from to <tt>&amp;optional</tt> checkp</i> =&gt; |</a>
<blockquote><br> Copies into <code><i>to</i></code> (a stream)
from <code><i>from</i></code> (also a stream) until the end
of <code><i>from</i></code> is reached. The streams should have the
same <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/f_stm_el.htm">element
type</a> unless they are bivalent. If <code><i>checkp</i></code> is
true (which is the default), the function will signal an error if the
element types aren't the same.
</blockquote>
<p><br>[Function]
<br><a class=none name="delete-directory-and-files"><b>delete-directory-and-files</b> <i> dirname <tt>&amp;key</tt> if-does-not-exist</i> =&gt; |</a>
<blockquote><br>
<p>
Recursively deletes all files and directories within the directory
designated by the non-wild <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_p.htm#pathname_designator">pathname designator</a> <code><i>dirname</i></code> including
<code><i>dirname</i></code> itself. <code><i>if-does-not-exist</i></code> must be one of <code>:ERROR</code> or <code>:IGNORE</code>
where <code>:ERROR</code> (the default) means that an error will be signaled if the directory
<code><i>dirname</i></code> does not exist.
</p>
<p>
<b>Warning:</b> this function <em>might</em> remove files from outside the
directory, if the directory that you are deleting contains links to
external files. This is currently fixed for SBCL and CCL.
</p>
</blockquote>
<h4>The <code>PATH</code> package</h4>
<p><br>[Package]
<br><a class=none name="package-path">(defpackage <b>path</b>)</a>
<blockquote>
Provides a set of short names for commonly used pathname manipulation
functions (these are all functions from the <code>cl-fad</code>
package which are being exported under different names):
<dl>
<dt><code>dirname</code></dt> <dd><a href="#pathname-as-directory">pathname-as-directory</a></dd>
<dt><code>basename</code></dt> <dd><code>cl:file-namestring</code></dd>
<dt><code>-e</code></dt> <dd><a href="#file-exists-p">file-exists-p</a></dd>
<dt><code>-d</code></dt> <dd><a href="#directory-exists-p">directory-exists-p</a></dd>
<dt><code>catfile</code></dt> <dd><a href="#merge-pathnames-as-file">merge-pathnames-as-file</a></dd>
<dt><code>catdir</code></dt> <dd><a href="#merge-pathnames-as-directory">merge-pathnames-as-directory</a></dd>
<dt><code>rm-r</code></dt> <dd><a href="#delete-directory-and-files">delete-directory-and-files</a></dd>
<dt><code>=</code></dt> <dd><a href="#pathname-equal">pathname-equal</a></dd>
<dt><code>absolute-p</code></dt> <dd><a href="#pathname-absolute-p">pathname-absolute-p</a></dd>
<dt><code>relative-p</code></dt> <dd><a href="#pathname-relative-p">pathname-relative-p</a></dd>
<dt><code>root-p</code></dt> <dd><a href="#pathname-root-p">pathname-root-p</a></dd>
</dl>
</blockquote>
<br>&nbsp;<br><h3><a class=none name="ack">Acknowledgements</a></h3>
The original code for this library was written by Peter Seibel for his
book <a href="http://www.gigamonkeys.com/book/"><em>Practical Common
Lisp</em></a>. I added some stuff and made sure it worked properly on
Windows, specifically with CCL. Thanks to James Bielman, Maciek
Pasternacki, Jack D. Unrue, Gary King, and Douglas Crosher who sent
patches for OpenMCL, ECL, ABCL, MCL, and Scieneer&nbsp;CL.
<p>
$Header: /usr/local/cvsrep/cl-fad/doc/index.html,v 1.33 2009/09/30 14:23:12 edi Exp $
<p><a href="http://weitz.de/index.html">BACK TO MY HOMEPAGE</a>
</body>
</html>