Documentation fixes

8 files changed, 54 insertions, 14 deletions

diff --git a/doc/crosscompile.html b/doc/crosscompile.html
index ba22b31..4098afd 100644
--- a/doc/crosscompile.html
+++ b/doc/crosscompile.html
@@ -76,8 +76,10 @@ architecture</strong>, via the --with-sysdeps option to configure.
 
 <p>
  The easiest way to get the correct sysdeps for a target achitecture is
-to natively compile skalibs on that target, and steal the produced sysdeps
-files. It can be easily done with a virtual machine, qemu for instance.
+to natively run skalibs' <tt>./configure</tt> script on that target,
+and steal the produced sysdeps files, which are normally written to the
+<tt>./sysdeps.cfg</tt> directory.
+Doing this is easy with a virtual machine, qemu for instance.
 You could also (politely) ask for precompiled sysdeps on the
 skaware mailing-list, if you cannot find them anywhere on the Internet.
 </p>
diff --git a/doc/flags.html b/doc/flags.html
index 2579d5a..2357445 100644
--- a/doc/flags.html
+++ b/doc/flags.html
@@ -43,6 +43,10 @@ symbolic link to point to <tt>skalibs-$v</tt> </li>
 to the installed skalibs shared libraries. </li>
 </ul>
 
+<p>
+ If this option is not given, no slashpackage support will be provided.
+</p>
+
 <a name="replace-libc"><h3> --enable-libc-replacements </h3></a>
 
 <p>
@@ -260,8 +264,8 @@ support.
 
 <p>
  If you don't set this option, then skalibs will include IPv6 support in the
-relevant networking functions, if the target architecture supports it.
-The safe option is to let this flag clear.
+relevant networking functions if and only if the target architecture supports it.
+This is the default, and it is safe.
 </p>
 
 <a name="forcedevr"><h3> --enable-force-devr </h3></a>
@@ -275,9 +279,9 @@ its autodetection.
 <p>
  Otherwise, <tt>/dev/random</tt> will be autodetected
 and tested; if entropy generation is low on the host, the compilation
-process might hang for several minutes. It is safe to let this flag
+process might hang for several minutes. It is safe to leave this flag
 clear; it should only be set to speed up the compilation process in a
-known environment and for testing purposes.
+known environment or for testing purposes.
 </p>
 
 <p>
diff --git a/doc/libskarnet.html b/doc/libskarnet.html
index 9c29514..6c8b559 100644
--- a/doc/libskarnet.html
+++ b/doc/libskarnet.html
@@ -60,14 +60,14 @@ to run <tt>ldconfig</tt> if needed. </li>
 installed: by default, it's <tt>/usr/lib/skalibs/sysdeps</tt>. Let's call
 it <tt>$sysdeps</tt>. </li>
  <li> Link with <tt>-lskarnet</tt>. If you are using socket functions, you
-will also need to link with <tt>-l$sysdeps/socket.lib</tt>. If you are using
+will also need to link with <tt>-l`cat $sysdeps/socket.lib`</tt>. If you are using
 time functions such as <tt>tain_now()</tt>, you will also need to link with
-<tt>-l$sysdeps/tainnow.lib</tt>. </li>
+<tt>-l`cat $sysdeps/tainnow.lib`</tt>. </li>
 </ul>
 
 <p>
- The <em>skarnet</em> library as a whole is big (833k for libskarnet.a and
-257k for libskarnet.so.2.0.0.0 on x86_64), but the utmost care has been
+ The <em>skarnet</em> library as a whole is big (826k for libskarnet.a and
+233k for libskarnet.so.2.0.0.0 on x86_64-linux), but the utmost care has been
 given to separate functions so that linkers never pull in any more than they
 need. Linking against the static version of libskarnet actually produces
 very small executables, and if your libc is suited for that, since skalibs
diff --git a/doc/libstddjb/djbunix.html b/doc/libstddjb/djbunix.html
index 37235c1..02f8624 100644
--- a/doc/libstddjb/djbunix.html
+++ b/doc/libstddjb/djbunix.html
@@ -465,7 +465,7 @@ The parent's end of the pipe will be stored in *<em>fd</em>.
 </p>
 
 <p>
-<code> pid_t child_spawn1_socket (char const *file, char const *const *argv, char const *const *envp, int *fd)
+<code> pid_t child_spawn1_socket (char const *file, char const *const *argv, char const *const *envp, int *fd) </code> <br />
 Like <tt>child_spawn0</tt>, except that a socket is created between the parent
 and the child. Both the child's stdin and stdout point to the socket; the parent has
 its end of the socket available in *<em>fd</em>.
diff --git a/doc/libstddjb/genwrite.html b/doc/libstddjb/genwrite.html
index 33ff4d3..3bf6fd8 100644
--- a/doc/libstddjb/genwrite.html
+++ b/doc/libstddjb/genwrite.html
@@ -46,7 +46,7 @@ stuff to the target without flushing it
 (which can be <tt>genwrite_put_stralloc</tt>, <tt>genwrite_put_buffer</tt>,
 <tt>genwrite_put_bufalloc</tt> or any
 compatible user-defined function) in <tt>.put</tt>, a pointer to a function
-that flushes the target (which can be <tt>genwrite_flush_stralloc</tt>,
+that flushes the target (which can be <tt>genwrite_flush_stralloc</tt>,
 <tt>genwrite_flush_buffer</tt>, <tt>genwrite_flush_bufalloc</tt> or any
 compatible user-defined function) in <tt>.flush</tt>, and a pointer to
 the target writing structure in <tt>.target</tt>.
diff --git a/doc/libstddjb/iopause.html b/doc/libstddjb/iopause.html
index 72eedde..da6ff48 100644
--- a/doc/libstddjb/iopause.html
+++ b/doc/libstddjb/iopause.html
@@ -188,7 +188,7 @@ if it is available; but if it is not, then <tt>iopause_poll</tt> defaults to
 which has a more comfortable API than
 <a href="http://pubs.opengroup.org/onlinepubs/9699919799/functions/select.html">select()</a>,
 but a maximum precision of 1 millisecond which might not be enough for some applications; whereas
-<tt>iopause_select<tt> defaults to select(), which incurs some CPU overhead for the
+<tt>iopause_select</tt> defaults to select(), which incurs some CPU overhead for the
 API conversion, but has a 1 microsecond precision.
 </p>
 
diff --git a/doc/libstddjb/safewrappers.html b/doc/libstddjb/safewrappers.html
index 37d2425..da8608b 100644
--- a/doc/libstddjb/safewrappers.html
+++ b/doc/libstddjb/safewrappers.html
@@ -70,7 +70,7 @@ signal is caught. And that means SA_RESTART should not be used.
 </p>
 
 <p>
- Which means that other system calls performed when the signal handler has
+ Which also means that other system calls performed when the signal handler has
 been installed, for instance in the body of the loop, will <em>not</em> be
 protected, and can return -1 EINTR if a signal is caught at the wrong time.
 </p>
diff --git a/doc/libunixonacid/skaclient.html b/doc/libunixonacid/skaclient.html
index 7deb38b..f4b3020 100644
--- a/doc/libunixonacid/skaclient.html
+++ b/doc/libunixonacid/skaclient.html
@@ -27,6 +27,40 @@ and implemented in the <tt>libskarnet.a</tt> or <tt>libskarnet.so</tt> library.
 <h2> General information </h2>
 
 <p>
+ <tt>skaclient</tt> is a client-server initiation protocol framework, allowing a process
+(the "client") to either connect to another process (the "server") via a Unix domain socket,
+or spawn such a "server" process itself. The client and the server then communicate
+via <em>two</em> sockets, one for synchronous data, the other for asynchronous
+data; they exchange <a href="unixmessage.html">unixmessages</a>.
+</p>
+
+<p>
+ The <tt>skaclient</tt> framework is used in several places in skarnet.org software,
+whenever a server can send asynchronous data to its client. For instance:
+</p>
+
+<ul>
+ <li> Communication between notification subscribers such as
+<a href="http://skarnet.org/software/s6/s6-ftrig-wait.html">s6-ftrig-wait</a>
+and their own <a href="http://skarnet.org/software/s6/s6-ftrigrd.html">s6-ftrigrd</a>
+daemon </li>
+ <li> Communication between clients of the
+<a href="http://skarnet.org/software/s6/libs6lock/">s6lock library</a> and a
+<a href="http://skarnet.org/software/s6/libs6lock/s6lockd.html">s6lockd
+server</a> </li>
+ <li> Communication between a client such as
+<a href="http://skarnet.org/software/s6-dns/s6-dnsip4-filter.html">s6-dnsip4-filter</a>,
+in need of asynchronous DNS service, and its own
+<a href="http://skarnet.org/software/s6-dns/skadns/skadnsd.html">skadnsd</a>
+daemon </li>
+ <li> The <a href="http://skarnet.org/software/skabus/">skabus</a> Unix bus
+infrastructure is entirely based on the skaclient framework to transmit
+unixmessages across processes. </li>
+</ul>
+
+<h2> Programming </h2>
+
+<p>
 FIXME: to be completed.
 </p>