add hack-a-round for language specified markdown code blocks

When language is specified in a markdown code block such as:

````
```c
    printf("Hello, world!\n");
```
````

is used, the resulting HTML as produced by `pandoc(1)` combined
with our revised `ioccc.css` is not intent the code block well.

A temporary "_hack-a-round_" is to turn the "markdown code language
specifier" into an HTML comment.  For example:

````
```<!---c-->
    printf("Hello, world!\n");
```
````

This commits turns off existing markdown code language specifiers
for now, while preserving our ability to restore them in the future
when a better  "_hack-a-round_" to the "markdown code language
specifier" problem is found.

The HTML impact if this "_hack-a-round_" is to produce HTML that
looks like this:

```
<pre class="&lt;!---c--&gt;"><code> .. </code></pre>
```

We might be able to create a complex set of classes for each markdown
code language (there won't be very many) in `ioccc.css`, or we might
find a more general CSS (or maybe `pandoc(1)`?) solution.

For now, this hack-a-round will have to do.

1990/dg/README.md

@@ -1,20 +1,20 @@
 ## To build:
 
-```sh
+```<!---sh-->
     make all
 ```
 
 
 ## To use:
 
-```sh
+```<!---sh-->
     echo foo bar | ./dg
 ```
 
 
 ## Try:
 
-```sh
+```<!---sh-->
     ./try.sh
 ```
 
@@ -24,13 +24,13 @@
 Some C preprocessors got confused about single quotes in certain cases.
 In particular, the following source line:
 
-```c
+```<!---c-->
     b12(ase(bco(acp(bi(acp(bcp(acp(bpl(aqu(bqu(A))))))))))))))))))))))))))))))
 ```
 
 sometimes has to be changed to:
 
-```c
+```<!---c-->
     b12(ase(bco(acp(bi(acp(bcp(acp(bpl('A'))))))))))))))))))))))))))))
 ```
 
@@ -56,7 +56,7 @@ As for the rot13 algorithm, it relies on `index(3)` to find if the
 character is in the set `A-Za-z`. `l[]` is an array that gets filled
 with:
 
-```c
+```<!---c-->
     "nopqrstuvwxyzzzzabcdefghijklmmmmNOPQRSTUVWXYZZZZABCDEFGHIJKLMMMM"
 ```