Merge pull request #2255 from xexyl/author_handle_re
Fix regexp in faq.md
This commit is contained in:
2 files changed
+11
-9
No files matched your search
@@ -2319,10 +2319,11 @@ form:</p>
|
||||
<p>See <a href="#author_json">FAQ 6.6</a> for information about the
|
||||
contents of these JSON file and how they are used.</p>
|
||||
<p>Because the <code>author_handle</code> is used to form a JSON
|
||||
filename, the string must be POSIX safe string. Furthermore, the
|
||||
<code>author_handle</code> must be an ASCII string that matches this
|
||||
regexp:</p>
|
||||
<pre class="re"><code>^[0-9A-Za-z][0-9A-Za-z_]*$</code></pre>
|
||||
filename, the string must be POSIX safe string plus the use of
|
||||
<code>+</code> (for technical reasons beyond this answer). In
|
||||
particular, the <code>author_handle</code> must be an ASCII string that
|
||||
matches this regexp:</p>
|
||||
<pre class="re"><code>^[0-9A-Za-z][0-9A-Za-z._+-]*$"</code></pre>
|
||||
<p>Default <code>author_handle</code>’s do not have multiple consecutive
|
||||
<code>_</code> (underscore) characters. Nevertheless, multiple
|
||||
consecutive <code>_</code> (underscore) characters are allowed. Contest
|
||||
@@ -2814,7 +2815,7 @@ not see who wrote the code.</p>
|
||||
IOCCC becomes an <em>entry</em>. Only then do the IOCCC judges look into
|
||||
the special JSON file to discover who the <em>author</em>(s) are.</p>
|
||||
<p><strong>An apology for being inconsistent</strong>: In the past, the
|
||||
IOCCC used the term <em>entry</em> to refer to what is now an
|
||||
IOCCC used the term <em>entry</em> to refer to what is now a
|
||||
<em>submission</em>. In a number of historical cases, such as old rules
|
||||
and old guidelines, terms such as <em>entry</em> may still be found.
|
||||
Moreover, out of habit, the IOCCC judges sometimes use old names such as
|
||||
|
||||
@@ -2504,11 +2504,12 @@ See [FAQ 6.6](#author_json) for information about the contents of these JSON fil
|
||||
how they are used.
|
||||
|
||||
Because the `author_handle` is used to form a JSON filename, the string must be
|
||||
POSIX safe string. Furthermore, the `author_handle` must be an ASCII
|
||||
string that matches this regexp:
|
||||
POSIX safe string plus the use of `+` (for technical reasons beyond this
|
||||
answer). In particular, the `author_handle` must be an ASCII string that
|
||||
matches this regexp:
|
||||
|
||||
```re
|
||||
^[0-9A-Za-z][0-9A-Za-z_]*$
|
||||
^[0-9A-Za-z][0-9A-Za-z._+-]*$"
|
||||
```
|
||||
|
||||
Default `author_handle`'s do not have multiple consecutive `_` (underscore)
|
||||
@@ -3125,7 +3126,7 @@ becomes an _entry_. Only then do the IOCCC judges look into
|
||||
the special JSON file to discover who the _author_(s) are.
|
||||
|
||||
**An apology for being inconsistent**: In the past, the IOCCC used the term _entry_ to
|
||||
refer to what is now an _submission_. In a number of historical cases,
|
||||
refer to what is now a _submission_. In a number of historical cases,
|
||||
such as old rules and old guidelines, terms such as _entry_ may still be
|
||||
found. Moreover, out of habit, the IOCCC judges sometimes use old
|
||||
names such as _entry_ when they should use _submission_. Sorry (tm Canada)! :-)
|
||||
|
||||
Reference in new issue
Block a user