In this issue

Some readers subscribed to the text version of the Apache Week mailing list received the HTML version of issue 97. This was caused by one particular subscriber whose mail system was setup to relay mail based on the To: address in the headers (mail relays should always use the address in the mail "envelope", not the header). Like all other mailing lists, Apache Week uses the address of the list in its To: field, so this user's mail system sent the message back to the list. The list is configured to prevent mail from subscribers going out onto the list, but in this case the mail was from the user's machine, not the user themselves.

The user in question was on the HTML version of the list. However because of the way that the Apache Week headers are configured, when the mail was relayed back from their site it was sent out to the text list. Starting from this issue, the headers are configured to ensure that the lists are kept separate. The user involved and their connectivity provider are both aware of what happened and have fixed the problem.

If you did not receive the text version of Apache Week issue 97 (dated 9th January 1998), you can obtain it by:

• Visiting http://www.apacheweek.com/issues/98-01-09.
• Getting it via mail by sending a message to majordomo@apacheweek.com containing the text get apacheweek 98-01-09.txt (note: send the index apacheweek command to get a list of all the back-issues available by mail)

We have not altered the subscriber lists for either the HTML or text versions of Apache Week at all. If you subscribed to the "apacheweek" list you will get the text version, and if you subscribed to the "apacheweek-html" list you will get the HTML version (suitable for reading in any HTML-enabled mail reader).

To unsubscribe from the apacheweek list, send a message to majordomo@apacheweek.com containing just the text

  unsubscribe apacheweek

To unsubscribe from apacheweek-html, the contents of the message should be unsubscribe apacheweek-html instead.

Apache Site: www.apache.org
Release: 1.2.5 (Released 5^th January 1998) (local download sites)
Beta: 1.3b3 (Released 20^th November 1997) (local download sites)

Apache 1.2.5 is the current stable release. Users of Apache 1.2.4 and earlier should upgrade to this version since it fixes a number of bugs and potential security problems. The next major release will be 1.3. A beta test release of 1.3 is available now for both Unix and Windows 95/NT systems.

These bugs have been reported in 1.3b3 and are either not fixed, or have not been reproduced by the Apache developers.

• TABs or multiple spaces in CustomLog formats get converted to a single space.

These bugs have been found and fixed in 1.3b4.

Because of the major differences between Windows and Unix, these are separated into bugs which affect Windows systems only, and other bugs (which may affect Windows as well). Unix users can ignore the bugs listed in the Windows section.

Windows-specific Bugs

• Output from CGI programs is buffered within Apache. This is because Windows does not support non-blocking output to sockets using the normal (BSD) sockets interface.
• Apache goes into an infinite loop if the number of threads set by ThreadsPerChild is greater than 64.

Other Bugs

• Some unnecessary memory allocation has been removed.
• If mod_speling found an ambiguous URL, it would return a list of choices containing invalid links to the actual files.
• Removed warnings when compiling on SVR4 systems.
• Apache now compiles on Siemens Nixdorf BS2000-OSD/POSIX mainframes which use the EBCDIC charset (instead of ASCII). This might make it easier to port to othe EBCDIC systems as well.
• Apache uses a case-sensitive match to determine whether to use basic authentication. The word "basic" has to be given by the client as "Basic" (not, for example, "basic" or "BASIC").
• If a HeaderFile contains plain text (not HTML), Apache outputs the <PRE> tag before the preamble instead of after it.

Patches for bugs in Apache 1.2.5 may be made available in the apply to 1.2.5 directory on the Apache site. Some new features and other unofficial patches are available in the 1.2 patches directory (these may not apply cleanly to 1.2.5). For details of all previously reported bugs, see the Apache bug database and known bugs pages. Many common configuration questions are answered in the Apache FAQ.

Development has slowed down to prepare for the release of Apache 1.3. During the beta release cycle Apache is in a "feature freeze" where no major new features will be added.

From the next beta of 1.3, it will be possible to give directives on the command line. The -C option gives a directive to execute before reading the configuration files (httpd.conf, etc), and -c gives a directive to execute after reading the configuration files. Both of these can be given multiple times to specify multiple directives, but sections (such as <VirtualHost> cannot be used.

These options were added to enhance support for the mod_perl module. It will now be possible to configure Apache completely within perl, without using configuration files at all, using:

  httpd -C "PerlModule Apache::httpd_conf"

This would cause mod_perl's PerlModule directive to be executed before any configuration files are read.

A bug in Netscape Navigator can cause it to repeatedly request the same GIF file. This occurs when an animated GIF contains a loop and when it has an expiry date. Navigator will request the GIF every time around the animation loop, even if the GIF file came with an expiry date sometime in the future. This bug was first reported in Summer 1996, but it is still in the latest versions of Navigator (4.02).

A patch is in progress to work around this behaviour. The workaround is to remove the Expires header from GIFs when the client is a buggy version of Navigator (i.e. versions 1 through 4).

Various recent press reports have incorrectly described the licence conditions for using Apache. In addition, some users are under the impression that Apache is covered by the GNU General Public Licence (GPL). This article will attempt to clarify this issue.

Apache is free. This does not mean, of course, that it is free from copyright and licence, nor does it mean that it is totally free of all associated costs, such as disk space costs. Because the use of Apache is restricted by a licence, it is not public domain. Since it is free, Apache is often described as "freeware". But it might be more appropriate to regard Apache as just like any other commercial software, except that it costs nothing to buy.

Apache is not shareware. The term "shareware" is not a term normally applied to Unix software. It is widely used to describe personal computer software that is distributed at low cost, and where the user can try it before they must pay. Because of the low-cost distribution, many small companies and individuals distribute their software as shareware. However, there is essentially no difference between "shareware" and "commercial" software -- you are expected to pay for both. Today, many software providers make their software available for testing before purchase -- including both Microsoft and Netscape -- so their software could also be described as "shareware".

Apache is not covered by the GPL licence. The GPL would prevent Apache's use in third party products which did not themselves provide source code. People can build Apache into their products (either commercial or free) and there is no requirement for the source to be provided with these products. (The licence does place restrictions on people who use Apache code, but these apply to advertising and documentation, not the use of the code itself). Incidently, the GNU project has a list of software categories (free, commercial, copyleft, etc) the might be of interest on their site. Note that GNU uses "free" in relation to continued availability of source code, not cost.

While the above attempts to simplify the issue, in practise it is more complex. For example, there are commercial and shareware products that come with no support, which is often associated with free products. Similarly, some commercial software is available for free, and some free software (such as Apache) is viably competing with traditional commercial software.

In Apache Week issue 96 we reported on a story in ZD Net's Internet Magazine that incorrectly described Apache. It also featured a Zona Research survey that gave Apache an 8% share of the Internet server market. The author of this story has replied, in Into Hot Air: Online Editor Albert Pang Examines Debates over Apache's Market Share, Understated or Otherwise.

This article is spread over several pages, and features a number of reactions from readers. While there are still some dubious assertions -- such as more Zona Reseach figures that claim that 52% of Internet servers are today running on NT -- the most serious issue surrounds how to describe Apache. To help clarify this, we have described the Apache licence above. The article makes the following points about the use of the term "freeware" when applied to Apache:

"I have a hard time accepting the fact that Apache is freeware because there are many vendors and developers that are profiting from Apache by selling add-ons and other services".

Third parties are actually encouraged to develop products based on or for use with Apache. This does not affect the fact that Apache code continues to be available for free.

"Another problem with the term freeware is that one still has to enter into a licensing agreement with the author before using the product. My question is how can there be strings attached to something that is ostensibly free."

A licence agreement on a free product does not stop the product itself from being free. The answer to the question is to prevent people from doing whatever they want with the code - even saying that they developed it themselves.

"Even when it's free to download Apache, it requires disk space to store it".

This is getting too absurd to comment on.