README, Configure, and Orwell
It would be a good idea for the README's "Building BIND" section to
mention the option to use a cache for configure
.
One of:
-C
--cache-file=../bind-config.cache
It is likely that many people who build from source are unaware of these
options - the only mention is in configure
's --help
, which is terse to the
point of being cryptic.
While there are some risks, this can dramatically reduce the time needed to configure (and thus build) BIND.
Perhaps this language:
If you frequently build BIND from source in a stable environment, you
should consider using a cache for configure
, which can dramatically
reduce the time required. To do this, use either:
-
-C
- forconfigure
to cache its results inconfig.cache
or -
--cache-file=FILE
(e.g.--cache-file=../bind-config.cache
) to specify the file
If your environment changes (e.g. your distribution, C compiler, or libraries), you may need to delete the cache file in order for BIND to build correctly.
If you build for more than one environment, use a unique cache file for each.
It is generally not necessary to delete the cache when upgrading BIND. But if you encounter build failures, deleting the cache is likely to resolve them.
In any case, before submitting a bug report, please delete or disable the cache file and re-build bind to ensure that your issue is not due to a stale cache.
For additional information on the configure cache
, see the autoconf doc
While on the README: the "Acknowledgements" section says
" * The original development of BIND 9 was underwritten by the following organizations:
Sun Microsystems, Inc.
Hewlett Packard
Compaq Computer Corporation
..."
My recollection is that Digital Equipment Corporation was one of the original (and long-time) underwriters. Certainly Vixie's salary was paid by DEC when it worked on it in the V4.9 timeframe.
Perhaps Compaq contributed, but it wasn't visible to me. Compaq didn't exist until 1982, and didn't become interested in enterprise software until much later.
While it is true that DEC was "merged" with Compaq, it is also true that Compaq was absorbed by HP.
As a long-time employee of DEC (and then Compaq, HP, and Intel), I'd prefer not to have history revised in this way. Preferably, put DEC back on the list (and remove CPQ unless they actually did underwrite BIND).
Or remove CPQ on the theory that DEC's goodwill (an accounting term) was ultimately acquired by HP.
If history is to be revised, at least the methodology should be consistent...
This same text appears on your website, in the ARM, and elsewhere. All references should, of course, match.