the htdocs/ directories.
While here, fix the hostname(1) evaluation for non-en_US
translations.
These scripts used to fetch the ports.tgz file from FTP, but
this file no longer exists. Since we already reference the
size of the ports tree (via share/xml/freebsd.ent) for the
Handbook, I do not plan to bring back this functionality, as
the information it was producing was incorrect anyway.
Note to translators of de_DE, ja_JP, and ru_RU trees:
The relevant files in the htdocs/ directories also needed to
be updated as part of this commit to avoid breaking the build.
Because my editor does not seem to properly handle ja_JP or
ru_RU, editing the installing.xml files within those trees was
not an option, as it mangled the file after saving. A sed(1)
substitution to comment the '&ports.size;' macro was done here
instead, but the entire line referencing ftp.freebsd.org should
be removed instead.
Approved by: doceng (implicit)
Sponsored by: The FreeBSD Foundation
Switch the 'packages.exist' target to use pkg.freebsd.org, since
we do not publish packages on FTP mirrors anymore.
Approved by: doceng (implicit)
Sponsored by: The FreeBSD Foundation
- remove article about releng packages
- we no longer produce multi-cd releases
- CVS is not our VCS system
- we havn't releaseed with xfree86 is years
- section 1 about the 'ports cluster' is now entirely untrue and
been disclaimed by portmgr
Discussed with: peter, bdrewery, Daniel O'Connor
- references mostly outdated hardware
- doesn't use modern technologies (IPMI, nmdm, bhyve)
- the SUN hardware peculiarities are not useful anymore
- there are other better references on crimping custom serial cables
- MAKEDEV is no a thing anymore
there might be room for a modern version of this article in the handbook
but there is little that is salvageable from this article.
Discussed with: bcr, xmj, sbruno, cem
- using a custom gcc should be done within poudriere
- there is a framework for choosing a ports compiler anyways
- this article is incomplete w.r.t. modern gcc verisons
- encouraging people to add global options like `mssse3` is dangerous
- the original reason to use a custom gcc is gone (our old gcc was
outdated)
- comparing 'Linux' vs 'FreeBSD' is comparing apples and fruit: one is
just the kernel and the other is a specific distribution.
- the license 'debate' is well covered elsewhere
- it still references CVS
- comparing the development style of '300 committers' vs 'going through
the maintainer' is a straw man and doesn't reflect how Linux works (or
how BSD works)
- the supported platforms list is old and incorrect
- the "Extensible Frameworks" section belongs elsewhere
- the 'security' section is a feature list, not a comparison and is
missing important features
etc.
Discussed with: mjg
If I understand correctly the build broke because the non-english articles
import from the english version of fbsd-from-scratch.
Remove the translated articles for the articles I removed in English
- Replace /XML/{doc,www}/ with /XML/ in SysId.
- Remove empty stylesheets in share/xsl and point share/xml/empty.xsl via
XML catalog instead.
- Change the L10N layer in freebsd-*.xsl not to use localized XSLT
stylesheets directly.
- Move share/xsl/* to share/xml and remove share/xsl.
- Remove obsolete share/web2c/pdftex.def.
now picked up via share/xsl/*.xsl and they are empty when there is no
localized content in it. This prevents inconsistency between share/xsl/*.xsl
and the localized ones.
picked up via XML catalog in freebsd50.ent or xhtml10-freebsd.dtd,
not entities.ent. The L10N entities always come first to be able to
override everything.
- Define as   in EUC-JP encoding. This is a workaround to
prevent an invalid character in EUC-JP caused by converting 0xa0 in
UTF-8 (EUC-JP does not allow 0xa0) in XSLT processing. Theoretically
it should be   still in the final XML output.
- Make XML catalog resolution consistent for l10n.ent. This should be
revisited for the others later.
