summaryrefslogtreecommitdiff
path: root/pkgtools
AgeCommit message (Collapse)AuthorFilesLines
2007-11-19Even on non-BSD systems, the "-R all" option should disable printing ofrillig1-3/+5
the "time" attribute. Noticed by Malcolm Herbert on tech-pkg.
2007-11-17Don't install corrupted catpages on Solaris and AIX. PKGREVISION++rillig1-19/+12
Cleaned up the Makefile.
2007-11-16The type of BUILTIN_FIND_GREP.* had been wrong. It's not List ofrillig1-2/+2
ShellWord, but a simple String. This has led to some confusion in db1.builtin.mk.
2007-11-14Add two missing .Ek.wiz1-1/+3
2007-11-14Properly check the return value of load_lines("PLIST.common").rillig1-2/+9
2007-11-14Some shells, for example /bin/ksh on Solaris, don't like "else fi"rillig1-2/+2
without any commands in between.
2007-11-13Provide more useful details in the error message for conflicting packages:rillig1-5/+4
- The package that cannot be installed - The pattern of the conflict - The name of the installed (conflicting) package.
2007-11-13There are several error messages with the same words. To make findingrillig1-2/+2
the cause easier, add more information to one of them.
2007-11-11Added commented-out LICENSE=something.martti5-5/+10
2007-11-08Don't use the build time version to find packages, we want to pick upjoerg2-20/+7
the newest version. Bump version to 20071109.
2007-11-08Bump version or revision of all packages that have a runtime dependencyjoerg2-4/+4
on gzip.
2007-11-08Supports DESTDIR.joerg1-1/+2
2007-11-07Mention that RUN does more error checking than _PKG_SILENT + _PKG_DEBUG.rillig1-3/+3
2007-11-07In variable assignments in the shell you can sometimes leave out therillig1-1/+9
double quotes.
2007-11-07Filemasks and Pathmasks may very well contain special characters, so therillig1-3/+3
:Q operator should not be removed from them.
2007-11-07In shell commands, handle backticks inside double quotes. By the way,rillig1-3/+15
double quotes inside backticks inside double quotes are error prone.
2007-11-07When checking a single *.mk file, determine the used variables. Thisrillig1-1/+3
avoids many false positive warnings when checking files in the mk/ directory.
2007-11-07Fixed the XXX from the last commit. PLIST conditionals are read in whenrillig1-1/+19
a package is checked and discarded for the directory and filename checks.
2007-11-07Directories may be removed in a PLIST if there is a corresponding @execrillig1-2/+9
${MKDIR}. In PLIST.common_end, directories that have been populated in PLIST.common may also be removed. XXX: The PLIST_SUBST macros that may expand to "@comment " are currently not distinguished from pathname macros. This leads to some false positive warnings.
2007-11-07Warn about removal of unrelated directories in the PLIST.rillig1-3/+16
2007-11-06The "short" names for tools may only be used in the predefinedrillig1-2/+2
{pre,do,post}-* targets, not in user-defined ones.
2007-11-05Also add -E to usage(). From PR 37265.joerg1-4/+4
2007-11-04The fifth field in a CVS/Entries file is the keyword mode. It doesn'trillig1-3/+3
have to be always empty.
2007-11-04Teach bootstrap-mk-files about groff-1.19.2 on Darwin.minskim2-4/+6
2007-11-03Reverted the deprecation of USE_NCURSES.rillig2-3/+4
XXX: To discuss: How is this variable different from PREFER_PKGSRC=ncurses? Why do we need it at all? Why isn't it documented anywhere?
2007-11-03pkg_install does the right things with the files, so CHECK_PERMS is notjoerg1-1/+2
useful and creates cyclic dependencies if pkg_install is outdated. From Peter Bex.
2007-11-03Tag pkg_install as 20071103.joerg1-2/+2
2007-11-03USE_NCURSES is deprecated. It has no effects anymore.rillig2-3/+5
2007-11-02Remove obsolete option r from usage() and man page. I can't evenjoerg3-6/+6
remember what is was for. From Yakovetsky Vlad in PR 37265.
2007-11-01Added quotes around the filenames, as suggested by joerg@.rillig1-7/+5
While here, removed an unnecessary variable.
2007-11-01When rename(2) fails, it is useful to know some details about therillig1-4/+5
filenames that have been involved. Do not use assert() for error checking.
2007-10-31"Safe" variables should not have the :Q operator, since it only makesrillig1-4/+9
the code harder to read and does not have any effect anyway.
2007-10-31Do not print warnings about missing PLIST definitions for Perl modulesrillig1-1/+2
and other packages that set GENERATE_PLIST.
2007-10-31Added a function for checking bmake conditionals.rillig1-2/+11
2007-10-31Second try to deprecate LICENCE. ACCEPTABLE_LICENCES continues to exist,rillig1-3/+3
since the reasons for keeping two variables for one purpose seem to be greater than the ones for simplicity. Don't ask _me_ why.
2007-10-27Update to version 1.30.seb3-10/+24
Approved by MAINTAINER, jmmv@. Add option -N: with the expection of pkgtools/libkver (see manual page about NETBSD_RELEASE) avoid installation of default packages as well as INSTALL_PACKAGES and BUILD_PACKAGES during the creation of the chroot. Note: This is useful for running pbulk in a sandbox created by pkg_comp as binary packages , possibly very low in the build dependencies tree like pkgtools/digest, won't be built by pkg_comp while creating the sandbox hence pbulk won't have to rebuild nearly all packages otherwise already up-to-date. The maintainer notes that this seems like abusing pkg_comp to just create sandboxes. And well... yes indeed!
2007-10-27bump PKGREVISION for archivers/pax fix (largefile support).grant1-2/+2
2007-10-25* If PLIST_SRC is explicitly set to an empty value in a package Makefile,jlam2-3/+2
then automatically generate a PLIST that says "${PKGNAME} has no files". * If PLIST_SRC and GENERATE_PLIST are not set in a package Makefile, and no PLIST files exist, then fail during the package build with PKG_FAIL_REASON. * Remove "intentionally empty" PLISTs again. Now, the easy way to say that a package installs no files is to just add the following to the package Makefile: PLIST_SRC= # empty
2007-10-25Re-add "intentionally empty" PLISTs for meta-packages and other packagesjlam1-0/+2
that directly manipulate empty PLISTs. Modify plist/plist.mk so that if the PLIST files are missing and no GENERATE_PLIST is defined, then the package fails to build.
2007-10-25Remove empty PLISTs from pkgsrc since revision 1.33 of plist/plist.mkjlam2-2/+0
can handle packages having no PLIST files.
2007-10-25Reverted the deprecation of LICENCE and ACCEPTABLE_LICENCES, since thererillig1-3/+5
is no consensus among the pkgsrc developers about them. Some want to keep them (proper British spelling), others want to eliminate them (redundancy).
2007-10-25It's not very useful to have empty PLIST files, so warn about it. Thisrillig1-1/+15
may happen when a package has been newly created and the print-PLIST target has not yet been run.
2007-10-17Make language lawyers happy and rename report_recipient to plural.joerg3-6/+6
2007-10-17LICENCE and ACCEPTABLE_LICENCES (British English) are deprecated.rillig1-3/+5
2007-10-15I just removed the $bulklog directory once again, due to runningrillig1-1/+7
bin/bulkbuild when instead I wanted to resume the bulk build. To prevent this frustration, added a warning and a five second waiting period.
2007-10-14Fixed the problem with package name matching that Klaus Heinz discoveredrillig1-5/+21
in http://mail-index.netbsd.org/tech-pkg/2007/10/14/0004.html The pattern "pkg_install" matched "pkg-install-20070828", but "pkg*all" didn't. Now it does.
2007-10-13Updated the URLs for ftp.NetBSD.org, so that they will work after therillig2-15/+15
upcoming clean-up.
2007-10-13Whew. pkglint didn't know FAIL_MSG at all.rillig1-1/+2
2007-10-13FAIL_MSG may be a hidden command.rillig1-2/+32
2007-10-12As pointed out in PR 37113, it is dangerous to set MAINTAINERrillig1-2/+2
unconditionally in Makefile.common, so warn about it.
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-28 07:29:18 +0000