summaryrefslogtreecommitdiffstats
path: root/usr.sbin/pkg_add/OpenBSD (follow)
Commit message (Collapse)AuthorAgeFilesLines
* tweaks to system version handling:espie2021-03-151-5/+6
| | | | | | | | | - create an element that accumulates version values for when we have several - actually use compare on version values instead of hardcoding the difference no functional change
* finish GC old unintuitive interfaceespie2021-03-021-2/+2
| | | | now handle_continue and find_window_size are fully separated.
* in case we're not a tty, don't do anything elseespie2021-02-011-3/+6
| | | | this does fix the grep case
* turns out STDOUT may have been redirected, in which case tcgetpgrp willespie2021-01-311-2/+5
| | | | | | return -1, in which case we never need to suppress output. noticed by Mark Patruck
* do not print to STDOUT if we're in background, as requested by Theoespie2021-01-303-16/+36
| | | | who pointed at ssh code for process group handling. Thanks
* I'm such a doofus, of course I have to call base method as wellespie2021-01-301-1/+2
|
* switch ProgressMeter to clearer APIespie2021-01-303-9/+18
|
* rework API slightly, so that handle_continue is an explicit methodespie2021-01-302-3/+14
| | | | (to be overrideen by subclasses when needed)
* for now, do not try to install quirks in case we're running "not"espie2021-01-301-2/+2
| | | | | the better thing to do would be to pseudo-install it under /tmp so we can run it, but it requires way more changes
* remove test that's now bogus, as fullpkgpath takes precedence.espie2021-01-261-10/+6
| | | | thanks bluhm@
* - grab fullpkgpath early (to decide: hard error if not provided) ?espie2021-01-111-6/+16
| | | | | | | | | | | | | | | - tweak the look-up mechanism for dependencies to NOT look at installed packages for inter-dependencies between subpackages (and debug packages) of the SAME pkgpath. The rationale for looking at installed packages in the usual case is that it's normally much faster rather than have the ports tree spew out packing-lists. In the case of multi-packages updates, the new version may have been built successfully (and correctly) while the old version is still around. However, if pkg_create looks at the installed version, it may fail to create the package because the shared library versions will be different.
* avoid extra warnings if undefined values are encounteredespie2021-01-081-2/+4
|
* rename pkg_add's pod files to avoid filenames with colonsdaniel2020-12-2016-0/+2147
| | | | | | | | | Filenames with colons are not allowed on windows filesystems. So we rename the pod files to avoid this. Should get us closer to being able to check out the openbsd source code on a windows system. ok espie@ (who suggested the current approach); sthen@ was ok with a previous version of the diff
* speed up the dependency check pass significantly:espie2020-11-091-2/+22
| | | | | | | | | first pass will read packing-lists fully, use that to cache along dependency information, so we don't have to read packing-list again during second pass. we do not cache full plists for the file checks, because 1/ that is significantly larger 2/ that pass is disk io bound anyway
* fix indirect dependencies bug: match what pkg_add does by adding tagespie2020-11-091-10/+40
| | | | | | | | | | | | definitions to the list of RequiredBy. requires passing the pkgname around for the special case where tag and define-tag are in the same location, so use that info for better diagnostics Note that this loses the optimization where we only need read DependsOnly this can be addressed later by storing stub packing-lists with only relevant info during the first pass, so that we don't have to read them again
* reword confusing messageespie2020-11-091-2/+2
|
* More DESC -> DESCR; spotted by espie@ajacoutot2020-07-251-3/+3
|
* tyop: DESC -> DESCRajacoutot2020-07-251-2/+2
| | | | ok landry@ robert@
* Show a message while pkg_add updates the font cache. This occurs at thesthen2020-06-091-1/+3
| | | | | | end of an install and can take quite a while if you have certain fonts installed. Before this change the cursor was left at an empty line. tweak/ok espie@
* in order for PKG_CACHE to work, skipped entries at the end should stillespie2020-05-171-2/+8
| | | | | | be read. This happens for packages ending in symlinks for instance. this should fix some fringe cases where FETCH_PACKAGES doesn't work
* Add support for "alpha" suffixes in package versionsjca2020-03-261-3/+6
| | | | | | | | | | | | | We already supported "beta" releases but not "alpha", probably to discourage people from pushing "alpha-quality" software in the ports tree. We have 30 ports with distfiles matching "alpha", and people just work around the lack of explicit support for "alpha" releases with mixes of crafted PKGNAME and REVISION/EPOCH. This looks needlessly complicated and a bit unfair when compared to GH_COMMIT support. Pointers from landry@ and sthen@, lgtm landry@, no objection from espie@
* forgot to actually put "nodebug" in stateespie2020-02-201-2/+5
|
* make shell "with debug info" by default and add a bit of code to markespie2020-02-201-5/+10
| | | | stuff "no debug" eventually
* refactor opt('Q') into its own codeespie2020-02-191-19/+32
| | | | | | make sure we see -a along with opt('Q') tweak -aQ to use keep_all so that it shows every entry in every repository instead of stopping at the first one.
* allow search objects to say "keep_all" and make PackageRepositoryListsespie2020-02-192-4/+14
| | | | aware of that for match_locations
* Separate singular and pluralespie2020-02-191-4/+9
|
* future replacement for PackageInfo.pmespie2020-02-171-0/+222
| | | | | | | | | | - the name was a bit confusing (pkg_info) - one of the first modules I wrote, and the old one isn't OO at all, and full of globals, which can be a problem for displaying error messages, and for tweaking PKG_DB locally... This will replace PackageInfo.pm once I've gone through all the callers. New synopsis would make this part of State in most cases, without any singleton.
* add iwx to fw_updatestsp2020-02-111-2/+2
|
* banish implicit returnespie2020-01-281-2/+2
|
* match opendir with closedir... not that important as the thing was mostespie2020-01-281-2/+2
| | | | probably GC'd.
* give pkg_create knowledge of WRKOBJDIR, so it can error out if a symlinkespie2020-01-261-1/+8
| | | | | | points in there. okay aja@ sthen@
* nit, use "fatal" with parameters as intendedespie2020-01-131-2/+2
|
* debug-> pkg-debug definesespie2020-01-112-4/+4
| | | | | It's only used for debugging pkgtools, it shouldn't interfere with possible flavors in pkg_create
* have ask_tree be more specific in its error reporting.espie2020-01-091-6/+15
| | | | | | | | | make sure the child has the same effective and real ids, as perl interprets it as "setuid, refuse to debug" go to the end of the pipe so the child doesn't exit prematurely (didn't use to happen but now it does) help from afresh1@ in diagnosing that one.
* plain forgot about THAT optimization.espie2020-01-021-2/+2
| | | | | | Let .so/.a print again in pkg_info -L noticed by Mikolaj Kucharski
* partial backout of 1.166: we still need to compute the pkgname without theespie2019-12-291-7/+10
| | | | | | repo for debug packages, but we need the full path for the handle (for now) fixes regression noticed by bluhm@
* close the gap for the "final" hint that did not install debug packagesespie2019-12-081-6/+9
| | | | | | | | | (coincidentally, that's the pattern used for install from ports) This just requires always passing pkgname thru parse to extract the right name. There's definitely a bit of code to clean up in pkg_add afterwards :/
* have DEBUG_PKG_CACHE also apply to already installed packages, toespie2019-12-081-19/+24
| | | | | | | | help bootstrap the process Clean the code: - move the decision logic into individual may_grab_debug_for - acknowledge the PkgAdd object doesn't matter in there, just use class - for already installed stuff, we skip packages that already exist
* add DEBUG_PKG_CACHE functionality. Doesn't interfere with normal operation.espie2019-12-041-1/+61
| | | | | | | | | | Basically, if you set DEBUG_PKG_CACHE to a directory, pkg_add will download debug packages "on the sly" any time it installs packages. It is mostly silent (apart from fetching error or directory non existent). It *won't* download packages that are already installed, but this allows keeping debug packages "just in case" in synch with your current snapshot, as some people (mpi@) have worried that by the time you get to debugging something, the snapshots will have moved too much.
* allow error to be passed along, in case we need to save $? for laterespie2019-12-041-3/+3
|
* preliminary addition of -d option (semantics: add debug packages ifespie2019-11-162-8/+33
| | | | | | they're present along the packages we want to add/update) probably not perfect yet, but 100% a no-op if it's not used, so won't break existing usage.
* nicer order of multi-choice packages (by version if everything else isespie2019-11-151-2/+14
| | | | | | equal, alphabetically otherwise) prompted by claudio@, okay claudio@ kmos@
* tweaks to the file hierarchy to annotate anything with debuginfo, soespie2019-11-101-8/+23
| | | | | | | existing @bin @lib and new @static-lib @so as discussed with p2k19 people
* typoespie2019-11-101-2/+2
|
* added too much for -stable, if ftp says 404, we still need to save thatespie2019-11-081-3/+2
|
* when asking the ports tree, we fork a child.espie2019-11-071-2/+16
| | | | | | | | | instead of ditching stderr entirely, keep it in a temp file, and if the child exits with an error, we got something to display. (note that returning and undef'd plist is enough of an error, just we had no clue what went wrong previously) aja@ ran into this a few weeks ago.
* style nitespie2019-11-071-2/+2
|
* Add ogx to fw_update.visa2019-11-051-2/+2
|
* adapt to new ftp(1) retrieving error messagesderaadt2019-10-231-2/+2
|
* I didn't check robert's fix closely enough.espie2019-10-201-19/+28
| | | | | | fix it so commands are run when several fontdirs are involved also streamline the alias code a little bit, and have it report actual issues in opening files
Copyright © 1996 – 2023 Jason A. Donenfeld. All Rights Reverse Engineered.