pacman/lib/libalpm/deps.c

868 lines
24 KiB
C
Raw Normal View History

2005-03-14 20:51:43 -05:00
/*
* deps.c
*
* Copyright (c) 2006-2011 Pacman Development Team <pacman-dev@archlinux.org>
* Copyright (c) 2002-2006 by Judd Vinet <jvinet@zeroflux.org>
* Copyright (c) 2005 by Aurelien Foret <orelien@chez.com>
* Copyright (c) 2005, 2006 by Miklos Vajna <vmiklos@frugalware.org>
*
2005-03-14 20:51:43 -05:00
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
2005-03-14 20:51:43 -05:00
*/
#include "config.h"
2005-03-14 20:51:43 -05:00
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
/* libalpm */
#include "deps.h"
#include "alpm_list.h"
2005-03-14 20:51:43 -05:00
#include "util.h"
#include "log.h"
#include "graph.h"
2005-03-14 20:51:43 -05:00
#include "package.h"
#include "db.h"
#include "handle.h"
#include "trans.h"
/* global handle variable */
extern pmhandle_t *handle;
void _alpm_dep_free(pmdepend_t *dep)
{
FREE(dep->name);
FREE(dep->version);
FREE(dep);
}
static pmdepmissing_t *depmiss_new(const char *target, pmdepend_t *dep,
const char *causingpkg)
{
pmdepmissing_t *miss;
MALLOC(miss, sizeof(pmdepmissing_t), RET_ERR(PM_ERR_MEMORY, NULL));
STRDUP(miss->target, target, RET_ERR(PM_ERR_MEMORY, NULL));
miss->depend = _alpm_dep_dup(dep);
STRDUP(miss->causingpkg, causingpkg, RET_ERR(PM_ERR_MEMORY, NULL));
return miss;
}
void _alpm_depmiss_free(pmdepmissing_t *miss)
{
_alpm_dep_free(miss->depend);
FREE(miss->target);
FREE(miss->causingpkg);
FREE(miss);
}
/* Does pkg1 depend on pkg2, ie. does pkg2 satisfy a dependency of pkg1? */
static int _alpm_dep_edge(pmpkg_t *pkg1, pmpkg_t *pkg2)
{
alpm_list_t *i;
for(i = alpm_pkg_get_depends(pkg1); i; i = i->next) {
if(_alpm_depcmp(pkg2, i->data)) {
return 1;
}
}
return 0;
}
/* Convert a list of pmpkg_t * to a graph structure,
* with a edge for each dependency.
* Returns a list of vertices (one vertex = one package)
* (used by alpm_sortbydeps)
*/
static alpm_list_t *dep_graph_init(alpm_list_t *targets)
{
alpm_list_t *i, *j;
alpm_list_t *vertices = NULL;
/* We create the vertices */
for(i = targets; i; i = i->next) {
pmgraph_t *vertex = _alpm_graph_new();
vertex->data = (void *)i->data;
vertices = alpm_list_add(vertices, vertex);
}
/* We compute the edges */
for(i = vertices; i; i = i->next) {
pmgraph_t *vertex_i = i->data;
pmpkg_t *p_i = vertex_i->data;
/* TODO this should be somehow combined with alpm_checkdeps */
for(j = vertices; j; j = j->next) {
pmgraph_t *vertex_j = j->data;
pmpkg_t *p_j = vertex_j->data;
if(_alpm_dep_edge(p_i, p_j)) {
vertex_i->children =
alpm_list_add(vertex_i->children, vertex_j);
}
}
vertex_i->childptr = vertex_i->children;
}
return vertices;
}
2005-03-14 20:51:43 -05:00
/* Re-order a list of target packages with respect to their dependencies.
*
* Example (reverse == 0):
2005-03-14 20:51:43 -05:00
* A depends on C
* B depends on A
* Target order is A,B,C,D
*
* Should be re-ordered to C,A,B,D
*
* if reverse is > 0, the dependency order will be reversed.
*
* This function returns the new alpm_list_t* target list.
2005-03-14 20:51:43 -05:00
*
*/
alpm_list_t *_alpm_sortbydeps(alpm_list_t *targets, int reverse)
2005-03-14 20:51:43 -05:00
{
alpm_list_t *newtargs = NULL;
alpm_list_t *vertices = NULL;
alpm_list_t *vptr;
pmgraph_t *vertex;
2005-03-14 20:51:43 -05:00
if(targets == NULL) {
return NULL;
2005-03-14 20:51:43 -05:00
}
_alpm_log(PM_LOG_DEBUG, "started sorting dependencies\n");
vertices = dep_graph_init(targets);
vptr = vertices;
vertex = vertices->data;
while(vptr) {
/* mark that we touched the vertex */
vertex->state = -1;
int found = 0;
while(vertex->childptr && !found) {
pmgraph_t *nextchild = vertex->childptr->data;
vertex->childptr = vertex->childptr->next;
if(nextchild->state == 0) {
found = 1;
nextchild->parent = vertex;
vertex = nextchild;
}
else if(nextchild->state == -1) {
pmpkg_t *vertexpkg = vertex->data;
pmpkg_t *childpkg = nextchild->data;
const char *message;
_alpm_log(PM_LOG_WARNING, _("dependency cycle detected:\n"));
if(reverse) {
message =_("%s will be removed after its %s dependency\n");
} else {
message =_("%s will be installed before its %s dependency\n");
}
_alpm_log(PM_LOG_WARNING, message, vertexpkg->name, childpkg->name);
}
2005-03-14 20:51:43 -05:00
}
if(!found) {
newtargs = alpm_list_add(newtargs, vertex->data);
/* mark that we've left this vertex */
vertex->state = 1;
vertex = vertex->parent;
if(!vertex) {
vptr = vptr->next;
while(vptr) {
vertex = vptr->data;
if(vertex->state == 0) break;
vptr = vptr->next;
2005-03-14 20:51:43 -05:00
}
}
}
}
_alpm_log(PM_LOG_DEBUG, "sorting dependencies finished\n");
if(reverse) {
/* reverse the order */
alpm_list_t *tmptargs = alpm_list_reverse(newtargs);
/* free the old one */
alpm_list_free(newtargs);
newtargs = tmptargs;
}
alpm_list_free_inner(vertices, _alpm_graph_free);
alpm_list_free(vertices);
return newtargs;
2005-03-14 20:51:43 -05:00
}
static int no_dep_version(void)
{
int flags = alpm_trans_get_flags();
return flags != -1 && (flags & PM_TRANS_FLAG_NODEPVERSION);
}
static pmdepend_t *filtered_depend(pmdepend_t *dep, int nodepversion)
{
if(nodepversion) {
pmdepend_t *newdep = _alpm_dep_dup(dep);
ASSERT(newdep, return dep);
newdep->mod = PM_DEP_MOD_ANY;
dep = newdep;
}
return dep;
}
static void release_filtered_depend(pmdepend_t *dep, int nodepversion)
{
if(nodepversion) {
free(dep);
}
}
static pmpkg_t *find_dep_satisfier(alpm_list_t *pkgs, pmdepend_t *dep)
{
alpm_list_t *i;
for(i = pkgs; i; i = alpm_list_next(i)) {
pmpkg_t *pkg = i->data;
if(_alpm_depcmp(pkg, dep)) {
return pkg;
}
}
return NULL;
}
/** Find a package satisfying a specified dependency.
* The dependency can include versions with depmod operators.
* @param pkgs an alpm_list_t* of pmpkg_t where the satisfier will be searched
* @param depstring package or provision name, versioned or not
* @return a pmpkg_t* satisfying depstring
*/
pmpkg_t SYMEXPORT *alpm_find_satisfier(alpm_list_t *pkgs, const char *depstring)
{
pmdepend_t *dep = _alpm_splitdep(depstring);
pmpkg_t *pkg = find_dep_satisfier(pkgs, dep);
_alpm_dep_free(dep);
return pkg;
}
/** Checks dependencies and returns missing ones in a list.
* Dependencies can include versions with depmod operators.
* @param pkglist the list of local packages
* @param reversedeps handles the backward dependencies
* @param remove an alpm_list_t* of packages to be removed
* @param upgrade an alpm_list_t* of packages to be upgraded (remove-then-upgrade)
* @return an alpm_list_t* of pmdepmissing_t pointers.
2005-03-14 20:51:43 -05:00
*/
alpm_list_t SYMEXPORT *alpm_checkdeps(alpm_list_t *pkglist, int reversedeps,
alpm_list_t *remove, alpm_list_t *upgrade)
2005-03-14 20:51:43 -05:00
{
alpm_list_t *i, *j;
alpm_list_t *targets, *dblist = NULL, *modified = NULL;
alpm_list_t *baddeps = NULL;
int nodepversion;
2005-03-14 20:51:43 -05:00
targets = alpm_list_join(alpm_list_copy(remove), alpm_list_copy(upgrade));
for(i = pkglist; i; i = i->next) {
pmpkg_t *pkg = i->data;
if(_alpm_pkg_find(targets, pkg->name)) {
modified = alpm_list_add(modified, pkg);
} else {
dblist = alpm_list_add(dblist, pkg);
}
}
alpm_list_free(targets);
nodepversion = no_dep_version();
/* look for unsatisfied dependencies of the upgrade list */
for(i = upgrade; i; i = i->next) {
pmpkg_t *tp = i->data;
_alpm_log(PM_LOG_DEBUG, "checkdeps: package %s-%s\n",
alpm_pkg_get_name(tp), alpm_pkg_get_version(tp));
for(j = alpm_pkg_get_depends(tp); j; j = j->next) {
pmdepend_t *depend = j->data;
depend = filtered_depend(depend, nodepversion);
/* 1. we check the upgrade list */
/* 2. we check database for untouched satisfying packages */
if(!find_dep_satisfier(upgrade, depend) &&
!find_dep_satisfier(dblist, depend)) {
/* Unsatisfied dependency in the upgrade list */
pmdepmissing_t *miss;
char *missdepstring = alpm_dep_compute_string(depend);
_alpm_log(PM_LOG_DEBUG, "checkdeps: missing dependency '%s' for package '%s'\n",
missdepstring, alpm_pkg_get_name(tp));
free(missdepstring);
miss = depmiss_new(alpm_pkg_get_name(tp), depend, NULL);
baddeps = alpm_list_add(baddeps, miss);
}
release_filtered_depend(depend, nodepversion);
}
}
if(reversedeps) {
/* reversedeps handles the backwards dependencies, ie,
* the packages listed in the requiredby field. */
for(i = dblist; i; i = i->next) {
pmpkg_t *lp = i->data;
for(j = alpm_pkg_get_depends(lp); j; j = j->next) {
pmdepend_t *depend = j->data;
depend = filtered_depend(depend, nodepversion);
pmpkg_t *causingpkg = find_dep_satisfier(modified, depend);
/* we won't break this depend, if it is already broken, we ignore it */
/* 1. check upgrade list for satisfiers */
/* 2. check dblist for satisfiers */
if(causingpkg &&
!find_dep_satisfier(upgrade, depend) &&
!find_dep_satisfier(dblist, depend)) {
pmdepmissing_t *miss;
char *missdepstring = alpm_dep_compute_string(depend);
_alpm_log(PM_LOG_DEBUG, "checkdeps: transaction would break '%s' dependency of '%s'\n",
missdepstring, alpm_pkg_get_name(lp));
free(missdepstring);
miss = depmiss_new(lp->name, depend, alpm_pkg_get_name(causingpkg));
baddeps = alpm_list_add(baddeps, miss);
2005-03-14 20:51:43 -05:00
}
release_filtered_depend(depend, nodepversion);
2005-03-14 20:51:43 -05:00
}
}
}
alpm_list_free(modified);
alpm_list_free(dblist);
2005-03-14 20:51:43 -05:00
return baddeps;
2005-03-14 20:51:43 -05:00
}
static int dep_vercmp(const char *version1, pmdepmod_t mod,
const char *version2)
{
int equal = 0;
if(mod == PM_DEP_MOD_ANY) {
equal = 1;
} else {
int cmp = alpm_pkg_vercmp(version1, version2);
switch(mod) {
case PM_DEP_MOD_EQ: equal = (cmp == 0); break;
case PM_DEP_MOD_GE: equal = (cmp >= 0); break;
case PM_DEP_MOD_LE: equal = (cmp <= 0); break;
case PM_DEP_MOD_LT: equal = (cmp < 0); break;
case PM_DEP_MOD_GT: equal = (cmp > 0); break;
default: equal = 1; break;
}
}
return equal;
}
int _alpm_depcmp(pmpkg_t *pkg, pmdepend_t *dep)
{
alpm_list_t *i;
int satisfy = 0;
/* check (pkg->name, pkg->version) */
if(pkg->name_hash && dep->name_hash
&& pkg->name_hash != dep->name_hash) {
/* skip more expensive checks */
} else {
satisfy = (strcmp(pkg->name, dep->name) == 0
&& dep_vercmp(pkg->version, dep->mod, dep->version));
if(satisfy) {
return satisfy;
}
}
/* check provisions, format : "name=version" */
for(i = alpm_pkg_get_provides(pkg); i && !satisfy; i = i->next) {
const char *provision = i->data;
const char *provver = strchr(provision, '=');
if(provver == NULL) { /* no provision version */
satisfy = (dep->mod == PM_DEP_MOD_ANY
&& strcmp(provision, dep->name) == 0);
} else {
/* This is a bit tricker than the old code for performance reasons. To
* prevent the need to copy and duplicate strings, strncmp only the name
* portion if they are the same length, since there is a version and
* operator in play here. Cast is to silence sign conversion warning;
* we know provver >= provision if we are here. */
size_t namelen = (size_t)(provver - provision);
provver += 1;
satisfy = (strlen(dep->name) == namelen
&& strncmp(provision, dep->name, namelen) == 0
&& dep_vercmp(provver, dep->mod, dep->version));
}
}
return satisfy;
}
pmdepend_t *_alpm_splitdep(const char *depstring)
2005-03-14 20:51:43 -05:00
{
pmdepend_t *depend;
const char *ptr, *version = NULL;
2005-03-14 20:51:43 -05:00
if(depstring == NULL) {
return NULL;
}
CALLOC(depend, 1, sizeof(pmdepend_t), RET_ERR(PM_ERR_MEMORY, NULL));
2005-03-14 20:51:43 -05:00
/* Find a version comparator if one exists. If it does, set the type and
* increment the ptr accordingly so we can copy the right strings. */
if((ptr = strstr(depstring, ">="))) {
depend->mod = PM_DEP_MOD_GE;
version = ptr + 2;
} else if((ptr = strstr(depstring, "<="))) {
depend->mod = PM_DEP_MOD_LE;
version = ptr + 2;
} else if((ptr = strstr(depstring, "="))) {
/* Note: we must do =,<,> checks after <=, >= checks */
depend->mod = PM_DEP_MOD_EQ;
version = ptr + 1;
} else if((ptr = strstr(depstring, "<"))) {
depend->mod = PM_DEP_MOD_LT;
version = ptr + 1;
} else if((ptr = strstr(depstring, ">"))) {
depend->mod = PM_DEP_MOD_GT;
version = ptr + 1;
2005-03-14 20:51:43 -05:00
} else {
/* no version specified, leave version and ptr NULL */
depend->mod = PM_DEP_MOD_ANY;
2005-03-14 20:51:43 -05:00
}
/* copy the right parts to the right places */
STRNDUP(depend->name, depstring, ptr - depstring,
RET_ERR(PM_ERR_MEMORY, NULL));
depend->name_hash = _alpm_hash_sdbm(depend->name);
if(version) {
STRDUP(depend->version, version, RET_ERR(PM_ERR_MEMORY, NULL));
}
2005-03-14 20:51:43 -05:00
return depend;
2005-03-14 20:51:43 -05:00
}
pmdepend_t *_alpm_dep_dup(const pmdepend_t *dep)
{
pmdepend_t *newdep;
CALLOC(newdep, 1, sizeof(pmdepend_t), RET_ERR(PM_ERR_MEMORY, NULL));
STRDUP(newdep->name, dep->name, RET_ERR(PM_ERR_MEMORY, NULL));
newdep->name_hash = dep->name_hash;
STRDUP(newdep->version, dep->version, RET_ERR(PM_ERR_MEMORY, NULL));
newdep->mod = dep->mod;
return newdep;
}
/* These parameters are messy. We check if this package, given a list of
* targets and a db is safe to remove. We do NOT remove it if it is in the
* target list, or if if the package was explictly installed and
* include_explicit == 0 */
static int can_remove_package(pmdb_t *db, pmpkg_t *pkg, alpm_list_t *targets,
int include_explicit)
{
alpm_list_t *i;
if(_alpm_pkg_find(targets, alpm_pkg_get_name(pkg))) {
return 0;
}
if(!include_explicit) {
/* see if it was explicitly installed */
if(alpm_pkg_get_reason(pkg) == PM_PKG_REASON_EXPLICIT) {
_alpm_log(PM_LOG_DEBUG, "excluding %s -- explicitly installed\n",
alpm_pkg_get_name(pkg));
return 0;
}
}
/* TODO: checkdeps could be used here, it handles multiple providers
* better, but that also makes it slower.
* Also this would require to first add the package to the targets list,
* then call checkdeps with it, then remove the package from the targets list
* if checkdeps detected it would break something */
/* see if other packages need it */
for(i = _alpm_db_get_pkgcache(db); i; i = i->next) {
pmpkg_t *lpkg = i->data;
if(_alpm_dep_edge(lpkg, pkg) && !_alpm_pkg_find(targets, lpkg->name)) {
return 0;
}
}
/* it's ok to remove */
return 1;
}
/**
* @brief Adds unneeded dependencies to an existing list of packages.
* By unneeded, we mean dependencies that are only required by packages in the
* target list, so they can be safely removed.
* If the input list was topo sorted, the output list will be topo sorted too.
*
* @param db package database to do dependency tracing in
* @param *targs pointer to a list of packages
* @param include_explicit if 0, explicitly installed packages are not included
2005-03-14 20:51:43 -05:00
*/
void _alpm_recursedeps(pmdb_t *db, alpm_list_t *targs, int include_explicit)
2005-03-14 20:51:43 -05:00
{
alpm_list_t *i, *j;
2005-03-14 20:51:43 -05:00
if(db == NULL || targs == NULL) {
return;
2005-03-14 20:51:43 -05:00
}
for(i = targs; i; i = i->next) {
pmpkg_t *pkg = i->data;
for(j = _alpm_db_get_pkgcache(db); j; j = j->next) {
pmpkg_t *deppkg = j->data;
if(_alpm_dep_edge(pkg, deppkg)
&& can_remove_package(db, deppkg, targs, include_explicit)) {
_alpm_log(PM_LOG_DEBUG, "adding '%s' to the targets\n",
alpm_pkg_get_name(deppkg));
/* add it to the target list */
targs = alpm_list_add(targs, _alpm_pkg_dup(deppkg));
2005-03-14 20:51:43 -05:00
}
}
}
}
/**
* helper function for resolvedeps: search for dep satisfier in dbs
*
* @param dep is the dependency to search for
* @param dbs are the databases to search
* @param excluding are the packages to exclude from the search
* @param prompt if true, will cause an unresolvable dependency to issue an
* interactive prompt asking whether the package should be removed from
* the transaction or the transaction aborted; if false, simply returns
* an error code without prompting
* @return the resolved package
**/
static pmpkg_t *resolvedep(pmdepend_t *dep, alpm_list_t *dbs,
alpm_list_t *excluding, int prompt)
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
{
alpm_list_t *i, *j;
int ignored = 0;
alpm_list_t *providers = NULL;
int count;
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
/* 1. literals */
for(i = dbs; i; i = i->next) {
pmpkg_t *pkg = _alpm_db_get_pkgfromcache(i->data, dep->name);
if(pkg && _alpm_depcmp(pkg, dep) && !_alpm_pkg_find(excluding, pkg->name)) {
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
if(_alpm_pkg_should_ignore(pkg)) {
int install = 0;
if(prompt) {
QUESTION(handle->trans, PM_TRANS_CONV_INSTALL_IGNOREPKG, pkg,
NULL, NULL, &install);
} else {
_alpm_log(PM_LOG_WARNING, _("ignoring package %s-%s\n"), pkg->name, pkg->version);
}
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
if(!install) {
ignored = 1;
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
continue;
}
}
return pkg;
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
}
}
/* 2. satisfiers (skip literals here) */
for(i = dbs; i; i = i->next) {
for(j = _alpm_db_get_pkgcache(i->data); j; j = j->next) {
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
pmpkg_t *pkg = j->data;
if(_alpm_depcmp(pkg, dep) && strcmp(pkg->name, dep->name) != 0 &&
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
!_alpm_pkg_find(excluding, pkg->name)) {
if(_alpm_pkg_should_ignore(pkg)) {
int install = 0;
if(prompt) {
QUESTION(handle->trans, PM_TRANS_CONV_INSTALL_IGNOREPKG,
pkg, NULL, NULL, &install);
} else {
_alpm_log(PM_LOG_WARNING, _("ignoring package %s-%s\n"), pkg->name, pkg->version);
}
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
if(!install) {
ignored = 1;
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
continue;
}
}
_alpm_log(PM_LOG_DEBUG, "provider found (%s provides %s)\n",
pkg->name, dep->name);
providers = alpm_list_add(providers, pkg);
/* keep looking for other providers in the all dbs */
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
}
}
}
/* first check if one provider is already installed locally */
for(i = providers; i; i = i->next) {
pmpkg_t *pkg = i->data;
if(_alpm_pkghash_find(_alpm_db_get_pkgcache_hash(handle->db_local), pkg->name)) {
alpm_list_free(providers);
return pkg;
}
}
count = alpm_list_count(providers);
if(count >= 1) {
/* default to first provider if there is no QUESTION callback */
int index = 0;
if(count > 1) {
/* if there is more than one provider, we ask the user */
QUESTION(handle->trans, PM_TRANS_CONV_SELECT_PROVIDER,
providers, dep, NULL, &index);
}
if(index >= 0 && index < count) {
pmpkg_t *pkg = alpm_list_getdata(alpm_list_nth(providers, index));
alpm_list_free(providers);
return pkg;
}
alpm_list_free(providers);
providers = NULL;
}
if(ignored) { /* resolvedeps will override these */
pm_errno = PM_ERR_PKG_IGNORED;
} else {
pm_errno = PM_ERR_PKG_NOT_FOUND;
}
return NULL;
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
}
/** Find a package satisfying a specified dependency.
* First look for a literal, going through each db one by one. Then look for
* providers. The first satisfier found is returned.
* The dependency can include versions with depmod operators.
* @param dbs an alpm_list_t* of pmdb_t where the satisfier will be searched
* @param depstring package or provision name, versioned or not
* @return a pmpkg_t* satisfying depstring
*/
pmpkg_t SYMEXPORT *alpm_find_dbs_satisfier(alpm_list_t *dbs, const char *depstring)
{
pmdepend_t *dep;
pmpkg_t *pkg;
ASSERT(dbs, return NULL);
dep = _alpm_splitdep(depstring);
ASSERT(dep, return NULL);
pkg = resolvedep(dep, dbs, NULL, 1);
_alpm_dep_free(dep);
return pkg;
}
/**
* Computes resolvable dependencies for a given package and adds that package
* and those resolvable dependencies to a list.
2005-03-14 20:51:43 -05:00
*
* @param localpkgs is the list of local packages
* @param dbs_sync are the sync databases
* @param pkg is the package to resolve
* @param packages is a pointer to a list of packages which will be
* searched first for any dependency packages needed to complete the
* resolve, and to which will be added any [pkg] and all of its
* dependencies not already on the list
* @param remove is the set of packages which will be removed in this
* transaction
* @param data returns the dependency which could not be satisfied in the
* event of an error
* @return 0 on success, with [pkg] and all of its dependencies not already on
* the [*packages] list added to that list, or -1 on failure due to an
* unresolvable dependency, in which case the [*packages] list will be
* unmodified by this function
2005-03-14 20:51:43 -05:00
*/
int _alpm_resolvedeps(alpm_list_t *localpkgs, alpm_list_t *dbs_sync, pmpkg_t *pkg,
alpm_list_t *preferred, alpm_list_t **packages,
alpm_list_t *remove, alpm_list_t **data)
2005-03-14 20:51:43 -05:00
{
int ret = 0;
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
alpm_list_t *i, *j;
alpm_list_t *targ;
alpm_list_t *deps = NULL;
alpm_list_t *packages_copy;
2005-03-14 20:51:43 -05:00
if(_alpm_pkg_find(*packages, pkg->name) != NULL) {
return 0;
}
/* Create a copy of the packages list, so that it can be restored
on error */
packages_copy = alpm_list_copy(*packages);
/* [pkg] has not already been resolved into the packages list, so put it
on that list */
*packages = alpm_list_add(*packages, pkg);
_alpm_log(PM_LOG_DEBUG, "started resolving dependencies\n");
for(i = alpm_list_last(*packages); i; i = i->next) {
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
pmpkg_t *tpkg = i->data;
targ = alpm_list_add(NULL, tpkg);
deps = alpm_checkdeps(localpkgs, 0, remove, targ);
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
alpm_list_free(targ);
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
for(j = deps; j; j = j->next) {
pmdepmissing_t *miss = j->data;
pmdepend_t *missdep = alpm_miss_get_dep(miss);
/* check if one of the packages in the [*packages] list already satisfies
* this dependency */
if(find_dep_satisfier(*packages, missdep)) {
_alpm_depmiss_free(miss);
continue;
}
/* check if one of the packages in the [preferred] list already satisfies
* this dependency */
pmpkg_t *spkg = find_dep_satisfier(preferred, missdep);
if(!spkg) {
/* find a satisfier package in the given repositories */
spkg = resolvedep(missdep, dbs_sync, *packages, 0);
}
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
if(!spkg) {
pm_errno = PM_ERR_UNSATISFIED_DEPS;
char *missdepstring = alpm_dep_compute_string(missdep);
_alpm_log(PM_LOG_WARNING,
_("cannot resolve \"%s\", a dependency of \"%s\"\n"),
missdepstring, tpkg->name);
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
free(missdepstring);
if(data) {
*data = alpm_list_add(*data, miss);
}
ret = -1;
} else {
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
_alpm_log(PM_LOG_DEBUG, "pulling dependency %s (needed by %s)\n",
alpm_pkg_get_name(spkg), alpm_pkg_get_name(tpkg));
*packages = alpm_list_add(*packages, spkg);
_alpm_depmiss_free(miss);
2005-03-14 20:51:43 -05:00
}
}
Resolvedeps rework I divided resolvedeps into 2 functions. The new _alpm_resolvedep function will resolve one dependency, for example the 'foo>=1.0-1' dependency. It can be useful in sync_addtarget refactoring. The resolvedeps parameters were changed, to be coherent with recursedeps: * the target-list is an alpm_list* instead of alpm_list**. This is OK, because alpm_list_add == alpm_list_add_last * syncpkg param was removed. list contains the to-be-installed packages, resolvedeps will add the required dependencies into this list * trans param was removed, it was used in QUESTION() only, which can be used on the main (handle->trans) transaction only (because the front-end cannot access our pseudo-transactions at all!). The patch fixes some wrong dynamic pmdepmissing_t usage. I did a behavior change (and sync1003.py was modified accordingly), which needs some explanation: The old resolvedeps didn't elect packages from 'remove' list. I've dropped this because I don't want that 2nd excluding list param. In fact, in real life, we ~never need this rule. Resolvedeps is called before checkconflicts, so only -Su's %REPLACES% packages are sitting in 'remove' list. This means, that we have the replacement packages in our target list. Usually "foo replaces bar" means, that bar isn't in our repos any more, so resolvedeps *cannot* elect it; but usually it won't try it at all, because foo is in the target list, and it is expected to satisfy 'bar>=1.0-1'-like dependencies too. Since checkdeps and checkconflicts is done after resolvedeps, this cannot cause any harm. Signed-off-by: Nagy Gabor <ngaba@bibl.u-szeged.hu> Signed-off-by: Dan McGee <dan@archlinux.org>
2008-07-05 19:22:13 -04:00
alpm_list_free(deps);
2005-03-14 20:51:43 -05:00
}
if(ret != 0) {
alpm_list_free(*packages);
*packages = packages_copy;
} else {
alpm_list_free(packages_copy);
}
_alpm_log(PM_LOG_DEBUG, "finished resolving dependencies\n");
return ret;
2005-03-14 20:51:43 -05:00
}
const char SYMEXPORT *alpm_miss_get_target(const pmdepmissing_t *miss)
{
/* Sanity checks */
ASSERT(miss != NULL, return NULL);
return miss->target;
}
const char SYMEXPORT *alpm_miss_get_causingpkg(const pmdepmissing_t *miss)
{
/* Sanity checks */
ASSERT(miss != NULL, return NULL);
return miss->causingpkg;
}
pmdepend_t SYMEXPORT *alpm_miss_get_dep(pmdepmissing_t *miss)
{
/* Sanity checks */
ASSERT(miss != NULL, return NULL);
return miss->depend;
}
pmdepmod_t SYMEXPORT alpm_dep_get_mod(const pmdepend_t *dep)
{
/* Sanity checks */
ASSERT(dep != NULL, return -1);
return dep->mod;
}
const char SYMEXPORT *alpm_dep_get_name(const pmdepend_t *dep)
{
/* Sanity checks */
ASSERT(dep != NULL, return NULL);
return dep->name;
}
const char SYMEXPORT *alpm_dep_get_version(const pmdepend_t *dep)
{
/* Sanity checks */
ASSERT(dep != NULL, return NULL);
return dep->version;
}
/** Reverse of splitdep; make a dep string from a pmdepend_t struct.
* The string must be freed!
* @param dep the depend to turn into a string
* @return a string-formatted dependency with operator if necessary
*/
char SYMEXPORT *alpm_dep_compute_string(const pmdepend_t *dep)
{
const char *name, *opr, *ver;
char *str;
size_t len;
/* Sanity checks */
ASSERT(dep != NULL, return NULL);
if(dep->name) {
name = dep->name;
} else {
name = "";
}
switch(dep->mod) {
case PM_DEP_MOD_ANY:
opr = "";
break;
case PM_DEP_MOD_GE:
opr = ">=";
break;
case PM_DEP_MOD_LE:
opr = "<=";
break;
case PM_DEP_MOD_EQ:
opr = "=";
break;
case PM_DEP_MOD_LT:
opr = "<";
break;
case PM_DEP_MOD_GT:
opr = ">";
break;
default:
opr = "";
break;
}
if(dep->mod != PM_DEP_MOD_ANY && dep->version) {
ver = dep->version;
} else {
ver = "";
}
/* we can always compute len and print the string like this because opr
* and ver will be empty when PM_DEP_MOD_ANY is the depend type. the
* reassignments above also ensure we do not do a strlen(NULL). */
len = strlen(name) + strlen(opr) + strlen(ver) + 1;
MALLOC(str, len, RET_ERR(PM_ERR_MEMORY, NULL));
snprintf(str, len, "%s%s%s", name, opr, ver);
return str;
}
2005-03-14 20:51:43 -05:00
/* vim: set ts=2 sw=2 noet: */