summaryrefslogtreecommitdiffstats
path: root/urpm.pm
Commit message (Collapse)AuthorAgeFilesLines
* move create_transaction() and prepare_transaction() from urpm.pm into ↵Pascal Rigaux2006-11-211-38/+0
| | | | urpm/install.pm
* move urpm::install() into urpm::install::install() (new module)Pascal Rigaux2006-11-211-145/+2
|
* create urpm::md5sum, move md5sum functions into it: Pascal Rigaux2006-11-211-47/+8
| | | | | | - urpm::util::md5sum() -> urpm::md5sum::compute() - urpm::local_md5sum() -> urpm::md5sum::on_local_medium()
* move open_safe, opendir_safe and syserror to urpm::sysPascal Rigaux2006-11-211-121/+28
|
* - rename get_md5sum() into from_MD5SUM()Pascal Rigaux2006-11-211-9/+7
| | | | | | - rename parse_md5sum() into from_MD5SUM__or_warn() - in from_MD5SUM__or_warn(), display the name of the MD5SUM file used
* when computing the md5sum of an hdlist, display the file name in the logPascal Rigaux2006-11-211-1/+1
|
* fix an optimisation i introduced which breaks generated /var/lib/urpmi/MD5SUMPascal Rigaux2006-11-211-1/+3
|
* 20 is the minimal valid size for an empty synthesis (which should be valid)Pascal Rigaux2006-11-211-7/+7
|
* move lock functions to urpm::sysPascal Rigaux2006-11-211-58/+9
|
* move urpm::parallel_remove to urpm::parallel::removePascal Rigaux2006-11-211-11/+2
| | | | | (but keep urpm::parallel_remove for backward compatibility since rpmdrake uses it)
* drop urpm::parallel_install(), directly use {parallel_handler}->parallel_installPascal Rigaux2006-11-211-7/+0
|
* create urpm::parallel and move some things from urpm.pmPascal Rigaux2006-11-211-45/+3
|
* cleanupPascal Rigaux2006-11-211-5/+3
|
* create _parse_media() out of configure()Pascal Rigaux2006-11-211-53/+56
|
* create _compute_flags_for_skiplist() and _compute_flags_for_instlist() out ↵Pascal Rigaux2006-11-211-25/+35
| | | | of configure()
* only do generate_medium_names() on really_modified (ie updated) mediaPascal Rigaux2006-11-211-17/+12
|
* introduce $medium->{really_modified} which is what $medium->{modified} shouldPascal Rigaux2006-11-211-4/+6
| | | | | be... but {modified} is over-used (eg: it often means "to update")
* use function names more appropriatePascal Rigaux2006-11-211-4/+4
|
* adapt function names to pubkey and descriptions being done laterPascal Rigaux2006-11-211-5/+5
|
* handle pubkey and descriptions after the whole hdlist/synthesis/list mess is ↵Pascal Rigaux2006-11-211-5/+6
| | | | settled down
* restore removing {list} (it was done in _write_rpm_list_if_needed)Pascal Rigaux2006-11-211-16/+11
|
* set $medium->{list} only when neededPascal Rigaux2006-11-211-5/+2
|
* create _get_pubkey_and_descriptions() out of various code that was done in thePascal Rigaux2006-11-211-14/+15
| | | | | middle of code parsing hdlist/synthesis
* don't pass rpm_files around, use $medium->{rpm_files}Pascal Rigaux2006-11-211-16/+10
|
* don't pass around $retrieved_md5sum, set $medium->{md5sum} as soon as wePascal Rigaux2006-11-211-8/+17
| | | | | | | | checked against {md5sum}. but remove it if parsing fail later on. also restore parsing current synthesis when new hdlist/synthesis is buggy. (it was in _update_medium_first_pass_failed())
* create _read_rpms_from_dir() out of ↵Pascal Rigaux2006-11-211-46/+56
| | | | _update_medium__parse_if_unmodified__or_get_files__local
* no {with_hdlist} is valid (it will end up setting {force_building_hdlist} ↵Pascal Rigaux2006-11-211-1/+1
| | | | and using *.rpm)
* remove hopefully now non useful reparsing of synthesis on errorPascal Rigaux2006-11-211-5/+0
| | | | | (since URPM 1.48, parse_synthesis doesn't modify {depslist} if it doesn't succeed)
* create get_hdlist_or_synthesis_and_check_md5sum__local() out ofPascal Rigaux2006-11-211-26/+29
| | | | | | _update_medium__parse_if_unmodified__or_get_files__local() and so make things clearer
* rename recompute_local_md5sum() to local_md5sum()Pascal Rigaux2006-11-211-3/+3
|
* simplifyPascal Rigaux2006-11-211-8/+5
|
* handle error by returningPascal Rigaux2006-11-211-15/+14
|
* _update_medium_second_pass: move what is not $second_pass dependent in ↵Pascal Rigaux2006-11-211-6/+14
| | | | _update_medium_build_hdlist_synthesis()
* separate what is dependent on $second_pass and what is notPascal Rigaux2006-11-211-12/+14
|
* use _parse_maybe_hdlist_or_synthesis()Pascal Rigaux2006-11-211-5/+1
|
* revert buggy changePascal Rigaux2006-11-211-2/+5
|
* use _parse_maybe_hdlist_or_synthesis()Pascal Rigaux2006-11-211-5/+2
|
* fully clean {cachedir}/partial/ on errorPascal Rigaux2006-11-211-2/+1
|
* handle error by returning, it allows removing many unless ($error) ...Pascal Rigaux2006-11-211-30/+29
|
* factorize code in _parse_maybe_hdlist_or_synthesis()Pascal Rigaux2006-11-211-27/+28
|
* don't blindly probe for named hdlist. This feature is only available throughPascal Rigaux2006-11-201-6/+3
| | | | | media.cfg which gives things precisely without having to probe in the dark
* cleanup (at last!) the f*cking $with_hdlist_dir, use the better _hdlist_dir(),Pascal Rigaux2006-11-201-23/+26
| | | | | which really returns a dir, with reduce_pathname already called.
* $with_hdlist_dir really is something simple, so simplify ↵Pascal Rigaux2006-11-201-7/+7
| | | | _parse_hdlist_or_synthesis__virtual()
* {with_dir} is more important than {probe_with}Pascal Rigaux2006-11-201-8/+3
|
* {probe_with} and {with_hdlist} should not be together. In any case, we don't ↵Pascal Rigaux2006-11-201-1/+1
| | | | want to workaround a bad {with_hdlist}
* simplify call to _probe_with_try_list()Pascal Rigaux2006-11-201-3/+5
|
* have same _guess_hdlist_suffix parameters(), and simplifyPascal Rigaux2006-11-201-4/+3
|
* by safePascal Rigaux2006-11-201-1/+1
|
* fix previous commitPascal Rigaux2006-11-201-1/+1
|
* - remove code handling remote "list" files (we don't support it anymore)Pascal Rigaux2006-11-201-18/+9
| | | | | - cleanup