Re: [Yum-devel] [PATCH 2/2] Uncompress 'updateinfo.xml.gz' after downloading.

2012-04-17 Thread James Antill
On Tue, 2012-04-17 at 14:34 +0200, Zdeněk Pavlas wrote: > This works around the problem with non-root process > trying to gen_decompress() it in a read-only directory. NAK. I'm not sure workarounds like this are worth it ... we'd need to add ones for pkgtags and groups (and deltainfo, etc.) Rea

Re: [Yum-devel] [PATCH 1/2] yum -C should use the system cachedir.

2012-04-17 Thread James Antill
On Tue, 2012-04-17 at 14:34 +0200, Zdeněk Pavlas wrote: > > --cacheonly > If you're using this as a user yum will not use the tempcache for > the user but will only use the system cache in the system cachedir. > > > setCacheDir() sets up the user tempcache when UID != 0. > Thus, we should avoid

Re: [Yum-devel] [PATCH 1/2] Check CWD earlier. Related: rhbz#711358

2012-04-17 Thread James Antill
On Tue, 2012-04-17 at 13:45 +0200, Zdeněk Pavlas wrote: > Check the current directory before instantiating YumBase, > and before processing config and options. This avoids > a possible traceback with '-cfg' option. ACK. ___ Yum-devel mailing list Yum-

Re: [Yum-devel] [PATCH 2/2] Touch repoXML later, to handle RepoError properly. Related: rhbz#769864

2012-04-17 Thread James Antill
On Tue, 2012-04-17 at 13:45 +0200, Zdeněk Pavlas wrote: > The removed loop does not handle skip-if-unavailable option, > while repos.populateSack() does. > > Forward port from RHEL6.3 ACK. > --- > yumcommands.py |2 -- > 1 files changed, 0 insertions(+), 2 deletions(-) > > diff --git a/yu

Re: [Yum-devel] [PATCH] repo.async = 'auto' option

2012-04-17 Thread James Antill
On Tue, 2012-04-17 at 10:04 +0200, Zdeněk Pavlas wrote: > Make 'async' option 3-state. 'auto' evaluates to '1' > only in base Yum repository parsed from INI file. My first instinct is leave it as a boolean and have auto == on. What is the desired behaviour difference between "on" and "auto" ...

[Yum-devel] [PATCH 2/2] Uncompress 'updateinfo.xml.gz' after downloading.

2012-04-17 Thread Zdeněk Pavlas
This works around the problem with non-root process trying to gen_decompress() it in a read-only directory. http://lists.baseurl.org/pipermail/yum-devel/2012-April/009208.html --- yum/yumRepo.py |2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/yum/yumRepo.py b/yum/yumRepo

[Yum-devel] [PATCH 1/2] yum -C should use the system cachedir.

2012-04-17 Thread Zdeněk Pavlas
--cacheonly If you're using this as a user yum will not use the tempcache for the user but will only use the system cache in the system cachedir. setCacheDir() sets up the user tempcache when UID != 0. Thus, we should avoid calling it with --cacheonly. --- cli.py |4 +--- 1 files changed, 1

[Yum-devel] [PATCH 1/2] Check CWD earlier. Related: rhbz#711358

2012-04-17 Thread Zdeněk Pavlas
Check the current directory before instantiating YumBase, and before processing config and options. This avoids a possible traceback with '-cfg' option. Forward port from RHEL6.3 --- yummain.py | 24 1 files changed, 12 insertions(+), 12 deletions(-) diff --git a/yumm

[Yum-devel] [PATCH 2/2] Touch repoXML later, to handle RepoError properly. Related: rhbz#769864

2012-04-17 Thread Zdeněk Pavlas
The removed loop does not handle skip-if-unavailable option, while repos.populateSack() does. Forward port from RHEL6.3 --- yumcommands.py |2 -- 1 files changed, 0 insertions(+), 2 deletions(-) diff --git a/yumcommands.py b/yumcommands.py index 9c181ec..477d16a 100644 --- a/yumcommands.py +

[Yum-devel] [PATCH] repo.async = 'auto' option

2012-04-17 Thread Zdeněk Pavlas
Make 'async' option 3-state. 'auto' evaluates to '1' only in base Yum repository parsed from INI file. When plugins create subclasses they may inherit options from Yum repository so we should not just modify them. Used a private _async attribute instead. --- docs/yum.conf.5 |4 ++-- yum/__in