[anaconda][RFC][?master/f21-branch/rhel7-branch?][PATCH] Use tty1 for useful hints
by Martin Kolman
A while ago during an IRC discussion a friend mentioned that he recently tried to install Fedora
but had some weird issues with his graphics hardware that left him staring at a black screen
(apparently X kinda started but not completely, so the text mode fallback did not trigger). He tried
various things (multiple media version, checking media checksum etc.) but always got the same behavior.
I asked him for logs and told him about text mode - unfortunately he already reinstalled the machine long ago
with some other distro. But he mentioned that he tried to switch ttys, but he couldn't find any hints what
might be wrong, where could he find the log files, about trying text mode, etc.
That was when I realized that tty1, which could be the obvious choice of installer users in trouble currently
only shows Anaconda version and has a lot of unused screen real estate. From there I got the idea to use tty1
for displaying important hints that might help users that have installation issues find what when wrong and what
to do now - that the shell is running TMUX, where to find logs, how to force text mode - and importantly -
how to properly file bugreports. ;-)
I see the patch as kind for request for comments/proof of concept and I'm looking forward to you feedback! :)
Martin Kolman (1):
initial implementation
anaconda | 12 ++++++++++++
1 file changed, 12 insertions(+)
--
1.9.3
9 years, 5 months
[PATCH] test the reclaim dialog
by Chris Lumens
Here's a test of the reclaim dialog when it should succeed due to having
plenty of reclaimable space. That part's easy. The real work here is
making sure multiple tests run in parallel so when we start having a lot
of GUI tests, they don't take all day to run.
You may notice I am being lame and using parted+mkfs to create the filesystem
I then try to reclaim later. I should probably be using blivet directly to
do this. Patches welcome.
Also note that I am giving each VM 2 GB of memory so there's always going to be
a limit on the number of VMs that can be run. I can perhaps not hardcode the
number of processes and instead figure it out based on the amount of free
memory. I can hopefully also get us more memory for our test systems.
So, where should I be going next?
- Chris
9 years, 5 months
[PATCH] Do not translate empty strings, gettext translates them into system information (basically PO file header)
by Vojtěch Trefný
---
blivet/i18n.py | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/blivet/i18n.py b/blivet/i18n.py
index 1f07266..2f033bf 100644
--- a/blivet/i18n.py
+++ b/blivet/i18n.py
@@ -45,8 +45,8 @@ N_ = lambda x: x
# evaluated on every call.
# pylint: disable=unnecessary-lambda
if six.PY2:
- _ = lambda x: _get_translations().ugettext(x)
+ _ = lambda x: _get_translations().ugettext(x) if x else ""
P_ = lambda x, y, z: _get_translations().ungettext(x, y, z)
else:
- _ = lambda x: _get_translations().gettext(x)
+ _ = lambda x: _get_translations().gettext(x) if x else ""
P_ = lambda x, y, z: _get_translations().ngettext(x, y, z)
--
2.1.0
9 years, 5 months
[anaconda][master][PATCH] Reset All - Revert All Changes
by Martin Kolman
A user has suggested in #1163701 that it might not be very evident what
the Reset All button in custom partitioning spoke (he though it will
reset his partition table or do something else destructive, even though there
the we-won't-touch-your-disks label right above it) does. It is also kinda
confusing what's the difference between the Reset All an reload storage configuration
buttons.
So I've renamed the button to "Revert All Changes" which should make quite apparent what it does
and how it is different from the reload button. I still feel that "Revert" might a bit too technical
but have not been able to come with anything better and "Reset All Changes" does not sound right either.
As we are well past the string freeze this is master only.
Martin Kolman (1):
Make it more evident what the Revert changes button does (#1163701)
pyanaconda/ui/gui/spokes/custom.glade | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--
1.9.3
9 years, 5 months
[blivet:master 0/2] Two patches about the failing selinux tests
by Anne Mulhern
mulhern (2):
Make selinux test less precise.
Make logging a little less verbose and more useful in FS.mount()
blivet/formats/fs.py | 6 ++++--
tests/formats_test/selinux_test.py | 28 ++++++++--------------------
2 files changed, 12 insertions(+), 22 deletions(-)
--
1.9.3
9 years, 5 months
[rhel7-branch][PATCH] Convert LVM_PE_SIZE to KiB (#1163081)
by Vratislav Podzimek
It is a Size instance not a plain number of MiB now.
Signed-off-by: Vratislav Podzimek <vpodzime(a)redhat.com>
---
pyanaconda/kickstart.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/pyanaconda/kickstart.py b/pyanaconda/kickstart.py
index 379a573..4122c02 100644
--- a/pyanaconda/kickstart.py
+++ b/pyanaconda/kickstart.py
@@ -1592,7 +1592,7 @@ class VolGroupData(commands.volgroup.RHEL7_VolGroupData):
if self.pesize == 0:
# default PE size requested -- we use blivet's default in KiB
- self.pesize = LVM_PE_SIZE * 1024
+ self.pesize = LVM_PE_SIZE.convertTo("KiB")
pesize = Size("%d KiB" % self.pesize)
if pesize not in getPossiblePhysicalExtents():
--
1.9.3
9 years, 5 months
[PATCH] Check for a GLib source ID of None in unwatchAllProcesses.
by David Shea
This check is in unwatchProcess but didn't make it to the other method.
The source ID will be None if the switch to GLib process watching done
by the GUI has not completed.
---
pyanaconda/iutil.py | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/pyanaconda/iutil.py b/pyanaconda/iutil.py
index bcd5b55..5549e5a 100644
--- a/pyanaconda/iutil.py
+++ b/pyanaconda/iutil.py
@@ -541,7 +541,8 @@ def unwatchAllProcesses():
"""Clear the watched process list."""
global _forever_pids
for child_pid in _forever_pids:
- GLib.source_remove(_forever_pids[child_pid][1])
+ if _forever_pids[child_pid][1]:
+ GLib.source_remove(_forever_pids[child_pid][1])
_forever_pids = {}
def getDirSize(directory):
--
2.1.0
9 years, 5 months
[master/rhel7/f21][PATCH] Check if we read something when emptying stdin queue
by Vratislav Podzimek
In some cases (looking at you s390x again) select may report there's something
on stdin while there's nothing and stdin.read(1) returns nothing. If that
happens, we are good to stop reading from stdin because there's nothing queued.
Related: rhbz#1162702
Signed-off-by: Vratislav Podzimek <vpodzime(a)redhat.com>
---
pyanaconda/ui/lib/entropy.py | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/pyanaconda/ui/lib/entropy.py b/pyanaconda/ui/lib/entropy.py
index 52e595d..401053b 100644
--- a/pyanaconda/ui/lib/entropy.py
+++ b/pyanaconda/ui/lib/entropy.py
@@ -116,9 +116,10 @@ def _tui_wait(msg, desired_entropy):
# and then just read everything from the input buffer and revert the
# termios state
- while sys.stdin in select.select([sys.stdin], [], [], 0)[0]:
+ data = "have something"
+ while sys.stdin in select.select([sys.stdin], [], [], 0)[0] and data:
# just read from stdin and scratch the read data
- sys.stdin.read(1)
+ data = sys.stdin.read(1)
if termios_attrs_changed:
termios.tcsetattr(fd, termios.TCSAFLUSH, old)
--
1.9.3
9 years, 5 months
[master/rhel7/f21] Two more entropy-related issues
by Vratislav Podzimek
These two patches fixes two more entropy-related issues discovered by QA --
getting/setting terminal attributes is not supported on s390x and 'part
--encrypted' in kickstart should result in the same behaviour as far as entropy
requirements are concerned.
Both bugs still wait for their acks, but I really think they deserve them.
Vratislav Podzimek (2):
Don't rely on terminal attributes being configurable (#1162702)
Require min entropy for 'part --encrypted' devices (#1162695)
pyanaconda/kickstart.py | 14 +++++++++-----
pyanaconda/ui/lib/entropy.py | 24 ++++++++++++++++++------
2 files changed, 27 insertions(+), 11 deletions(-)
--
1.9.3
9 years, 5 months
[blivet/f21/master] Add more arguments to mpathconf (#1154347)
by David Shea
Setting find_multipaths to yes will keep multipath from creating a
multipath device unless one of the following conditions is met:
- two non-blacklisted paths have the same wwid
- a path has the same wwid as a previously created multipath device
- the user forced the creation of a device
In other words, only create a multipath device if there is a device with
multiple paths. This would be helpful.
---
blivet/devicelibs/mpath.py | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/blivet/devicelibs/mpath.py b/blivet/devicelibs/mpath.py
index 3f108a3..7bacc52 100644
--- a/blivet/devicelibs/mpath.py
+++ b/blivet/devicelibs/mpath.py
@@ -25,5 +25,7 @@ def set_friendly_names(enabled=True):
else:
val = "n"
- cmd = ["mpathconf", "--user_friendly_names", val, "--with_multipathd", "y"]
+ # --find_multipaths is important to keep multipath from making up multipath devices
+ # that aren't really multipath
+ cmd = ["mpathconf", "--find_multipaths", "y", "--user_friendly_names", val, "--with_multipathd", "y"]
return (util.run_program(cmd) == 0)
--
2.1.0
9 years, 5 months