I took a look at the blivet commit 45b1e921 and bug and I'm now quite uncertain that
the constructor added in that call is doing the right thing for blivet. I mean, if I were
using blivet w/out anaconda, and I did
```
>> from blivet import formats
>> z = formats.getFormat("macefi")
```
is it really right that the label should be:
```
>> z.label
'Linux HFS+ ESP'
```
In every other case, not specifying the label means accept the default label that mkfs
will set. In this one case, it means set the label to "Linux HFS+ ESP". Is that
the behavior that any client of blivet ought to expect? If not, it would be best to get
rid of that constructor in blivet, and just go with a slightly extended patch in anaconda,
that sets the format label, like:
```
diff --git a/pyanaconda/ui/gui/spokes/custom.py b/pyanaconda/ui/gui/spokes/custom.py
index 625cf0a..dd53bfa 100644
--- a/pyanaconda/ui/gui/spokes/custom.py
+++ b/pyanaconda/ui/gui/spokes/custom.py
@@ -2404,7 +2404,9 @@ class CustomPartitioningSpoke(NormalSpoke, StorageChecker):
if new_type is None:
return
log.debug("fs type changed: %s", new_type)
- fmt = getFormat(new_type)
+ fmt = getFormat(new_type, label="Linux HFS+ ESP" if new_type ==
"macefi" else None)
+ if fmt.label:
+ self._labelEntry.set_text(fmt.label)
fancy_set_sensitive(self._mountPointEntry, fmt.mountable)
def _populate_container(self, device=None):
```
--
To view this pull request on github, visit
https://github.com/rhinstaller/anaconda/pull/33