Jira (PUP-9305) Built-in unique function behaves differently with undef
Title: Message Title Sean Millichamp commented on PUP-9305 Re: Built-in unique function behaves differently with undef Just wanted to add my $0.02: I wish more of the stdlib functions handled types more strictly than they have and I think the behavior of the core unique makes a lot more sense. If I call unique on something I expect that it will be an array. If it is not an array then I probably haven't handled something correctly earlier in the code (either a bug or failed to validate input correctly) and I would want it to raise an error. Please don't change the core unique behavior as added in Puppet 5. It is the right direction for the language/platform. Add Comment This message was sent by Atlassian JIRA (v7.7.1#77002-sha1:e75ca93) -- You received this message because you are subscribed to the Google Groups "Puppet Bugs" group. To unsubscribe from this group and stop receiving emails from it, send an email to puppet-bugs+unsubscr...@googlegroups.com. To post to this group, send email to puppet-bugs@googlegroups.com. Visit this group at https://groups.google.com/group/puppet-bugs. For more options, visit https://groups.google.com/d/optout.
Jira (PUP-9305) Built-in unique function behaves differently with undef
Title: Message Title
Henrik Lindberg commented on PUP-9305
Re: Built-in unique function behaves differently with undef
Sean Millichamp That is what I am leaning towards (not changing) since "give me a unique set of undef" is a strange thing to ask for and most likely masks some other problem if it was to return undef. We had a similar request to make empty() accept undef and return true - there we had lots of breakage of legacy code when it did not accept undef, and it seemed reasonably harmless to allow that even if you can argue that undef is indeed undefined and you cannot even know if it is something "empty or not". OTOH, undef is often referred to as "the empty set". In this case (for unique) this ticket was filed long after 5.0.0 was released - so cannot be a common problem and is easily avoided by using type checking. You can argue that if undef is indeed "the empty set", then a unique should return that empty set - e.g. an empty array. But, if undef represents the "empty set", then it would be reasonable to return undef - but the problem with returning undef is that it should return something that is iterable (for example a String if given a String, and that works also on an empty string). So to comply with that requirement it would need to return an empty iterable (for example an empty Array). So, my position is:
I don't like the idea of turning undef into an Array since I think this masks a problem
I don't like the idea of turning undef into undef since it will move the error to where you try to iterate over the undef result - so this goes against "catch the error early"
I don't like making undef iterable as if it was an empty array - I think that would just mask other errors and move the problem making it more difficult to find the cause.
If accepting something like Optional[Array] and wanting to do a unique it is easy to use conditional logic or calling then
# Three different ways to check with conditional logic
$result = if $maybe_an_array =~ Array { $maybe_an_array.unique } # results in undef if not an array
$result = $maybe_an_array ? { Array => $maybe_an_array.unique, default => undef }
$result = $maybe_an_array.then |$x| { $x.unique }
Jira (PDB-4182) Use beaker 4 in pdb acceptance tests
Title: Message Title Zachary Kent created an issue PuppetDB / PDB-4182 Use beaker 4 in pdb acceptance tests Issue Type: Task Assignee: Unassigned Created: 2018/11/09 9:49 AM Priority: Normal Reporter: Zachary Kent Add Comment This message was sent by Atlassian JIRA (v7.7.1#77002-sha1:e75ca93) -- You received this message because you are subscribed to the Google Groups "Puppet Bugs" group. To unsubscribe from this group and stop receiving emails from it, send an email to puppet-bugs+unsubscr...@googlegroups.com. To post to this group, send email to puppet-bugs@googlegroups.com. Visit this group at https://groups.google.com
Jira (BOLT-969) bolt-server clears its tmpdir during cache purging
Title: Message Title Nick Lewis created an issue Puppet Task Runner / BOLT-969 bolt-server clears its tmpdir during cache purging Issue Type: Bug Assignee: Unassigned Created: 2018/11/09 1:52 PM Priority: Normal Reporter: Nick Lewis By default, bolt-server clears out any cached tasks (directories in cache) which haven't been touched in a week. Unfortunately, bolt-server also uses a tmp dir in the cache as the temporarily download location for tasks until copying them into the proper location. That tmp dir is also subject to the cache purge, which means it will be deleted if no new tasks have been downloaded in the past week. Once that directory has been deleted, new task downloads will fail until bolt-server is restarted, at which point it will recreate the directory. The cache/tmp directory should either be moved outside the cache directory or excluded from the purge. Add Comment
Jira (BOLT-969) bolt-server clears its tmpdir during cache purging
Title: Message Title Alex Dreyer updated an issue Puppet Task Runner / BOLT-969 bolt-server clears its tmpdir during cache purging Change By: Alex Dreyer Priority: Normal Major Add Comment This message was sent by Atlassian JIRA (v7.7.1#77002-sha1:e75ca93) -- You received this message because you are subscribed to the Google Groups "Puppet Bugs" group. To unsubscribe from this group and stop receiving emails from it, send an email to puppet-bugs+unsubscr...@googlegroups.com. To post to this group, send email to puppet-bugs@googlegroups.com. Visit this group at https://groups.google.com/group/puppet-bugs. For more options, visit https://groups.google.com/d/optout.
Jira (BOLT-969) bolt-server clears its tmpdir during cache purging
Title: Message Title Alex Dreyer updated an issue Puppet Task Runner / BOLT-969 bolt-server clears its tmpdir during cache purging Change By: Alex Dreyer Sprint: Bolt Ready for Grooming Add Comment This message was sent by Atlassian JIRA (v7.7.1#77002-sha1:e75ca93) -- You received this message because you are subscribed to the Google Groups "Puppet Bugs" group. To unsubscribe from this group and stop receiving emails from it, send an email to puppet-bugs+unsubscr...@googlegroups.com. To post to this group, send email to puppet-bugs@googlegroups.com. Visit this group at https://groups.google.com/group/puppet-bugs. For more options, visit https://groups.google.com/d/optout.
Jira (PUP-9307) Account for trailing slashes in mount resource titles
Title: Message Title
Adrian Parreiras Horta created an issue
Puppet / PUP-9307
Account for trailing slashes in mount resource titles
Issue Type:
Bug
Assignee:
Unassigned
Created:
2018/11/09 2:14 PM
Labels:
cspapercuts
Priority:
Minor
Reporter:
Adrian Parreiras Horta
Consider the following example mount resource:
mount {'/test/':
ensure => mounted,
fstype => "iso9660",
device => "/root/vim.iso",
options => "loop,ro",
}
Jira (PUP-9308) stack level too deep (SystemStackError)
Title: Message Title ken ma created an issue Puppet / PUP-9308 stack level too deep (SystemStackError) Issue Type: Bug Affects Versions: PUP 5.5.6 Assignee: Unassigned Components: Compiler Created: 2018/11/09 4:30 PM Fix Versions: PUP 5.5.z Priority: Major Reporter: ken ma Puppet Version:5.5.x Puppet Server Version:5.5.6 OS Name/Version: puppet agent: windows server 2012r2 /puppet server: centos 7 when i running puppet apply pp file and pp size more than 200k, i got error: C:/Program Files/Puppet Labs/Puppet/sys/ruby/lib/ruby/vendor_ruby/puppet/parser/compiler.rb:33:in `new': stack level too deep (SystemStackError) same pp i was running on puppet 3.7, it's work well. Add Comment
