Present: me, chinmoy

Minutes in the attachment.

Cheers,
Elvis
2017-07-15 10:00:18     eang    Meeting?
2017-07-15 10:00:39     chinmoy[m]      i am here
2017-07-15 10:01:24     eang    let's start, as arnav said he wouldn't make it
2017-07-15 10:01:35     chinmoy[m]      sure.
2017-07-15 10:02:19     chinmoy[m]      this week i mostly created patches.
2017-07-15 10:03:01     chinmoy[m]      however i found one serious flaw in my 
logic in showing the warning from ioslave
2017-07-15 10:03:47     chinmoy[m]      in short with the available 
infrastructure showing warning at proper times is not just possible
2017-07-15 10:04:31     chinmoy[m]      my previous patches just happen to work
2017-07-15 10:05:34     eang    what's the problem?
2017-07-15 10:07:29     chinmoy[m]      the warning from ioslave is only shown 
properly when a single folder is being deleted and it contains one or more read 
restricted file.
2017-07-15 10:08:33     chinmoy[m]      with arnav's kio-stash and in copy 
operation the warning would have been shown as many times as there are 
read-restricted files
2017-07-15 10:11:12     eang    I see, how are you planning to address this?
2017-07-15 10:13:48     chinmoy[m]      In case you are interested, here's the 
cause. KIO::del and copy are high level jobs. they create number of subjobs 
according to the number of items, so, there are that many ioslaves as well. And 
there is no way for an ioslave to check for existence of its other siblings.
2017-07-15 10:14:46     chinmoy[m]      Right now, I removed the code for 
warning from ioslave
2017-07-15 10:16:14     chinmoy[m]      added a job flag PrivilegeExecution for 
application to use and a public method enablePrivilegeExecution which will be 
used from the job's side if the above flag is encountered.
2017-07-15 10:17:02     chinmoy[m]      and those jobs which use this method 
will also show warning dialogs as well
2017-07-15 10:19:11     eang    the problem of checking whether other ioslaves 
are running could be solved with a kded plugin
2017-07-15 10:19:13     chinmoy[m]      and when execAsPrivilegeUser is called 
a signal will be sent to job requessting if the method should be executed and 
it will be the responsibility of the job to send back a message containing 
"ElevatePrivilege"
2017-07-15 10:19:33     eang    (in case you want to give it a shot)
2017-07-15 10:19:59     chinmoy[m]      can you explain it a bit
2017-07-15 10:20:48     eang    kded is always running, so an ioslave can query 
it any time
2017-07-15 10:21:10     eang    the plugin could have dbus methods such as 
"registerJob", "isJobRunning", etc.
2017-07-15 10:22:22     chinmoy[m]      the jobs are scheduled synchronously 
(addSubJob) so i dont think it would serve the purpose
2017-07-15 10:23:04     chinmoy[m]      the slaves will be destroyed as soon as 
the job finishes no?
2017-07-15 10:24:27     eang    yes, but maybe it's possible to query kded 
between schedule time and deletion time? I don't know
2017-07-15 10:24:47     eang    anyway, it would complicate things for sure
2017-07-15 10:25:18     chinmoy[m]      definitely
2017-07-15 10:25:36     chinmoy[m]      well give me few minutes I will update 
my patch
2017-07-15 10:26:14     chinmoy[m]      I think with my current approach the 
loopholes David mentioned should be no more
2017-07-15 10:27:06     eang    ok sounds good
2017-07-15 10:28:44     eang    chinmoy: anything else you wanted to discuss?
2017-07-15 10:31:36     chinmoy[m]      there's one thing, what does BIC stands 
for? I think i didn't get it properly.
2017-07-15 10:32:05     eang    have you read 
https://community.kde.org/Policies/Binary_Compatibility_Issues_With_C%2B%2B ?
2017-07-15 10:32:26     chinmoy[m]      ok BCI
2017-07-15 10:33:42     chinmoy[m]      I am slowly familiarising myself with 
abbreviations
2017-07-15 10:34:07     eang    I'm actually not sure what BIC stands for
2017-07-15 10:34:07     chinmoy[m]      so it is BIC after all
2017-07-15 10:34:23     eang    the important thing is you understand the 
concept
2017-07-15 10:34:45     eang    (it could be Binary Interface Compatibility or 
something similar)
2017-07-15 10:35:04     chinmoy[m]      whatever, now i understood
2017-07-15 10:35:44     eang    meeting over then?
2017-07-15 10:36:01     chinmoy[m]      i think yes

Reply via email to