**** BEGIN LOGGING AT Tue Nov 28 03:00:01 2017 Nov 28 07:29:54 good morning Nov 28 09:20:00 hello Nov 28 09:20:19 can we we set and unset vars like this do_fetch[noexec] = "" Nov 28 09:20:29 do_fetch[noexec] = "1" Nov 28 09:20:38 setting works Nov 28 09:21:23 I have a situation whaere if foo is set then I don't want to run fetch task Nov 28 09:21:43 and if it is set to 0 then I want to execute do_fetch Nov 28 09:21:55 is it possible using the above syntax Nov 28 09:58:04 Noor: hey Nov 28 09:58:21 Noor: I think the only way is to do it via anonymous python Nov 28 09:58:38 bluelightning: that is what I figured :) Nov 28 09:58:43 moving to that now Nov 28 10:49:30 bluelightning: isn't it a bug Varflag is not set to desired value using this syntax Nov 28 11:31:09 Noor: the varflag is being set as instructed though... the problem is that the code that checks for the noexec varflag only checks if it is set, not what its value is Nov 28 11:44:56 bluelightning: even in this case do_fetch[noexec] = "" flag will be set? Nov 28 11:45:19 and shouldn't it check the value instead of jsut checking it is set or not Nov 28 15:08:11 Noor: yes, it should check the value, it's a bug if it only checks existence, which i know was the case in the past and might still be the case. it'd be nice to switch it to use to_boolean() to be consistent with other flags/variables with boolean semantics Nov 28 20:09:09 I currently have a strange problem with mercurial fetching. I used the AUTOREV which worked so far without problems in morty. Now I updated to rocko and during parsing of recipes it spawns a lot of processes to get the latest revision from the HG repository. Although the sqlite cache is filed, it looks like first a lot of threads are doing a lookup before the first thread is able to fill the cache. But until then a lot of other lookups h Nov 28 20:12:33 So parsing takes a lot of time, because it seems like the evaluation of SRCPV set to "${@bb.fetch2.get_srcrev(d)}") in base.bbclass is executed in parallel. When I add rev = bb.fetch2.get_srcrev(d) to the base.bbclass after elif scheme == "hg": it resolves the problem as cache seems to work now **** ENDING LOGGING AT Wed Nov 29 03:00:02 2017