| Time |
S |
Nick |
Message |
| 19:22 |
|
|
IRCFrEAK joined #sourcefu |
| 23:11 |
|
dotplus |
Been writing python for years on and off, but just released my first project to pypi: https://pypi.python.org/pypi/sensu-cli |
| 23:13 |
|
pdurbin |
dotplus: nice! Congrats! |
| 23:14 |
|
dotplus |
I feel like I'm about 12 \o/ |
| 23:15 |
|
pdurbin |
heh, I'm kind of excited about the "feedcritic" project I've been hacking on the last week or so: https://github.com/pdurbin/feedcritic |
| 23:16 |
|
pdurbin |
this afternoon I switched http://podcasts.greptilian.com over to it so I guess it's good enough for production :) |
| 23:16 |
|
pdurbin |
now I'm fighting with my Makefile and wondering why I decided to use make anyway :) |
| 23:16 |
|
dotplus |
excellent! well done. |
| 23:17 |
|
dotplus |
make is like a god of Ancient Myth: both awesome and terrible |
| 23:18 |
|
dotplus |
I'll take a look if you have a specific problem |
| 23:21 |
|
pdurbin |
dotplus: sure, please check out this "FIXME only run this if TSV was passed in" at https://github.com/pdurbin/feedcritic/blob/080406a7fbde4118e494a79deae350a367b332a1/Makefile#L11 |
| 23:22 |
|
pdurbin |
As I try to explain in "usage" DIR is required but TSV should be optional. |
| 23:24 |
|
dotplus |
you just want to switch on whether the TSV var exists/is populated or not? |
| 23:24 |
|
pdurbin |
yeah |
| 23:24 |
|
pdurbin |
only copy the tsv if TSV is defined |
| 23:25 |
|
dotplus |
ok, PR incoming in a minute |
| 23:25 |
|
pdurbin |
\o/ |
| 23:29 |
|
dotplus |
there ya go. untested, but that should do it. subject to tab vs. space fail and the like because make(1) is awkward. |
| 23:30 |
|
pdurbin |
heh. maybe I'll try it out locally before merging. thanks :) |
| 23:32 |
|
bear |
you can one-line it by doing [[ "" == $TSV ]] && cp ($TSV) . (or something like that) |
| 23:32 |
|
dotplus |
yep, that too |
| 23:32 |
|
bear |
I find that is sometimes easier to avoid the multiline \ stuff |
| 23:36 |
|
dotplus |
actually, in this case, I would use bear's oneliner. But I wanted to show the general approach. |
| 23:36 |
|
bear |
good point |
| 23:36 |
|
pdurbin |
heh, I was just experimenting with putting this on one line |
| 23:36 |
|
pdurbin |
because the multiline doesn't seem to just work |
| 23:37 |
|
bear |
the "trick" is remembering that [[ ]] has a return value as it's a built-in function |
| 23:37 |
|
dotplus |
oh. and one other thing. isn't make(1) going to use sh(1) not bash? |
| 23:37 |
|
bear |
oh - yea - I think I relearn that everytime I write one of these |
| 23:37 |
|
dotplus |
so [[ would not be available. |
| 23:39 |
|
dotplus |
[ "" == "$TSV" ] && cp ($TSV) . |
| 23:39 |
|
bear |
dotplus++ |
| 23:39 |
|
pdurbin |
dotplus: ok, merged, with my edit. thanks! |
| 23:39 |
|
pdurbin |
now you folks can tell me if I'm crazy for using make for this :) |
| 23:40 |
|
bear |
`SHELL := /bin/bash` FTW ;) |
| 23:40 |
|
* dotplus |
looks askance at bear |
| 23:40 |
|
bear |
you will never (well very rarely) hear me complain about using a Makefile |
| 23:42 |
|
bear |
my suggestion would be to use make(1) fully and refactor your clean and cleanxml targets to use dependencies so it's make(1) that builds the list of files and they are not hardcoded |
| 23:42 |
|
dotplus |
I like make. it's almost always going to available, likely already installed. It's well known and yeah, it's got problems, but familiarity is worth a lot. |
| 23:43 |
|
pdurbin |
yeah, it's everywhere |
| 23:43 |
|
pdurbin |
I mean, longer term, I would make a deb and an RPM. |
| 23:43 |
|
bear |
the other nit I would pick is that lines 6 and 7 should be removed and change line 5 to read `all: clean cleanxml` |
| 23:45 |
|
dotplus |
not a nit, a sensible improvement. if nothing else, it's more idiomatic make. |
| 23:45 |
|
pdurbin |
I almost never use make so I'll defer to you guys. |
| 23:46 |
|
bear |
*nods* - I only used nit as I didn't know his level of make experience |
| 23:46 |
|
bear |
but yes, right now it's making me itch a bit |
| 23:47 |
|
bear |
is the deploy step supposed to be run if the output files are generated? |
| 23:47 |
|
bear |
where output files seem to be "index.html feedcritic.js rssfeed.svg podcasts.json latest.json" |
| 23:48 |
|
pdurbin |
it's complicated |
| 23:49 |
|
pdurbin |
the plan is to run this via cron, which will update podcasts.json and latest.json every night |
| 23:49 |
|
pdurbin |
the only files only change when I fix a bug or add a feature or whatever |
| 23:49 |
|
bear |
k, then leave it as is - adding make(1) complexity will probably not help |
| 23:49 |
|
pdurbin |
now I know where to come for the make fu :) |
| 23:51 |
|
pdurbin |
Does anyone here listen to podcasts besides sivoais ? Everyone is welcome to try it out. :) |
| 23:51 |
|
dotplus |
btw, I'm looking for a new job op. As you know, I have lots of "standard" unix sysadmin background particularly CM. Most recent experience is Ansible, Openstack, python and Jenkins/Gerrit/etc for CI for all of them. |
| 23:52 |
|
bear |
I used to listen to a lot of them |
| 23:52 |
|
bear |
dotplus - want to come work for me and help wrangle SRE stuff for CircleCI - lots of sysadmin, kubernetes, nomad work |
| 23:52 |
|
pdurbin |
dotplus: does the location matter? |
| 23:52 |
|
dotplus |
so if any of you know someone looking for someone like me (remote), I'd love an introduction. |
| 23:53 |
|
dotplus |
bear: er quite possibly. you're hiring? |
| 23:53 |
|
bear |
yes |
| 23:53 |
|
bear |
and we are very remote friendly |
| 23:54 |
|
dotplus |
I know. got a more complete description posted somewhere. From what I've heard circleci sounds very promising. |
| 23:54 |
|
bear |
it's a very challenging team because we are not your usual [Sys|Dev]Ops shop |
| 23:54 |
|
dotplus |
s/somewhere/\1?/ |
| 23:54 |
|
bear |
my team is tasked with making it so that devs can manage their own tools and services in production |
| 23:55 |
|
bear |
yea, we are trying to get the formal job description online - I can email you the draft if your curious |
| 23:55 |
|
dotplus |
I am very much interested! |
| 23:56 |
|
bear |
where are you located at? |
