Unnecessarily complicated?

[JustinGrote]

Hey great job on the omreport parsing, but why did you do all the substringing yourself? This works just as well:

.\omreport.exe storage pdisk controller=0 -fmt ssv | select -skip 4 | convertfrom-csv -Delimiter ';'

Just wondering what I'm missing.

[alx9r]

@JustinGrote Thanks. It's been some time since I looked at this. If I recall correctly you are right that this is unnecessarily complicated, at least for today's omreport.exe. I believe the most reasonable way to achieve this is to simply convert the xml representation of the omreport.exe output to the .Net XML object. I expect that would look something like [xml]$xmlStringProducedbyOmreport which would be less to go wrong than what's happening in this module.

I can't remember exactly what I was facing that I ended up down this path, but I'd definitely pursue the XML parsing route rather than parsing of un- or semi-structured strings.

[JustinGrote]

I'm working on it now, sadly it's not easy enough to use the xml as the objects are in tons of different formats (some use attributes, some use

text) , and they use the non-resolved status integers that I can't find a

reliable reference for anywhere.

I'm going the ssv format, with a regex to pull out the relevant sections as the pull is inconsistent too.

On Wed, Jul 17, 2019, 1:36 PM alx9r notifications@github.com wrote:

@JustinGrote https://github.com/JustinGrote Thanks. It's been some time since I looked at this. If I recall correctly you are right that this is unnecessarily complicated, at least for today's omreport.exe. I believe the most reasonable way to achieve this is to simply convert the xml representation of the omreport.exe output to the .Net XML object. I expect that would look something like [xml]$xmlStringProducedbyOmreport which would be less to go wrong than what's happening in this module.

I can't remember exactly what I was facing that I ended up down this path, but I'd definitely pursue the XML parsing route rather than parsing of un- or semi-structured strings.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/alx9r/PsOmsa/issues/10?email_source=notifications&email_token=ADUNKUQLJWKAHUMV5HAIHDDP757D7A5CNFSM4IERX622YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2GQLEI#issuecomment-512558481, or mute the thread https://github.com/notifications/unsubscribe-auth/ADUNKUQGPXLF3L2DOZGZCM3P757D7ANCNFSM4IERX62Q .

[JustinGrote]

OK, here's my SSV parser, it has to work around a couple dumb things like:

• omreport inserts random tabs for no reason in some output (storage enclosure)
• It can have "error messages" like no items found but still gives a zero (success) exit status
• There are multiple header deliniation formats depending on which query you run.

I only tested against storage queries (since most other OMReport info I can get from CIM, and against a system with multiple of all items. Feel free to swipe it.

Storage Commands Tested:

storage vdisk
storage battery
storage enclosure
storage controller
storage connector controller=0
storage connector controller=1
storage connector controller=2
storage connector controller=3
storage pdisk controller=0
storage pdisk controller=1
storage pdisk controller=2
storage pdisk controller=3

https://gist.github.com/JustinGrote/a1c71a0c0f4d30d69dccbedc29329521

EDIT: I revised it to match on SSV blocks rather than try to split on the super inconsistent headers. So far is much more reliable for the output.