McStas + Mantid on windows, seems to only work via mcdisplay.pl

[willend]

On Windows 10, when IDF xml file is generated via mcdisplay-mantid (py), it is not picked up 100% correctly by the subsequent mcrun --format=NeXus.

Mantid barks:

The XML itself loads correctly in Mantid using LoadEmptyInstrument, and is also present inside the HDF file:

I think the issue may be the highlighted filename ^ (red circle) that probably needs to match the name of the binary, in this case ILL_H16_IN5_Mantid.exe

I think this one is not important enough that it should stop us from releasing, but it would be good if you had a look @climbcat . - Idea: Assumption that instrument name is something.out? And hence nothing filled in the template?

[willend]

@trnielsen could you try to check how well this works / does not work in its current state

[climbcat]

The current version of VMWare needs to do some updates, but the auto-update mechanism fails with this (from the log file) which prevents compilation:

2018-10-16T13:13:44.838+02:00| vthread-1| I125: Failed to find /lib/modules/4.15.0-34-generic/build/include/linux/version.h

[willend]

Hi there,

You are likely missing the Linux kernel headers, try

sudo apt-get install linux-headers-generic

and have another go.

Best and greetings from a plane ✈️ over the Atlantic.

Peter

---

Fra: Jakob Garde notifications@github.com Sendt: tirsdag, oktober 16, 2018 7:19 AM Til: McStasMcXtrace/McCode Cc: Peter Kjær Willendrup; Assign Emne: Re: [McStasMcXtrace/McCode] McStas + Mantid on windows, seems to only work via mcdisplay.pl (#459)

The current version of VMWare needs to do some updates, but the auto-update mechanism fails with this (from the log file) which prevents compilation:

2018-10-16T13:13:44.838+02:00| vthread-1| I125: Failed to find /lib/modules/4.15.0-34-generic/build/include/linux/version.h

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHubhttps://github.com/McStasMcXtrace/McCode/issues/459#issuecomment-430198556, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AIMfy12BLvA8D3lrtr2B90B4MDtrMnGMks5ulb_OgaJpZM4NRipm.

[climbcat]

Have a nice journey!

[trnielsen]

rc07 win1064 does not install behind proxy. See issue #687

[willend]

@trnielsen while we wait for @ebknudsen and I to do something about #689, could you possibly proceed by evaluating if (most of) the geometry information is created correctly using the mcdisplay-mantid Python script?

(My expectation is that it close to works, but we should ensure that the @ebknudsen patches done in https://github.com/McStasMcXtrace/McCode/commit/b2c99a8bf9c4e39381f800379fad37f900d6e941#diff-274030ef2e9c5e657607e41533668fa4 and https://github.com/McStasMcXtrace/McCode/commit/b0dd14f7cc0de76bc89827eb86414ff4ad260727#diff-274030ef2e9c5e657607e41533668fa4 are ported to the Python version also)

[willend]

Have attempted to port the changes from https://github.com/McStasMcXtrace/McCode/commit/b0dd14f7cc0de76bc89827eb86414ff4ad260727#diff-274030ef2e9c5e657607e41533668fa4 over to Python by https://github.com/McStasMcXtrace/McCode/commit/ab0c247a0e2ca13885a7c2123ae3db92516f277e.

Currently my Mantid crashes on visualising the instrument geometry outputting a resulting dataset. We should consider if this is worth pursuing this side of release!

[ebknudsen]

A thought: possibly this relates to the 1-off error fixed by b2c99a8. I can't say for sure though. IIRC I fixed the issues in the opposite order for mcdisplay.pl.

[willend]

Right, could easily be. Let's have a look next week during release camp? (And decide if we do anything about it there - there is a solution if the Perl junk is used)

[willend]

A new set of binaries are available for test at http://tmp.mccode.org/Windows/

[willend]

Conclusion?

[willend]

@trnielsen if you would eventually have a look here it would be great :-)