CRAN compilation failed on Windows i386

[Jean-Romain]

@floriandeboissieu: The windows builder gives an error for rlas 1.2.0 on Windows for 32 bits architecture. https://win-builder.r-project.org/mQ1z5qA5JIY5/. This error is likely in testthat. I can't figure out what is the error. Do you have any idea? A way to reproduce?

[floriandeboissieu]

With so little clue in the log... I'll check tomorrow if I can find something

Cheers

On 03/04/2018 19:04, Jean-Romain Roussel wrote:

@floriandeboissieu https://github.com/floriandeboissieu: The windows builder gives an error for rlas 1.2.0 on Windows for 32 bits architecture. https://win-builder.r-project.org/mQ1z5qA5JIY5/. This error is likely in testthat. I can't figure out what is the error. Do you have any idea? A way to reproduce?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/Jean-Romain/rlas/issues/17, or mute the thread https://github.com/notifications/unsubscribe-auth/AP54IfeFJMi0tH8vbZ93SQdphD45ec_pks5tk6ukgaJpZM4TFcWs.

[floriandeboissieu]

Anything with rhub check?

On 05/04/2018 00:39, Florian de Boissieu wrote:

With so little clue in the log... I'll check tomorrow if I can find something

Cheers

On 03/04/2018 19:04, Jean-Romain Roussel wrote:

@floriandeboissieu https://github.com/floriandeboissieu: The windows builder gives an error for rlas 1.2.0 on Windows for 32 bits architecture. https://win-builder.r-project.org/mQ1z5qA5JIY5/. This error is likely in testthat. I can't figure out what is the error. Do you have any idea? A way to reproduce?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/Jean-Romain/rlas/issues/17, or mute the thread https://github.com/notifications/unsubscribe-auth/AP54IfeFJMi0tH8vbZ93SQdphD45ec_pks5tk6ukgaJpZM4TFcWs.

[Jean-Romain]

Yep an rhub does not provide more informations. It failed after 1h30.

[Jean-Romain]

https://builder.r-hub.io/status/rlas_1.2.0.tar.gz-42dbc6ac7089497d901ef8aa630423c4

[Jean-Romain]

I sequentially uploaded 3 versions

1. All the code in testthat was commented: :heavy_check_mark:
2. All the code in testthat was commented but test-head_tools.R : :x:
3. All the code in testthat was commented but test-readlas.R : :heavy_check_mark:

So the problem seems to come from test-head_tools.R. I slowily investigate. One clue each time the CRAN compile the package i.e. once each 15 minutes :smile:

The big deal is that is that header_tools is pure R stuff...

[Jean-Romain]

The issue occurs at several place but one of them is ligne 18 of test_header_tools.R

wheader = read.lasheader(write_path)

However line 5 of the same file this one works

header = read.lasheader(lazfile)

[floriandeboissieu]

I started the same yesterday, commenting all but:

 - readlas tests -> works

 - writelas tests -> doesn't work

I commented all writelas including back progressively the tests.

On 05/04/2018 21:37, Jean-Romain Roussel wrote:

The issue occurs at several place but one of them is ligne 18 of |test_header_tools.R|

wheader = read.lasheader(write_path)

However line 5 of the same file this one works

header = read.lasheader(lazfile)

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/Jean-Romain/rlas/issues/17#issuecomment-379052566, or mute the thread https://github.com/notifications/unsubscribe-auth/AP54IddC6Dur-X0JFVBhTb6lFqM48kZMks5tlnKCgaJpZM4TFcWs.

[floriandeboissieu]

I narrowed to line 43 of test-writelas:

header = read.lasheader(lazfile)

I believe it has something to do with vlrsreader function.

Maybe, it could has something to do with I64 type definition, in inst/include/mydefs.hpp (line 61-67)

I'm sorry I cannot go on today...

Hopping it helps

On 06/04/2018 07:38, Florian de Boissieu wrote:

I started the same yesterday, commenting all but:

- readlas tests -> works

- writelas tests -> doesn't work

I commented all writelas including back progressively the tests.

On 05/04/2018 21:37, Jean-Romain Roussel wrote:

The issue occurs at several place but one of them is ligne 18 of |test_header_tools.R|

wheader = read.lasheader(write_path)

However line 5 of the same file this one works

header = read.lasheader(lazfile)

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/Jean-Romain/rlas/issues/17#issuecomment-379052566, or mute the thread https://github.com/notifications/unsubscribe-auth/AP54IddC6Dur-X0JFVBhTb6lFqM48kZMks5tlnKCgaJpZM4TFcWs.

[Jean-Romain]

You're right. I reached the same conclusion.

This works

lazfile <- system.file("extdata", "example.laz", package="rlas")
header = read.lasheader(lazfile)

And this don't

lazfile <- system.file("extdata", "extra_byte.laz", package="rlas")
header = read.lasheader(lazfile)

[Jean-Romain]

I'm slowly converging. The bug comes from vlrsreader (you were right but I checked it by security). readheader.cpp has not been modified a lot thus I'm pretty sure it comes from this commit: fc0b1419