Commit the files that should have gone into #8612

[joa-quim]

(Fix the variable transparency case that was broken by the #8612 commit.)

[seisman]

The CI jobs tell me that test/grdimage/transp_mix.sh still passes, but test/grdimage/image_vartrans.sh fails.

[joa-quim]

I failed to commit also the modified tests. You should now see the opposite situation. test/grdimage/transp_mix.sh fails (because I removed the KNOWN_FAILURE) and test/grdimage/image_vartrans.sh passes. What I ask you is to replace the transp_mix.ps original by the new one.

[seisman]

I'm on macOS. Both fail for me.

test/grdimage/transp_mix.sh:

test/grdimage/image_vartrans.sh

[seisman]

It's like something different with macOS preview.

Using the Adobe PDF Reader, I've

[joa-quim]

But what do you get when running the tests? I get

C:\v\build>ctest -R image_vartrans
Test project C:/v/build
    Start 523: test/grdimage/image_vartrans.sh
1/1 Test #523: test/grdimage/image_vartrans.sh ...   Passed    3.44 sec

100% tests passed, 0 tests failed out of 1

Total Test time (real) =   3.47 sec

C:\v\build>ctest -R transp_mix
Test project C:/v/build
    Start 536: test/grdimage/transp_mix.sh
1/1 Test #536: test/grdimage/transp_mix.sh ......***Failed    5.17 sec

[seisman]

 ctest -R test/grdimage/image_vartrans.sh
Test project /Users/seisman/OSS/gmt/gmt/build/debug
    Start 532: test/grdimage/image_vartrans.sh
1/1 Test #532: test/grdimage/image_vartrans.sh ...***Failed   14.86 sec

[joa-quim]

Did you update for this PR? I latter added the modified scripts. Note that now to have variable transparency we must now use the option -Q+t. i.e., gmt grdimage rgba.tif -Q+t

[seisman]

Here is what I have in my log file:

Start testing: Nov 11 22:06 CST
----------------------------------------------------------
532/1186 Testing: test/grdimage/image_vartrans.sh
532/1186 Test: test/grdimage/image_vartrans.sh
Command: "/bin/bash" "gmtest" "test/grdimage/image_vartrans.sh"
Directory: /Users/seisman/OSS/gmt/gmt/build/debug/test
"test/grdimage/image_vartrans.sh" start time: Nov 11 22:06 CST
Output:
----------------------------------------------------------
Set GMT_SESSION_NAME = 70298
/usr/local/bin/gm compare: image difference exceeds limit (0.0140282 > 0.003).
ERROR: gmtest:171
memtrack errors: 0
exit status: 1
test/grdimage/image_vartrans.ps: RMS Error = 0.0140 [FAIL]
memtrack errors: 0
exit status: 1
<end of output>
Test time =  15.60 sec
----------------------------------------------------------
Test Failed.
"test/grdimage/image_vartrans.sh" end time: Nov 11 22:06 CST
"test/grdimage/image_vartrans.sh" time elapsed: 00:00:15
----------------------------------------------------------

End testing: Nov 11 22:06 CST

[joa-quim]

/Users/seisman/OSS/gmt/gmt/build/debug

Hmm, my runs were done with a release build. Could it be the reason?

[seisman]

Did you update for this PR? I latter added the modified scripts. Note that now to have variable transparency we must now use the option -Q+t. i.e., gmt grdimage rgba.tif -Q+t

I've updated to the latest version.

/Users/seisman/OSS/gmt/gmt/build/debug

Hmm, my runs were done with a release build. Could it be the reason?

I also tried the release build. The image is still different from the baseline image, so it fails.

[joa-quim]

Can't make sense out of this. I am using GMT_DATA_SERVER static at compile time (the only one that works for me). And before this PR the variable transparency had to fail always because code use completely broken.

[joa-quim]

How can we have this in the original PS? It's impossible. Nothing in the image_vartrans.sh puts a BACKGROUND word as a kind of watermark. And this disappears when we convert the orig in .png or .pdf

[seisman]

The "BACKGROUND" word is likely added by macOS Preview. I don't see it in the Adobe PDF reader.

[joa-quim]

Yes, likely but the SC I posted was taken in my computer as seen from Ghostscript.
The CI's help nothing. image_vartrans fails on Linux and passes on Mac & Windows.

[joa-quim]

Any reason to not to merge this PR?

[github-actions]

Summary of changed images

This is an auto-generated report of images that have changed on the DVC remote

Status	Path
modified	test/baseline/grdimage/
modified	test/baseline/grdimage/transp_mix.ps

Image diff(s)

Added images

Modified images

Path	Old	New
test/baseline/grdimage/transp_mix.ps

Report last updated at commit d0c540b

[seisman]

I've updated the baseline image for transp_mix.sh and the image comparison is available at #8613 (comment). The old and new images seem almost the same, except that the new image looks more blurry than the old one. Is it expected?

[joa-quim]

I can’t see a difference between the two. And note that only the second row, first col subplot has changed.

[joa-quim]

When I do a dvc pull my transp_mix.ps file does not get updated and dvc tells me that everything is Everything is up to date.

We need to find a solution/replacement to this DVC thing. I simply can't get it to work.

[seisman]

What's the dvc version you're using?

[joa-quim]

C:\progs_cygw\GMTdev\gmt5\master>dvc --version
3.28.0

But note that all 3 OSs fail this CI test.