WIP: Fix GSHHGDIR and DCWDIR

[seisman]

@PaulWessel Can you check if this works for you.

[PaulWessel]

OK trying now.

[seisman]

In this function, GMT tries to find DCW in GMT->session.DCWDIR (if set) first, then search via gmt_getsharepath.

gmt/src/gmt_dcw.c

Lines 71 to 77 in f75038a

	GMT_LOCAL bool dcw_get_path (struct GMT_CTRL *GMT, char *name, char *suffix, char *path) {
	bool found = false;
	/* This is the order of checking:
	* 1. Check in GMT->session.DCWDIR, if set
	* 2. Look via gmt_getsharepath.
	*/

When COPY_DCW is set to TRUE, we know that DCW data must be in SHAREDIR/dcw, so we should skip searching in GMT->session.DCWDIR.

Now the problem is, GMT->session.DCWDIR is always set to DCW_INSTALL_PATH:

gmt/src/gmt_init.c

Lines 5473 to 5475 in f75038a

	if (GMT->session.DCWDIR)
	gmt_M_str_free (GMT->session.DCWDIR);
	GMT->session.DCWDIR = strdup (DCW_INSTALL_PATH);

And DCW_INSTALL_PATH is always set to DCW_PATH in

gmt/src/config.h.in

Lines 68 to 69 in f75038a

	/* Name of DCW path */
	#define DCW_INSTALL_PATH "@DCW_PATH@"

.

[PaulWessel]

Your first change worked for me. Do I need to try ff7f807 as well?

[seisman]

The first commit breaks CI, because we don't set "COPY_GSHHG" and "COPY_DCW" in our CI.

[PaulWessel]

We still need this merged too, no?

[seisman]

No. #983 already works. This PR tries to make it better.

I think we can leave it after rc1.

[codecov-io]

Codecov Report

Merging #984 into 6.0 will decrease coverage by 1.84%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##              6.0     #984      +/-   ##
==========================================
- Coverage   58.27%   56.42%   -1.85%     
==========================================
  Files         230      230              
  Lines      132468   132469       +1     
==========================================
- Hits        77201    74751    -2450     
- Misses      55267    57718    +2451

Impacted Files	Coverage Δ
src/gmt_init.c	58.64% <ø> (-0.36%)	⬇️
src/testapi_matrix.c	0% <0%> (-100%)	⬇️
src/testapi_vector.c	0% <0%> (-100%)	⬇️
src/pssolar.c	0% <0%> (-63.86%)	⬇️
src/kml2gmt.c	0% <0%> (-56.37%)	⬇️
src/grdlandmask.c	24.48% <0%> (-48.69%)	⬇️
src/gmt2kml.c	0% <0%> (-36.48%)	⬇️
src/mgd77/mgd77track.c	0% <0%> (-27.52%)	⬇️
src/gmt_shore.c	50.22% <0%> (-27.12%)	⬇️
src/pscoast.c	46.74% <0%> (-18.5%)	⬇️
... and 35 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d6ac126...aaea978. Read the comment docs.

[joa-quim]

Ping on this. Implement it (after conflict solving)? Looks OK to me.