Issue running examples (edgefx/featfx error).

[chlnddev]

Issue running examples (edgefx/featfx error).

[krober10nd]

Hey Leo, What happens when you type plot(gdat) Do you see New Zealand? Are you using m_mapv1.4 (latest from the website)? This error means your bbox doesn't intersect with the shoreline dataset.

…

On Thu, Jul 18, 2019, 10:14 PM Leo ***@***.***> wrote: Hi, I have been following the guide and I am trying to run through the examples. I have installed m_map in the root folder and put the 'GSHHS_f_L1' shapefile in the datasets/ folder. However I am running into an issue, the following error occurs when running Example_1_NZ.m: Reading shapefile with m_shaperead 179837 Records in file, by 500: .......................................................................................................................................................................................................................................................................................................................................................................Partitioning the boundary into islands, mainland, ocean Smoothing coastline with 5 point window Reading shapefile with m_shaperead 5706 Records in file, by 500: ...........Partitioning the boundary into islands, mainland, ocean 'Read in meshing boundary: ' 'GSHHS_f_L1' Building feature size function... Warning: No mainland or inner In edgefx/featfx (line 262) In edgefx (line 212) In Example_1_NZ (line 20) Error using & Matrix dimensions must agree. Error in edgefx/finalize (line 775) hh_m(nearshore & hh_m > obj.max_el_ns) = obj.max_el_ns; Error in edgefx (line 246) obj = finalize(obj,feat); Error in Example_1_NZ (line 20) fh = edgefx('geodata',gdat,... — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=AEOBZ7DGZ3FUFATQOWVHIYDQAEIPLA5CNFSM4IFBLLKKYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HAE7GXA>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEOBZ7ELPUW3RFWUDC4JHSDQAEIPLANCNFSM4IFBLLKA> .

[DrakonianMight]

HI Keith,

Yes I downloaded m_map1.4, unzipped it, renamed the top directory m_map and put in the root folder today.
I get the following when I run plot(gdat) :

Undefined function 'startsWith' for input arguments of type 'char'.

Error in geodata/plot (line 685)
            if startsWith(projection,'ste')

[krober10nd]

Ok, you must be using an earlier version of MATLAB. Please replace the conditional if statement with "startsWith" in geodata.m to startsWith(projection,'ste') ~isempty(regexp(projtype,'ste'))

…

On Thu, Jul 18, 2019 at 10:29 PM Leo ***@***.***> wrote: HI Keith, Yes I downloaded m_map1.4, unzipped it, renamed the top directory m_map and put in the root folder today. I get the following when I run plot(gdat) : Undefined function 'startsWith' for input arguments of type 'char'. Error in geodata/plot (line 685) if startsWith(projection,'ste') — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=AEOBZ7BJZF766WRCPYQCBO3QAEKI7A5CNFSM4IFBLLKKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2KJLII#issuecomment-513054113>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEOBZ7BYO452YTH72TX3YS3QAEKI7ANCNFSM4IFBLLKA> .

-- -------------------- Keith Roberts Ph.D. (Eng.) University of Notre Dame [email protected]

[DrakonianMight]

Yes, looks like I am running 2016a and startsWith was introduced in 2016b.

Thanks for sending that fix through, I applied it, however I get a further error:

Building feature size function...
man, no medial points!
Error using  & 
Matrix dimensions must agree.

Error in edgefx/finalize (line 775)
                hh_m(nearshore & hh_m > obj.max_el_ns) = obj.max_el_ns;

Error in edgefx (line 246)
                    obj = finalize(obj,feat);

Error in Example_1_NZ (line 20)
fh = edgefx('geodata',gdat,...
 
>> plot(gdat)
Undefined function or variable 'projtype'.

Error in geodata/plot (line 685)
            if ~isempty(regexp(projtype,'ste'))

If its any help if I replace 'projtype' with 'projection' I get an error but plot(gdat) works OK.
Thanks for getting back to me so quickly.

[krober10nd]

No problem. Yea, sorry it was proj not projtype. But what does the image look like?

…

On Thu, Jul 18, 2019 at 10:51 PM Leo ***@***.***> wrote: Yes, looks like I am running 2016a and startsWith was introduced in 2016b. Thanks for sending that fix through, I applied it, however I get a further error: Building feature size function... man, no medial points! Error using & Matrix dimensions must agree. Error in edgefx/finalize (line 775) hh_m(nearshore & hh_m > obj.max_el_ns) = obj.max_el_ns; Error in edgefx (line 246) obj = finalize(obj,feat); Error in Example_1_NZ (line 20) fh = edgefx('geodata',gdat,... >> plot(gdat) Undefined function or variable 'projtype'. Error in geodata/plot (line 685) if ~isempty(regexp(projtype,'ste')) If its any help if I replace 'projtype' with 'projection' I get an error but plot(gdat) works OK. Thanks for getting back to me so quickly. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=AEOBZ7HFROS42J6XOJYL4I3QAEM3NA5CNFSM4IFBLLKKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2KKLVQ#issuecomment-513058262>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEOBZ7D6EA6PEGGWEPSFGYDQAEM3NANCNFSM4IFBLLKA> .

-- -------------------- Keith Roberts Ph.D. (Eng.) University of Notre Dame [email protected]

[DrakonianMight]

So If I replace line 685 in geodata/plot to if ~isempty(regexp(projection,'ste')) (proj didn't work)

I still get:

Building feature size function...
man, no medial points!
Error using  & 
Matrix dimensions must agree.

Error in edgefx/finalize (line 775)
                hh_m(nearshore & hh_m > obj.max_el_ns) = obj.max_el_ns;

Error in edgefx (line 246)
                    obj = finalize(obj,feat);

Error in Example_1_NZ (line 20)
fh = edgefx('geodata',gdat,...

but running plot(gdat) gives me

[krober10nd]

Thanks, this is interesting. Did you modify the example script at all? Which branch did you download and are using?

…

On Thu, Jul 18, 2019 at 10:14 PM Leo ***@***.***> wrote: Hi, I have been following the guide and I am trying to run through the examples. I have installed m_map in the root folder and put the 'GSHHS_f_L1' shapefile in the datasets/ folder. However I am running into an issue, the following error occurs when running Example_1_NZ.m: Reading shapefile with m_shaperead 179837 Records in file, by 500: .......................................................................................................................................................................................................................................................................................................................................................................Partitioning the boundary into islands, mainland, ocean Smoothing coastline with 5 point window Reading shapefile with m_shaperead 5706 Records in file, by 500: ...........Partitioning the boundary into islands, mainland, ocean 'Read in meshing boundary: ' 'GSHHS_f_L1' Building feature size function... Warning: No mainland or inner In edgefx/featfx (line 262) In edgefx (line 212) In Example_1_NZ (line 20) Error using & Matrix dimensions must agree. Error in edgefx/finalize (line 775) hh_m(nearshore & hh_m > obj.max_el_ns) = obj.max_el_ns; Error in edgefx (line 246) obj = finalize(obj,feat); Error in Example_1_NZ (line 20) fh = edgefx('geodata',gdat,... — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=AEOBZ7DGZ3FUFATQOWVHIYDQAEIPLA5CNFSM4IFBLLKKYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HAE7GXA>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEOBZ7ELPUW3RFWUDC4JHSDQAEIPLANCNFSM4IFBLLKA> .

-- -------------------- Keith Roberts Ph.D. (Eng.) University of Notre Dame [email protected]

[DrakonianMight]

Hi, No edits to the original Example_1_NZ.m file.
I downloaded and was using the Projection branch.
I have also tried using the Master and I get an error (perhaps the below is of more use (from running from master).

Building feature size function...
Invalid MEX-file '/home/matlab/oceanwaves2d/OceanMesh2D-master/OceanMesh2D-master/@ann/private/annmex.mexa64': /lib64/libc.so.6: version `GLIBC_2.14' not found (required by
/home/matlab/oceanwaves2d/OceanMesh2D-master/OceanMesh2D-master/@ann/private/annmex.mexa64)

Error in get_data_class (line 3)
d = annmex();

Error in ann (line 108)
[anno.ccls anno.cfun] = get_data_class();

Error in WrapperForKsearch (line 10)
anno = ann(dataset);

Error in dpoly (line 26)
[~,d] = WrapperForKsearch(pv1', p',1);

Error in edgefx/get_dis (line 296)
            d = feval(obj.fd,obj,feat);

Error in edgefx/featfx (line 226)
            [d,obj] = get_dis(obj,feat);

Error in edgefx (line 179)
                            obj = featfx(obj,feat);

Error in Example_1_NZ (line 19)
fh = edgefx('geodata',gdat,...

Thanks again.

[krober10nd]

Yes, the "master" branch on there is obsolete now. I can't replicate the error on my machine with R2016b using no toolboxes. Can we do a sanity check and completely fully close MATLAB, reopen it and make sure you have no conflicting installations of OceanMesh in the path?

…

On Thu, Jul 18, 2019 at 11:29 PM Leo ***@***.***> wrote: Hi, No edits to the original Example_1_NZ.m file. I downloaded and was using the Projection branch. I have also tried using the Master and I get an error (perhaps the below is of more use (from running from master). Building feature size function... Invalid MEX-file ***@***.***/private/annmex.mexa64': /lib64/libc.so.6: version `GLIBC_2.14' not found (required by ***@***.***/private/annmex.mexa64) Error in get_data_class (line 3) d = annmex(); Error in ann (line 108) [anno.ccls anno.cfun] = get_data_class(); Error in WrapperForKsearch (line 10) anno = ann(dataset); Error in dpoly (line 26) [~,d] = WrapperForKsearch(pv1', p',1); Error in edgefx/get_dis (line 296) d = feval(obj.fd,obj,feat); Error in edgefx/featfx (line 226) [d,obj] = get_dis(obj,feat); Error in edgefx (line 179) obj = featfx(obj,feat); Error in Example_1_NZ (line 19) fh = edgefx('geodata',gdat,... Thanks again. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=AEOBZ7CIYAAYTPAJYINMKVTQAERHRA5CNFSM4IFBLLKKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2KMFVQ#issuecomment-513065686>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEOBZ7CQVMJZBM7L6ISIRS3QAERHRANCNFSM4IFBLLKA> .

-- -------------------- Keith Roberts Ph.D. (Eng.) University of Notre Dame [email protected]

[DrakonianMight]

Hi,

Yes I have restarted and can confirm still getting the same issue.

It seems to be linked to the use of '&' in line 775 of edgefx.m

On doing some debugging it seems that 'nearshore' is empty which is sourced from obj.boudist (which is also empty).

Any ideas?

[krober10nd]

Leo, It's indicating you have no shoreline geometry in your meshing domain, so it can't calculate the variable nearshore to enforce the element size bounds in edgefx.m Something is odd here. Your geodata has data and you pass it to the edgefx class constructor. Could you wrap your geodata class instance gdat in a cellarray in the call to edgefx? "geodata",{gdat} Matlab changed around the treatment of cell arrays in some precious versions... I would also put a breakpoint inside the name value arg parser when it tries to parse the geodata class object and see what it sees.

…

On Fri, Jul 19, 2019, 12:50 AM Leo ***@***.***> wrote: Hi, Yes I have restarted and can confirm still getting the same issue. It seems to be linked to the use of '&' in line 775. On doing some debugging it seems that 'nearshore' is empty which is sourced from obj.boudist (which is also empty). Any ideas? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=AEOBZ7BJBI65LVU664647XDQAE2XTA5CNFSM4IFBLLKKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2KPV5Q#issuecomment-513080054>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEOBZ7BGVDRMOINHSJCWPODQAE2XTANCNFSM4IFBLLKA> .

[WPringle]

It could be the issue that old matlab does not automatically resize matrices to fit dimensions when making logical statements.

nearshore = abs(obj.boudist) < 100; size is M x N
hh_m(nearshore & hh_m > obj.max_el_ns) = obj.max_el_ns; obj.max_el_ns, size of obj.max_el_ns is 1 x 1, size of hh_m is M x N

try changing hh_m > obj.max_el_ns to _hh_m > repmat(obj.max_el_ns,size(hh_m))

Although it doesn't explain why obj.boudist is empty

[WPringle]

Actually I see now that if you are still getting the "No mainland or inner" warning on line 262 then it will not calculate the obj.boudist so you are just getting an error at line 775 but the issue is happening before that because of the no mainland or inner problem.

[DrakonianMight]

Hi both,
Thanks for looking at this issue.

So as suggested I wrapped the gdat in a cellarray and added a breakpoint to look at gdat.
Sadly still getting an error.
Gdat properties:

The error:

feat looks like it is defined in line 144, in some sort of default value constructor.

[krober10nd]

Leo, Never seen this before. I have the same version of Matlab and no toolboxes too.. Try stepping through edgefx and find out why the feat.h0 isn't passed through to edgefx.

…

On Sun, Jul 21, 2019, 7:37 PM Leo ***@***.***> wrote: Hi both, Thanks for looking at this issue. So as suggested I wrapped the gdat in a cellarray and added a breakpoint to look at gdat. Sadly still getting an error. Gdat properties: [image: image] <https://user-images.githubusercontent.com/46990198/61597615-11f5d800-ac56-11e9-8737-f6879e7db683.png> The error: [image: image] <https://user-images.githubusercontent.com/46990198/61597626-2afe8900-ac56-11e9-9d8f-0b967a6a7b8c.png> feat looks like it is defined in line 144, in some sort of default value constructor. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=AEOBZ7BQVSGXLBGYBIOSKQTQATQKPA5CNFSM4IFBLLKKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2ONBWA#issuecomment-513593560>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEOBZ7CSMT2HQR6OUF63J6LQATQKPANCNFSM4IFBLLKA> .

[DrakonianMight]

Thanks for the help,

I am currently investigation getting a newer release of Matlab (2019), which I am helping with solve the issue. Before I proceed can you confirm that OceanMesh2D is functional on the 2019 version?

Thanks again,

[krober10nd]

Leo, Yes I test the software on windows and Mac each pull request on a variety of Matlab versions. I still don't understand how you can't run example 1. Did you try exploring changing the options or trying different examples? What commit did you use exactly?

…

On Tue, Jul 23, 2019, 1:01 AM Leo ***@***.***> wrote: Thanks for the help, I am currently investigation getting a newer release of Matlab (2019), which I am helping with solve the issue. Before I proceed can you confirm that OceanMesh2D is functional on the 2019 version? Thanks again, — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=AEOBZ7CHMIZKSVHI3O4TCFTQAZ7BNA5CNFSM4IFBLLKKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2R3I2Y#issuecomment-514045035>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEOBZ7DOSQVVXKTNH6FSJILQAZ7BNANCNFSM4IFBLLKA> .

[DrakonianMight]

I was using the Projection commit.

I did some testing and I still wasn't able to get it to run for example 1 or 2.

I also tried wrapping the object and making minor changes in edgfx, but I kept running into other issues. We were always planning on moving to a more recent version of Matlab anyway, if it solves the issue for me here that is a bonus.

[krober10nd]

Projection is the branch. Which commit? What's the error with example 2? Did you install breaks points in what was happening in the name/value parser in edgefx.m? I can't help without being able to replicate the error given the same workstation configuration.

…

On Tue, Jul 23, 2019, 1:47 AM Leo ***@***.***> wrote: I was using the Projection commit. I did some testing and I still wasn't able to get it to run for example 1 or 2. I also tried wrapping the object and making minor changes in edgfx, but I kept running into other issues. We were always planning on moving to a more recent version of Matlab anyway, if it solves the issue for me here that is a bonus. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=AEOBZ7CKXXAKCOJJ6GI74PLQA2EODA5CNFSM4IFBLLKKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2R5B7I#issuecomment-514052349>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEOBZ7FXVTM3HMLK6MXYB3LQA2EODANCNFSM4IFBLLKA> .

[DrakonianMight]

Apologies for taking so long to get back to you, I had some other priorities and I was awaiting our upgrade to a newer version of Matlab.

I have now started testing on 2019a, however unfortunately I am still getting the same error.

``179837` Records in file, by 500:

.......................................................................................................................................................................................................................................................................................................................................................................Partitioning the boundary into islands, mainland, ocean
Smoothing coastline with 5 point window
Reading shapefile with m_shaperead
5706 Records in file, by 500:

...........Partitioning the boundary into islands, mainland, ocean
'Read in meshing boundary: ' 'GSHHS_f_L1'

Building feature size function...
man, no medial points!
Error using &
Matrix dimensions must agree.

Error in edgefx/finalize (line 775)
hh_m(nearshore & hh_m > obj.max_el_ns) = obj.max_el_ns;

Error in edgefx (line 246)
obj = finalize(obj,feat);

Error in Example_1_NZ (line 19)
fh = edgefx('geodata',gdat,...``

I am using projection branch, commit: e434912

I have downloaded and placed m_map in the root folder and I do not have gdal installed.

[krober10nd]

I can't reproduce the error on their machine so there's no way to help you unless you present some more information other than indicating it doesn't work. Could you give me the full working directory as a zip file with the software, the data, and the scripts as you're executing it? Did you ensure the full m_map path is visible at the root directory using the genpath command as is described in the user guide?

…

On Sun, Jul 28, 2019 at 7:23 PM Leo ***@***.***> wrote: Apologies for taking so long to get back to you, I had some other priorities and I was awaiting our upgrade to a newer version of Matlab. I have now started testing on 2019a, however unfortunately I am still getting the same error. ``179837` Records in file, by 500: .......................................................................................................................................................................................................................................................................................................................................................................Partitioning the boundary into islands, mainland, ocean Smoothing coastline with 5 point window Reading shapefile with m_shaperead 5706 Records in file, by 500: ...........Partitioning the boundary into islands, mainland, ocean 'Read in meshing boundary: ' 'GSHHS_f_L1' Building feature size function... man, no medial points! Error using & Matrix dimensions must agree. Error in edgefx/finalize (line 775) hh_m(nearshore & hh_m > obj.max_el_ns) = obj.max_el_ns; Error in edgefx (line 246) obj = finalize(obj,feat); Error in Example_1_NZ (line 19) fh = edgefx('geodata',gdat,...`` I am using projection branch, commit: e434912 <e434912> I have downloaded and placed m_map in the root folder and I do not have gdal installed. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=AEOBZ7AILTY6AJMM32CO5LDQBYL4VA5CNFSM4IFBLLKKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD27HZPA#issuecomment-515800252>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEOBZ7CWLIIIOHBQPD5F7S3QBYL4VANCNFSM4IFBLLKA> .

-- -------------------- Keith Roberts Ph.D. (Eng.) University of Notre Dame [email protected]

[DrakonianMight]

I have zipped my directory (link below), to confirm I am running the linux edition of matlab.

https://send.firefox.com/download/ca60087ab82ec052/#KsUCxOg-N9C5ER8vcPnx0g

[krober10nd]

I just ran your exact code with the inputs and unfortunately for this issue, it worked just fine.

All I did is was right click the working directory and click Add to path --> Selected folders and subfolders.

I tried with and without the paid mapping toolbox and they both operated fine.

[DrakonianMight]

Well I guess that suggests that it might be a linux version of Matlab problem.

We have dug through it and isolated where it fails, get_dist (line371 in edfefx.m) fails specifically at line 374, when @dpoly is called. Sadly this fails silently, despite adding a number of comments throughout.

Going to track down a windows version in the next couple of days and try some testing on it.

[krober10nd]

Yes, it must be related to the hardware of the machine since everything I see in your file appears to be correct. What version of Linux are you using?

…

On Sun, Jul 28, 2019 at 10:00 PM Leo ***@***.***> wrote: Well I guess that suggests that it might be a linux version of Matlab problem. We have dug through it and isolated where it fails, get_dist (line371 in edfefx.m) fails specifically at line 374, when @dpoly is called. Sadly this fails silently, despite adding a number of comments throughout. Going to track down a windows version in the next couple of days and try some testing on it. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=AEOBZ7BRIAK4XF4OSAGAY3LQBY6JLA5CNFSM4IFBLLKKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD27KOOY#issuecomment-515811131>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEOBZ7A3KET6LMBCVIOOYYDQBY6JLANCNFSM4IFBLLKA> .

-- -------------------- Keith Roberts Ph.D. (Eng.) University of Notre Dame [email protected]

[DrakonianMight]

Scientific Linux, perhaps not one of the common distributions, but its based on Red Hat.

[krober10nd]

Ok Leo. Perhaps we can schedule a call over skype to discuss this issue using the debugger and the share screen option? Let me know.

…

On Sun, Jul 28, 2019 at 10:09 PM Leo ***@***.***> wrote: Scientific Linux, perhaps not one of the common distributions, but its based on Red Hat. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=AEOBZ7CYOSF3O53BZRZDXETQBY7MNA5CNFSM4IFBLLKKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD27KVPI#issuecomment-515812029>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEOBZ7FM5EXRYH72VV5UNO3QBY7MNANCNFSM4IFBLLKA> .

-- -------------------- Keith Roberts Ph.D. (Eng.) University of Notre Dame [email protected]

[HamishB]

What version of Scientific Linux are you using?

[HamishB]

I've just tried Example_1_NZ.m from the git repo on CentOS 7.6 with Matlab R2016b and it all worked ok. The only change I made to the code was changing the line endings from DOS to UNIX in @geodata/geodata.m.

I did get a couple warnings:
Warning: no polyshape or polybool available to merge possible overlapping of mainland and inner

---

When I tried using the min_el, max_el, etc. terms from the July 4th user guide everything took quite a bit longer and used a bunch more RAM (peak >3.5GB). But it worked.

Also to note it left a Temp_grid.mat file behind, don't know if that is desired.

regards,
Hamish

[krober10nd]

Hamish, The memory overhead depends on the choice of the minimum element size. This is why we allow you to local high resolution domains with the multiscale technique. The output of temp grid is expected, it saves the temporary mesh in case something goes awry.

…

On Fri, Aug 2, 2019, 2:10 AM HamishB ***@***.***> wrote: I've just tried Example_1_NZ.m from the git repo on CentOS 7.6 with Matlab R2016b and it all worked ok. The only change I made to the code was changing the line endings from DOS to UNIX in @geodata/geodata.m. I did get a couple warnings: Warning: no polyshape or polybool available to merge possible overlapping of mainland and inner ------------------------------ When I tried using the min_el, max_el, etc. terms from the July 4th user guide everything took a quite a bit longer and used a bunch more RAM (peak >3.5GB). But it worked. Also to note it left a Temp_grid.mat file behind, don't know if that is desired. regards, Hamish — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#19?email_source=notifications&email_token=AEOBZ7D23VZGH42I23FD3VLQCO6STA5CNFSM4IFBLLKKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3MTI3Y#issuecomment-517551215>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AEOBZ7BV5OT7FGFWPXZLXPLQCO6STANCNFSM4IFBLLKA> .

[HamishB]

By the way, if anyone wants a higher resolution coastline shapefile and bathy data for NZ:

https://koordinates.com/from/data.linz.govt.nz/layer/51153/

(250m res is aspirational in places)
https://koordinates.com/layer/8678-niwa-new-zealand-bathymetic-grid-2016/

both licensed as CC-By.

[DrakonianMight]

FYI I got everything working as expected on a fresh Matlab install on Windows.
It seems that there is a missing dependency on some Linux distributions (Redhat and Scientific Linux) which is causing the issue, sadly not something I was able to change on the HPC I was using. I will try and get some more details.

Thanks for the support and for responding quickly.

[krober10nd]

Thanks DrakonianMight. Glad you got it working!