-
Notifications
You must be signed in to change notification settings - Fork 348
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Let 6.1 have more flexibility to discover new remote data sets #3370
Conversation
Having this as a variable makes it easier to test features by changing the hash name.
This branch fails on Windows:
|
Perhaps @seisman can think of ways to update gmt_getremote.sh once we are not just serving earth_relief files? How do we find out that an entry in gmt_hash_server.txt is a cache file vs data file when we cannot use earth_relief? |
I added this too Per the web that should work for Windows to identify directories. |
OK. I'll try and open a PR for it. |
|
Working on this as part of the update. Is this what you expect from gmtwhich -G:
Of course, if the file already exists in cache or server/* we just copy to the requested place, but if the files do not exist locally, we do not place them in their natural directories as we do in all other situations. gmt which is special and only allows specific placements. Is this how you interpret gmtwhich? |
…ools/gmt into new-remote-files
Is the testing stuck? |
The CI is OK, but github didn't get the reports. |
Do we wait? |
The CI reports are here https://dev.azure.com/GenericMappingTools/GMT/_build/results?buildId=8258&view=results. No need to wait. |
I dont understand. I says several of the test failed? |
There is a segment fault for command
after commit 4e7c9b1. I think it's because GMT is using the old gmt_data_server.txt file. The error message is:
|
Yes, since I removed the special case that let us read the old format gmt_data_server.c. Perhaps we should move on a merge, then check again for issues |
Agreed. |
I will now change the data pointer on the server. |
gmtserver data now pointing to the new data_6.1 directory. Will build master and see if OK, then run 6.0. |
OK, local test with master still work as before (minus ex52.sh and gdal_nn.sh which crashes when run in ctest). |
How might I switch to 6.0.0 and run all tests locally? THere is no 6.0.0 branch (I think) and the tarball does not have the tests. There is probably a 6.0.0 tag but my GUI Git Desktop does not seem to know about tags? |
For command line, run |
That gave me an empty test dir so no tests could be run... |
Fails. This fails too:
|
|
That is pretty interesting. That is not even a GMT command so not affected by versions. I get a tiny file that contains
|
Sorry, I also got the "404 not found" HTML page, not the grid. |
Did you use |
He-he-he..... |
With GMT 6.0.0 and blank .gmt
due to downloading the 15s 2.6 Gb grid. With 6.1,
That is why are doing the tiling. |
|
I guess the only way it could work for 6.1 is to add another entry for 60m. I link won't do since it gives up by not finding it in the table. But it is not even listed on our dataset page - do we really need to go there? |
For some background, see GenericMappingTools/gmtserver-admin#37.
This branch will implement the changes needed for 6.1. First change is to replace the hardwired "gmt_hash_server.txt" name with a constant:
#define GMT_HASH_SERVER_FILE "gmt_hash_server.txt"
Using a arameter for data hash filename makes it easier to test features by temporarily changing the hash name.
I expect the hash name to remain the same upon release though. We want GMT <= 6.0.0 to function exactly as before, except get the updated grids for the data types known to <= 6.0.0. However, 6.1 will have access to more files.