-
Notifications
You must be signed in to change notification settings - Fork 28
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
GMT tests fail on Windows: Failure to open virtual file #59
Comments
That error message The rest of the story is that I am the person who creates the GMT Windows installers (I'm a GMT developer too) and when I do the all my GMT.jl testings, examples, etc I am actually using the current GMT development version, which is not exactly equal to the one you installed (time has passed meanwhile). So I can't be 100% sure that that version works the I think (it should but ... you know). If that turns out to be the case, and if you want to be so kind to test it, I can prepare another installer that reflects current state of GMT development, and the one that I'm really using now. |
No, GMT5 is completely uninstalled. There is nothing left where I installed it or in the PATH, and I restarted my computer. I now tried uninstalling and reinstalling GMT6dev. after the restart. This time I tried running the installer as Administrator. I can run If you send me a link I'll try the updated installer for the current master. Since you're the developer of the installer I'll also mention that the PATH too long error I get is very strange. There are only 12 folders in my PATH and I know Windows can handle a lot more than that. None of the 12 folder paths are very long, and I got the error even when I installed directly to C:. |
Yes, I'm sure that only the I'll prepare a new installer with an updated GMT version and will try to test it here on the condition that GMT was installed under "Program Files", which is something that we advise against. That space in dir name tends to cause never ending problems. |
OK, made a new installer and tested it. It works fine for me. BUT, this time I tested the installer by installing it in the original computer where it was created and -- damn it -- got the same warning about path being too long. However my path is a bit long, larger than 2k characters. I think the limit is bytes, not in number on entries. Googling for that warning message reveals a lot of answers coming from cmake, so likely this is still a cmake issue and the solutions continues to be to add the gmt6\bin dir to the path manually. |
With the new installer:
|
It's proving harder than imagined but thanks to this we caught a bug in the GMT development. Very recently we had to make changes related to transparency in the conversion from PostScript to PDF because Ghostscript made a breaking change in its last version (9.25) driven by security reasons. Apparently our internal GMT tests were not able to catch this problem, and neither did I yesterday night when I did only quick tests. In fact I did not run But it's not all bad news. The part that you are more interested now, works. I mean, this works for me
|
Fortunately your error message is much longer than mine
likely because you are on Julia v1.0 and I'm on v0.7. But this issue is more illusive that I thought. It only happens with GMT release builds. When I tested it with a debug build than it works. |
I've been doing lots of test, among them installing a 'cleaner' computer. There I started to also the "Cannot open virtual file" error too. Then I remembered to update the GMT.jl version and started to work. I have no recall to have done breaking changes in master but the fact is that solved the problems, so please one more (and simpler) attempt. Go to GMT.jl master ( |
Sorry for disappearing for a few days, life intervened. But no, even with GMT#master I still get the same error as in my last post. |
Now I get a different error when running
Seems to be a problem with the Windows registry? As far as I know, the only thing that changed since last time was pointing |
Ah, it's working. I just forgot to make it plain clear on the GMT install instructions (and probably try to come out with some test for this) that the ghostscript program is a mandatory dependency too. The thing is |
Continued from Discourse:
https://discourse.julialang.org/t/how-can-i-rasterize-the-polygons-of-a-shapefile-in-julia/15089/14
I have GMT6dev installed (x64 version) and have added
GMT6\bin
to my path. Running]test GMT
or the Hello World example in the GMT docs gets me this error:The text was updated successfully, but these errors were encountered: