-
Notifications
You must be signed in to change notification settings - Fork 40
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
not reading default.file parameter? #94
Comments
Thanks. I'm at a loss here, this will need a bit more digging. |
Same here. :) Can you think of anything that might be useful for troubleshooting or digging in under the hood more? |
I have a very similar issue and I arrived to similar conclusions (the .cnf file is not being loaded but defaults are) |
We could use |
I don't have root to this box, so haven't had luck with strace... but I may have come across another clue: I'm using macports, and just got this error from the command line:
It could be that this is a macports issue, not a RMariaDB issue... still a mystery, but I'll poke around on that front a bit and update here. |
Follow-up: not sure if related or not. My error from the command line arose because the first dash in my command above was an em-dash (stupid autocorrect and copy-paste!). The command works from the command line if I manually type it instead of copy/paste. |
Can you please try |
Same story. :(
with
|
In your console example you used |
Yep, it's always good to check:
|
potentially related: r-dbi/RMySQL#222 |
I am having the same problem as @bheavner (who is also my colleague, and thus we have a similar computing set up). I tried compiling RMariaDB from source, since it fixed the problem with the new version of Interestingly, |
I'm not sure if Lines 40 to 42 in 9f88fb6
Could you please try installing RMariaDB from my branch, and connect again? remotes::install_github("r-dbi/RMariaDB@18a071c") |
That worked! I can connect with a .cnf file with that branch. |
Weird! It did not work for me. @amstilp and I have found a difference in the installation step, and I'm digging more. I get this output:
(then much more weird homebrew stuff happens). I do not think I have brew installed on my machine. @amstilp does not get any brew output, and it works for her. Surprisingly, we get the same output from things like I am digging more into why my box is doing brew-y things, and will update here. I strongly suspect something weird in my library paths because of this brew-related mystery. |
IT WORKS FOR ME! Woot! Thanks very much for your patience and perseverance on this issue, @krlmlr! Here's what I found as the root cause of my weirdness and the difference with @amstilp 's first success: although we are both using RStudio, she always starts from the terminal, and I start using a pinned icon in my taskbar. Apparently, macOS doesn't read bash environmental variables when starting from the taskbar, so my RStudio did not have the correct $PATH. Critically, it was missing the path to I haven't figured out how to force RStudio to get the bash environmental variables and set the path correctly when starting from the taskbar, but that is not the problem of RMariaDB. If there was a change in RMariaDB that led to this behavior, I now have a way to work with the current implementation. I think it's more likely that there was a change in the underlying library on the Mac OS, so the things that fall within the default path weren't working. As one final point of information, I have the following $PATH in my RStudio when starting from the pinned icon:
Thanks again, very much! |
Hmm, it seems like the CRAN version (1.0.6) is also working for me, but I'm not sure why it wasn't working last week. I may have installed it differently or forgotten to restart my R session in between testing. Sorry about the confusion, I think we have it figured out here! |
To confirm, the CRAN version works for me as well - if I install from source. It does not work (and gives the familiar error), if I just do a vanilla CRAN install. |
Same for me. I should have said I had to install from source. Thanks again! We can probably close this issue. |
Though ideally, the vanilla CRAN install should work, too, yes? The $PATH issue seems to only matter when I'm compiling source. |
Yes, binary install from CRAN should work too. The Could you please set up your system so that |
I think you're asking for more output when I run the
With that, ... Is that helpful? |
Absolutely. What happens if you install |
Yes, it does! But with more compilation warnings:
|
Thanks. Does this mean that the change in 18a071c doesn't affect if the |
I think that's correct - the change in Building from source on mac seems to matter - the CRAN binary does not work for me in an R session started from the terminal or from RStudio (which I've now hacked to get my environmental variables). If I begin by deleting the
If I again delete the
Similarly, if I install a source package (after deleting
I see the same behavior in my RStudio session, now that I have have confirmed my path includes macports binaries:
If I install the source package from CRAN with the default RStudio path, I see the various brewing messages, and it does not work:
Similarly, it does not work if I have the default $PATH and install from
|
To me, this boils down to:
As @bheavner confirmed, even without having to change the code. Can we close this issue then? It remains to see why installing from binary fails. Unfortunately, I have a very limited (read: no) understanding of how these thins work on the Mac, only https://xkcd.com/1987/ comes to mind. What does a successful installation using CRAN binaries look like? |
This old thread has been automatically locked. If you think you have found something related to this, please open a new issue and link to this old issue if necessary. |
Something appears to have changed, and previously working code is not working for me now.
I have previously successfully established a database connection thus:
Where
~/.mysql-uwit.cnf
looks like this:But now that returns an error that appears to show an attempt to connect to a local MySQL, instead of the one running at my host:
I have confirmed that I can connect from the command line with
mysql --defaults-file=~/.mysql-uwit.cnf
, and I can connect without a .cnf file, like this:A further clue - if I try to specify the host in
dbConnect()
, but get other values from .cnf file, it appears to not load the username or password from the .cnf file:I'm working in RStudio on a mac, and here's my sessionInfo:
The text was updated successfully, but these errors were encountered: