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
Issue PRs changing toolchain over to jinja2 compiler syntax #162
Comments
Related: conda-forge/conda-smithy#781 |
Perfect, maybe this issue is just "have the bot crawl through all toolchain recipes and use @isuruf's awesome tool to fix everything". |
This is really a question of whether we want transient migration scripts as part of the bot. While the bot is a very cool thing that could push such changes with, wouldn't want to unnecessarily burden the bot with things needed for such scripts. Ultimately think it is up to you to make that decision. |
I think we have run into the migration issue (or "rebuild-the-world") a couple of times with conda-forge. The bot seems like a good place to put this capability and to manage these kinds of issues in the future. |
It would be hellpful to know how many recipes need to be rebuilt. #44 (comment) gives all descendants of toolchain, but we only need those recipes with toolchain as a build dep. |
This is in progress, we need to get #197 in, manually fix the graph, and then restart the bot and we should be good to go. |
I think this is done! |
I think there are still some PRs out there that have failed to build under the new syntax |
I don't think all recipes have been PRed yet. #204 |
Running import networkx as nx
from conda_forge_tick.migrators import Compiler
g = nx.read_gpickle('graph.pkl')
c = Compiler()
count = 0
for node, attrs in g.node.items():
if not c.filter(attrs):
count += 1
print(count) returns 263. So it looks like there are still 263 recipes to go. @CJ-Wright why did you think it was done? |
(I closed this because we are PRing things, not because all the PRs have be done) |
As a prelude to #44 it may be helpful to first issue a bunch of PRs into recipes currently using the
toolchain
syntax to move them over to the compiler syntax. Note that this will not move them over to the new compilers, the new syntax does the same thing as toolchain under the hood (I think? @msarahan?). This is needed for the new to the new compilers.I think the main issues here are:
toolchain
was an omnibus system everything was under toolchain (c, c++, fortran, etc.). The new syntax is more specific and specifies the exact compiler needed. One approach could be to get the code from the tarball and inspect the extensions of all the code? I don't know how reliable that would be, but it might be a good first guess.The text was updated successfully, but these errors were encountered: