|
From: | Dmitry Gutov |
Subject: | bug#61849: 29.0.60; Unable to use treesit-install-language-grammar because repo doesn't have parser.c |
Date: | Mon, 20 Mar 2023 22:04:41 +0200 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101 Thunderbird/102.8.0 |
On 20/03/2023 21:28, Eli Zaretskii wrote:
From: Misha Zharov<mishazharov1@gmail.com> Date: Sun, 19 Mar 2023 22:14:44 -0700 Cc:casouri@gmail.com,61849@debbugs.gnu.orgPlease chime in to try to convince them to go back to including the generated parser files.Thanks for the update. I've now done this, but I think that it will be difficult to make sure that community sticks to this standard of keeping the autogenerated files in the repo. Their arguments do have some merit regarding the difficulty of resolving merge conflicts and overall maintenance burden. At the same time, having the autogenerated file present is a nice QOL feature for users because it's easily reproducible.Bummer: https://github.com/DerekStride/tree-sitter-sql/issues/120#issuecomment-1476609242
Suppose the repositories remove the generated grammar files. What would be sufficient for us to regenerate them?
tree-sitter-sql apparently uses the tree-sitter-cli program. Would having it on the user's system suffice?
For a lot of developers NPM will already be installed. If the only remaining step will be 'npm install -g tree-sitter-cli', or dropping one of the pre-built binaries from https://github.com/tree-sitter/tree-sitter/releases into a directory on PATH, I think they could manage to do it once, to be used in all grammars which don't keep generated files in the repo.
[Prev in Thread] | Current Thread | [Next in Thread] |