-
Notifications
You must be signed in to change notification settings - Fork 749
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
[lex.phases] replace term 'input file' with 'source file' in phase 1 #7192
base: main
Are you sure you want to change the base?
Conversation
The only part of the standard that refers to input files is phase 1 of the phases of translation. Everywhere else, we use the term source file, that has a clear definition in the preceding [lex.separate] clarifying that we do not necessarily mean text files in a hosted filing system.
CWG and/or SG16 spent quite some time massaging [lex.phases] p1 into its present form (for the UTF-8 source support); I won't touch that wording editorially. Maybe "input file" is the subject of phase 1, and "source file" is what's handled in phases 2 and 3. And it's intentional that we only ever talk about the raw sources of translation in phase 1 and nowhere else. |
I am very happy to defer this to Core, but at the moment "input file" is not a defined term in the standard, and occurs in only this one paragraph. We are very careful to define "source file" as not necessarily implying a file in a filing system, but have no such notion of what "input file" should mean. I believe we need clarity here. |
Where in [lex.separate] do we expressly say we don't imply a "file"? I think we just do so by omission, and "input file" is no better or worse here. |
I stand corrected. The only reference I can find to "not necessarily a file" is a footnote in [headers], 16.4.2.3, and that applies to headers, not to source files. |
Can we please not do this change, then? |
I would still like the discussion in Core, because the term "input file" is not defined, unlike the term "source file". If you would prefer to create a core issue and close this PR, I would be content. |
Late comment: I actually just ran across the note that says source files need not be represented as physical files --- it is a note in phase 7 of translation, so not normative but clarifying what must otherwise be inferred. |
Is this a conversation we need to have here, or should this happen in CWG (or on the reflector)? |
Jens stated that such discussion belongs in Core. I will start a thread there this week, pointing back to this PR. |
The only part of the standard that refers to input files is phase 1 of the phases of translation. Everywhere else, we use the term source file, that has a clear definition in the preceding [lex.separate] clarifying that we do not necessarily mean text files in a hosted filing system.