Re: [vim/vim] fix(cli): parsing bug - require exact match for longopt name (PR #21088)

h-east left a comment (vim/vim#21088)

The diagnosis is right: `--servername=test` slips past `early_arg_scan()`,
which already compares exactly (`main.c:1906`), and is then matched by the
prefix comparison in `command_line_scan()` (`main.c:2268`), which reacts by
skipping the next argument.  Two things need work.

### Only half of the mismatch is fixed

`--servername` is compared exactly in both places after this change, which is
why the reported case works.  `--clientserver` is not: `early_arg_scan()`
still has `STRNICMP(argv[i], "--clientserver", 14)` at `main.c:1923`.  So
`--clientserverx socket` is still consumed there, sets `clientserver_method`
and eats the following argument, and only afterwards is rejected by
`command_line_scan()`.  It takes effect before it is refused.  Both places
should agree.  `--remote` at `main.c:1913` must stay a prefix comparison,
since `--remote-wait` and friends depend on it.

### The test depends on the locale

`Unknown option argument` is wrapped in `N_()` (`main.c:60`).  `lang mess C`
in `runtest.vim:155` only affects the Vim running the tests; the Vim started
through `system()` inherits `$LANG`, so the test fails wherever translations
are installed and the environment is not English.  Set `$LC_ALL` to "C"
around the call, or rely on `v:shell_error` alone.

Smaller points on the same test: `GetVimCommand()` already yields `-u NONE`
and `--not-a-term` (`util/shared.vim:349-355`), so `-Nu NONE` is redundant;
`2>&1` is shell syntax and `CheckNotGui` does not exclude the Windows
console; and only `--servername=test` is exercised, not `--clientserver=socket`.

### Two notes

`serversend` at `main.c:2269` has no implementation anywhere - not in
`early_arg_scan()`, the usage message or `runtime/doc/remote.txt`.  It
silently swallows its argument today and still will after this change.
Worth deleting rather than tightening, probably as a separate patch.

`--servername=test` now aborts startup instead of being quietly mishandled.
That is defensible - `remote.txt:62` only ever specified the form with a
space - but it deserves a line in the commit message.

-- 
Reply to this email directly or view it on GitHub:
https://github.com/vim/vim/pull/21088#issuecomment-5346695194
You are receiving this because you are subscribed to this thread.

Message ID: <vim/vim/pull/21088/[email protected]>

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups "vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
To view this discussion visit https://groups.google.com/d/msgid/vim_dev/vim/vim/pull/21088/c5346695194%40github.com.