Zsh Mailing List Archive
Messages sorted by:
Reverse Date,
Date,
Thread,
Author
Re: PATCH: Fix "35531: fallback on file completion"
- X-seq: zsh-workers 38183
- From: Oliver Kiddle <okiddle@xxxxxxxxxxx>
- To: zsh-workers@xxxxxxx
- Subject: Re: PATCH: Fix "35531: fallback on file completion"
- Date: Sat, 19 Mar 2016 00:20:47 +0100
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.co.uk; s=s2048; t=1458343249; bh=sbAMryICqSxqFqFuYyyRSQtzqkvQm2JDD5Kgy9SM1Zw=; h=In-reply-to:From:References:To:Subject:Date:From:Subject; b=fstVLJMJlmfUT6C/lLHBHzz1op95jPQSsF39cs6VLdno9v8Y8w2WnBADO+U4+jhfnijnBbt3ngxEnTwzM8PuFtZQUFEuYf4L4eZmeYxRjikFAEWsTs3AQeQLlwpqhRcWS65tWTGOTFnT4hNzdtreUv5QYRAhKiDrfW26sxF09C9uXyijSIS2sHHsKg8GNS+pehSIDy41VQX9Kg3r6MsDVQQlc2bWIvIxvRtXGMU4qk8bCf/cNaU0BBhnQn1ztDk+Qf8FHzbeTf0eFkmv1R5+la7n0NnrLSbNYxO/y//jdw04KTYgzZdfGbf4OjbwVseNPLW94yObuyUxLkNjxx1xvQ==
- In-reply-to: <1458317336-27434-1-git-send-email-mikachu@gmail.com>
- List-help: <mailto:zsh-workers-help@zsh.org>
- List-id: Zsh Workers List <zsh-workers.zsh.org>
- List-post: <mailto:zsh-workers@zsh.org>
- Mailing-list: contact zsh-workers-help@xxxxxxx; run by ezmlm
- References: <1458317336-27434-1-git-send-email-mikachu@gmail.com>
Mikael Magnusson wrote:
> The above commit just changed a bunch of stuff without testing, the
> result was that the completer didn't work. It now works, to an extent.
Specifically, in what way didn't it work?
Testing will have been limited because I don't have any device adb could
connect to (still using an ancient Nokia) and I don't typically use or
install it. There's a different adb command on Solaris which is perhaps
why I was annoyed at it not falling back to default file completion.
> The one remaining caveat is that if you enter a subcommand the completer
> doesn't recognize, it will go on to offer subcommands instead of falling
> back to _default. I'm not sure what it's trying to do with the
> "sanitize context" stuff, or "shift words" etc which just makes the rest
> of the code not know where it is, but I don't care enough to rewrite it.
The "sanitize context" stuff was because it was putting the sub-command
into the wrong component of $curcontext. Completion contexts are
supposed to be:
:completion:<function>:<completer>:<command>:<arg>:<tag>
We typically put subcommands into <command> by changing it to
<command>-<subcommand>. If you use a colon instead of a dash, the
subcommand will be in <arg> and thereafter, everything gets messed up.
> -(( $+functions[_adb_check_log_redirect] )) ||
> +(( $+functions[adb_check_log_redirect] )) ||
> _adb_check_log_redirect () {
Are you sure you want to back that out?
I'm not actually fond of these guards on function definitions myself
because they prevent the function from being reloaded when the
underlying file in $fpath is modified and reloaded. I only use them in
cases where it seems likely that someone would want to override the
function.
> - LOG_REDIRECT=${$(adb ${=ADB_DEVICE_SPECIFICATION} shell getprop log.redirect-stdio 2>/dev/null)//
> + LOG_REDIRECT=${$(adb ${=ADB_DEVICE_SPECIFICATION} shell getprop log.redirect-stdio)//
> /}
This should really be using _call_program but I'm sure I redirected
stderr for a reason. It was probably spewing out errors to the terminal
because it couldn't find an attached device.
I also don't really like _adb being chatty with lots of _message -r
calls. I tend to prefer completion functions to just do the best they
can of generating matches and otherwise keep quiet.
Oliver
Messages sorted by:
Reverse Date,
Date,
Thread,
Author