From 63a16b706303c5036172bcebe359ff677a2631e8 Mon Sep 17 00:00:00 2001
From: Albert Graef <aggraef@gmail.com>
Date: Sun, 14 Aug 2022 13:31:02 +0200
Subject: [PATCH] Handle some more corner cases of help patches, bugfixes.
---
pd/nw/pdgui.js | 64 +++++++++++++++++++++++++++++++++++++++-----------
1 file changed, 50 insertions(+), 14 deletions(-)
diff --git a/pd/nw/pdgui.js b/pd/nw/pdgui.js
index e5ed40179..729e2187d 100644
--- a/pd/nw/pdgui.js
+++ b/pd/nw/pdgui.js
@@ -158,6 +158,8 @@ function index_entry_esc(s) {
// keywords, and description of help patches.
function add_doc_details_to_index(filename, data) {
var title = path.basename(filename, "-help.pd"),
+ // AG: This is confusing. I'm not sure why we just replace the first
+ // newline here. Maybe there's a reason to do so, but I don't get it.
big_line = data.replace("\n", " "),
keywords,
desc,
@@ -232,32 +234,66 @@ function add_doc_details_to_index(filename, data) {
keywords = keywords && keywords.length > 1 ? keywords[1].trim() : null;
desc = desc && desc.length > 1 ? desc[1].trim() : null;
// Remove the Pd escapes for commas
- desc = desc ? desc.replace(" \\,", ",") : null;
+ // AG: We want to do a global replacement here.
+ desc = desc ? desc.replace(/ \\,/g, ",") : null;
if (desc) {
- // format Pd's "comma atoms" as normal commas
- desc = desc.replace(" \\,", ",");
+ // AG: And we want to get rid of the newlines, there's no reason
+ // whatsover to preserve the original formatting.
+ desc = desc.replace(/\n/g, " ");
}
- /* AG: Deal with a bunch of special cases which have multiple objects
- documented in them, listing the object names in the NAME field of the
- META data. */
+ // AG: Deal with a bunch of special cases which have multiple objects
+ // documented in them, listing the object names (and arguments) in the
+ // NAME field of the META data.
var names = big_line
.match(/#X text \-?[0-9]+ \-?[0-9]+ NAME ([\s\S]*?);/);
- names = names && names.length > 1 ? names[1].trim().split(" ") : [];
+ // Some NAME fields span multiple lines, pretend that they're spaces.
+ names = names && names.length > 1
+ ? names[1].trim().replace(/\n/g, " ").split(" ")
+ : [];
if (names.length > 1) {
- // special help file, not an actual object, remove from completions
+ // special help file, not a single object, remove from completions
let obj_result = obj_exact_match(title);
if (obj_result.length !== 0) {
let obj_ref = obj_result[0].refIndex;
//post("scanning "+filename);
completion_index.removeAt(obj_ref);
}
- // Special cases which don't have information in their NAME fields
- // that we can use. We might add more as we become aware of them.
- let skip = title == "list" | title == "midi";
- if (!skip) {
- // enter the completion entries for each of the objects named in
- // the NAME field
+ // Some NAME entries (e.g., list) list different variations of the
+ // same object (list append, list trim etc.), we try to be clever
+ // about these.
+ let prefix = names[0];
+ let count = names.filter(name => name == prefix).length;
+ if (count > 3) {
+ // Chances are good that we're looking at variations of the same
+ // command, and not just some oddball NAME list with repeated
+ // entries. We first check for existing entries, and skip those
+ // (TODO: we might want to update their arguments instead, but
+ // this code is already complicated enough).
+ obj_result = obj_exact_match(prefix);
+ if (obj_result.length === 0) {
+ // Now collect the arguments.
+ let args = [];
+ for (let i = 0; i < names.length; i++) {
+ if (names[i] == prefix) {
+ let a = [];
+ while (i+1 < names.length && names[i+1] != prefix) {
+ a.push(names[i+1]); i++;
+ }
+ if (a.length > 0) {
+ args.push({"occurrences" : 0, "text" : a.join(' ')});
+ }
+ }
+ }
+ //post("add completion "+prefix+" "+args.map(a => a.text).join(' '));
+ completion_index.add({
+ "occurrences" : 0,
+ "title" : prefix,
+ "args" : args
+ });
+ }
+ } else {
+ // just a list of object names here, add them all
for (let i = 0; i < names.length; i++) {
let title = names[i];
// check for existing entries, skip those
--
GitLab