User:SDrewthbot/trim trailing LF

SDrewthbot

trim trailing LF[edit]

An amendment to the code to Proofread Page extension has fixed one problem and caused another. Rather than stop proofreading or to cause too much confusion to users, it was suggested that we get a bot to correct the error quietly, and the bot operation linked to here is that fix. If you are really interested in the lead-in discussion, please follow the reference link. This bot will continue to run this task on a regular basis until the mediawiki application has been fixed. — billinghurst sDrewth 05:40, 19 October 2011 (UTC)

Synopsis[edit]

The below data relates to fixes initially undertake for enWS, and per a request for the fix to be implemented at laWS. Subsequently there was a mulWS request for this fix to be applied more widely through the xxWS community. … "cross-subdomain bot repairing bug 26028 per oldwikisource:Wikisource:Scriptorium"

The bot will run through the broader WS wiki space with this bot account, though not with a bot right slowly fixing these edits. To note that the communities fr/de/it will not be part of the fix due to local fixes being undertaken.

Query being run to determine Page:s edited.

https://xx.wikisource.org/w/api.php?action=query&list=recentchanges&rcnamespace=104&rcstart=2011-10-28T21:00:00Z&rcend=2011-10-16T23:00:00Z&rcshow=!bot&rclimit=500&format=xml

noting that where the Page: namespace is not 104, that will be modified as required. As runs are being undertaken they will appear below.

Technical[edit]

The query to grab edits is based on a query of the API

//en.wikisource.org/w/api.php?action=query&list=recentchanges&rcnamespace=104&rcstart=2011-10-21T12:00:00Z&rcend=2011-10-19T12:00:00Z&rcshow=!bot&rclimit=5000&format=xml

where the time is varied to get the next group of edits. This query is used with w:Wikipedia:AutoWikiBrowser — to make the list, utilise the HTML Scraper (with advanced Regex) , the url is plugged in and the filter used title="([^"]+) with group set to 1. Note that with AWB HTML scraper that it will grab a maximum of 500 hits, so if the list is going to be bigger than 5000, then shorter timer periods should be searched in spans to build the great list.

The replacement to undertake is

\s?\n(\<noinclude\>) with regex yes replaced with $1

Update — As there was noted that it would be preferable for their to be a line feed between <<references/>, I will add one in. — billinghurst sDrewth 13:07, 29 October 2011 (UTC)

To note, that at enWS the first error noted was at 16 October 2011‎ 23:13 UTC this equates to &rcend=2011-10-16T23:00:00Z and fix occurred sometime prior to 21:21, 28 October 2011 &rcstart=2011-10-28T21:00:00Z

[toolserver.org/~phe/statistics.php?diff=14&daysago=0]

Difference between Sat Oct 15 and Sat Oct 29

language	all pages	not proof.	problem.	w/o text	proofread	validated	all pages	with scans	w/o scans	disamb	percent
	Page namespace						Main namespace
fr	8499	2074	3	656	5766	1484	820	732	76	12	0.36
en	3205	303	28	315	2559	531	1060	902	151	7	0.29
de	784	-298	0	70	1012	1470	265	218	47	0	0.20
es	1122	-25	-4	42	1109	210	94	67	27	0	0.11
it	1248	529	104	52	563	258	31	50	-20	1	0.11
pl	533	265	-2	0	270	703	114	117	-4	1	0.31
sv	30	-76	-1	0	107	2	3	3	0	0	0.04
no	218	-14	0	9	223	5	38	42	-4	0	0.38
ca	184	0	0	0	184	7	0	13	-13	0	0.52
ru	117	5	-3	20	95	11	980	150	820	10	0.08
hy	412	335	0	1	76	69	75	63	10	2	2.03
da	224	150	0	1	73	0	4	3	1	0	0.09
vec	0	0	0	0	0	5	0	0	0	0	0.00
pt	1	2	-1	0	0	0	9	0	9	0	-0.00
br	28	-44	0	5	67	7	28	27	1	0	0.15
sl	77	-487	-1	0	565	0	485	0	485	0	-0.00
old	53	2	0	1	50	0	29	19	10	0	0.14
la	246	209	0	0	37	0	-3	0	-3	0	0.00
hr	0	0	0	0	0	0	4	0	4	0	-0.00
hu	0	0	0	0	0	0	13	0	13	0	-0.00
et	0	0	0	0	0	0	0	0	0	0	0.00
id	0	0	0	0	0	0	3	0	3	0	-0.01
vi	0	0	0	0	0	0	116	0	116	0	-0.02
el	0	0	0	0	0	0	158	0	157	1	-0.01
zh	0	0	0	0	0	0	125	0	127	-2	-0.00
te	0	0	0	0	0	0	60	0	60	0	-0.00
he	0	0	0	0	0	0	50	0	50	0	-0.00
total	16981	2930	123	1172	12756	4762	4561	2406	2123	32