Post reply

Name:
Email:
Subject:
Tags:

Seperate each tag by a comma
Message icon:

Attach:
(Clear Attachment)
(more attachments)
Allowed file types: apk, doc, docx, gif, jpg, mpg, pdf, png, txt, zip, xls, 3gpp, mp2, mp3, wav, odt, ods, html, mp4, amr, apk, m4a, jpeg
Restrictions: 50 per post, maximum total size 150000KB, maximum individual size 150000KB
Note that any files attached will not be displayed until approved by a moderator.
Anti-spam: complete the task

shortcuts: hit alt+s to submit/post or alt+p to preview


Topic Summary

Posted by: Johann
« on: September 21, 2018, 01:46:31 PM »

The result-page does no more display the amount of matched pages and matches. Only the amout of replacements, after execution, would be displayed.
Posted by: Johann
« on: September 17, 2018, 09:40:43 AM »

  • Just to report: the multiline option (when working without delimiter) seems not to work proper. All well by simple using delimiters.
  • Data-capacity: the limit is some over 100MB, which is problematic to replace to large things on the ati pages:


1. Spaces at line-beginning and linebreaks before tags, using find: \n[\s]+< and replace with \n\n, one one hand because replaced would nevertheless give a match and doing folder by folder would need long and has it's and at the root lang. (just lib:thai: could be managed so far, namespace thanissaro would require 300MB+, all in a lang-space propable some 10GB). A possible way, if nothing else found, is maybe 2 two step way, replacing firts with any special character and this later with two line-break. In this way matches can be reduced, slowly, slowly, step by step (about 20-50h).

2. p-tags with two line-breaks by using something like find <\/p>[\s]*<p> and replace with \n\n.

3. the many spaces and tabs between tags without touching/destroying unformated textpages (not thought in detail about it, but would be a mass-problem as well)

4. later on things like em, i, b, br, u, s-tags, while these matches can of cause be reduced step by step.

5. of cause the will be other mass-replacements harder to manage, but can be all of cause done by beggar-"tricks" and effort and patient like always.

* Johann : switching back to huge amount of pts-dictionary -> accessibility replacements for "dummies" and those not wishing to become schoolars or x.y.z., ax4 language speaker, Brahmans or depending on them, before or rather then gaining awakening.
Posted by: Johann
« on: September 06, 2018, 02:43:58 AM »

Sadhu!

The options "multiline", upper/lower case... replace the use of delimiter.
Usually putting the search between /{string}/x. x would give definitions to lower/upeer case, more the one line...

So more user-friendly. Let's see if it can match with the previous "hack" in regard of 10.000 and more matches.

Seems to work fine, shortly tested.
Posted by: Moritz
« on: September 05, 2018, 11:39:30 PM »

I installed the new version of BatchEdit plugin.
There are some new useful options next to the search input:

  • Search mode: "regular expression" (the usual case from previous version) or "plain text" (for simply searching text without regular expression syntax)
  • Case sensitive: yes/no (Klein-/Großschreibung beachten oder nicht)
  • Multiline: not sure how this works; maybe regular expressions would never match mutliple lines without using this option.


There is also a cog wheel (Zahnrad) symbol in the top right corner next to these options, which brings up some additional options:
  • Use delimiters and modifiers in regular expression (not sure what these are, but probably can be found out about here: https://www.regular-expressions.info/ )
  • Show match context of X characters or Y lines
  • Stop search after finding first Z matches (probably important: is set to 100 by default and can be deactivated)
  • Preserve marked matches on preview (wenn nicht gesetzt, verschwinden alle gesetzten Häkchen bei der Vorschau und müssen neu gesetzt werden)
  • Show confirmation on applying edits with no summary (not sure what this means; irgendeine Abfrage zur Sicherheit in manchen Fällen)

Also there is now a time limit on how long a search or replacement can take (can be changed in Admin settings). I have set this to 10 hours now. Should be enough usually.

Very helpful: there is now a progress bar for the search progress and replacement progress, helping estimating how much longer it will take (very light grey, difficult to distinguish from white).

Not tested much, hopefully not any new errors.


Edit: Just tested searching for "dhamma" with no limit of results; returns an empty result page. Probably too many results so that something gets broken.
Searching for "dhamma" with limit of 16000 results works, and takes a few minutes to complete.
Searching for "Johann" without limit works and gives 2076 results.
Posted by: Johann
« on: August 17, 2018, 02:15:37 AM »

Good to hear. As far for now, Atma is used to and knows it's capacity and ways well.
Posted by: Moritz
« on: August 17, 2018, 12:22:57 AM »

The first looks like an error on Greensta's side. The database was unavailable for a moment it seems.  :-| But if not happening more often, hopefully not a big problem.

The second bug was introduced by me. I just wanted to have different colors for match and replacement.
So instead of having yellow for both, I wanted to have yellow and green in the preview.
And red and green after the replacement.

But it seems I have not changed it for the first preview, where still both is yellow.

The original author was also wondering why I did this change. Now I see it's different for the first preview. Okay.

There has been a lot of new work been done in the meantime by the original author and others, including some really helpful new features like a progress bar, so that one can estimate how much more time a replacement will take for large updates. And much cleaner solutions to the small changes that I made.

I think I should update to the new version soon.
Posted by: Johann
« on: August 15, 2018, 11:16:51 AM »

strange error appeared while doing on replacement after another. But seems to be fine if just loging in again.

screenshort attached.

Info of a "bug": if selecting matches but push again on preview, it will return the replacements in green althought just reviewed.
Posted by: Johann
« on: August 12, 2018, 03:35:25 PM »

(...lookahead need something behind. Guess the issue is solved for my person)

The replacment tool has a problem with " . Put into search or replace, it would break the search or replace string after execution. But can be fixed by using [^\w] instead, at least for search.
Posted by: Johann
« on: July 26, 2018, 12:08:30 PM »

lookahead: no idea why there is recognition but no replacement for example with this regex:

Code: "find" [Select]
/\[\[([^\w]*)\/(?=lib\/|tipitaka\/|cdrom\/|extras\/|news\/|noncanon\/|ousources\/|pdf\/|s\/|tech\/)/


Code: "replace with [Select]
[[de:


a given [[../../../lib gets the match [[../../../ and the replacement look the same [[../../../lib ?

certainly total no more my persons sphere at all, this chess thinking...
Posted by: Johann
« on: July 02, 2018, 04:27:22 PM »

"$" comand seems no more working in replace-line. Maybe a change of plugin-code have been done. Havn't tested "\"

The replacement syntax has not been changed. Testing here, both "\" and "$" works for inserting match back-references.

Maybe another hint, not to forget the parantheses ().

\1, \2, \3 or $1, $2, $3 ... etc. are references to the groups inside parantheses. With no parantheses, there is no input for \1, \2, $1, $2 etc.

...
_/\_
Sadhu for acting zuvorkommend.
Posted by: Johann
« on: July 02, 2018, 04:18:11 PM »

Oh, I forgot to include the images that were added by the original author in his most recent updates. The DokuWiki download still has the version from February. The original author, Mykola Ostrovskyy has not yet created a new release version since February. It seems he is still working on some major changes he would like to add before the next "official" version.

So all errors here are just because I forgot to upload certain new files. But now I think it should be okay?

How ever, the resultpages misses now the amout of pages matched, and sum of matches, which is a useful controll and estimation of success point.

I'm not sure how this could be. Testing from here, I get infos like this:

After "Preview":
Quote
Search results: 9808 matches on 1019 pages

After "Apply":
Quote
Edit results: 9808 matches on 1019 pages, 2 replacements applied

Maybe it's a matter of display, caused by responsibility for mobil advices.
But what was just white before, contains now the matches.

* Johann : There seems to be a lot to understand in regard of "Zwischenspeicher"Also troubles with favicon, even on all places placed and a great deal that in cs-rm, the site takes the old version as the newer, meaning all "drafts" to recover, one by one.

"$" comand seems no more working in replace-line. Maybe a change of plugin-code have been done. Havn't tested "\"

The replacement syntax has not been changed. Testing here, both "\" and "$" works for inserting match back-references.

For example:

regex: "/(mindfulness)/"

replacement: "$1 test \1"

will replace "mindfulness" with "mindfulness test mindfulness".
Seems to be working without problem here.

_/\_

That's great. Might be again certain momentary personal handycap, here and there.
Posted by: Moritz
« on: July 02, 2018, 03:13:41 PM »

"$" comand seems no more working in replace-line. Maybe a change of plugin-code have been done. Havn't tested "\"

The replacement syntax has not been changed. Testing here, both "\" and "$" works for inserting match back-references.

Maybe another hint, not to forget the parantheses ().

\1, \2, \3 or $1, $2, $3 ... etc. are references to the groups inside parantheses. With no parantheses, there is no input for \1, \2, $1, $2 etc.

example:
regex: "/(\s[a-zA-Z]*) something in between (mindfulness)/"

replacement: "\1 something different \2"

would replace like this:

"satipatthana something in between mindfulness"
=> "satipatthana something different mindfulness"

"ariyasacca something in between mindfulness"
=> "ariyasomething something different mindfulness"

_/\_
Posted by: Moritz
« on: July 02, 2018, 02:55:00 PM »

Just came across another error that could happen when the amount of matches is really huge (for example, searching for "/is/" - must have millions of matches probably):

Quote
Fatal error: Allowed memory size of 134217728 bytes exhausted (tried to allocate 69632 bytes) in /var/www/clients/client2157/web5417/web/lib/plugins/batchedit/admin.php on line 336

So just to inform, if coming across that, that is because there are too many results to keep in memory.

I had some discussion with the author who is currently in the process of making some major changes, also thinking about how to deal with huge result sets . So I think I should mention that to him as well and maybe help and try to find a solution. But at the moment don't have much time for this.
Posted by: Moritz
« on: July 02, 2018, 02:41:48 PM »

Oh, I forgot to include the images that were added by the original author in his most recent updates. The DokuWiki download still has the version from February. The original author, Mykola Ostrovskyy has not yet created a new release version since February. It seems he is still working on some major changes he would like to add before the next "official" version.

So all errors here are just because I forgot to upload certain new files. But now I think it should be okay?

How ever, the resultpages misses now the amout of pages matched, and sum of matches, which is a useful controll and estimation of success point.

I'm not sure how this could be. Testing from here, I get infos like this:

After "Preview":
Quote
Search results: 9808 matches on 1019 pages

After "Apply":
Quote
Edit results: 9808 matches on 1019 pages, 2 replacements applied

"$" comand seems no more working in replace-line. Maybe a change of plugin-code have been done. Havn't tested "\"

The replacement syntax has not been changed. Testing here, both "\" and "$" works for inserting match back-references.

For example:

regex: "/(mindfulness)/"

replacement: "$1 test \1"

will replace "mindfulness" with "mindfulness test mindfulness".
Seems to be working without problem here.

_/\_
Posted by: Johann
« on: July 02, 2018, 10:44:52 AM »

After update Indexing finished, the "layout error" still exists, like pic above. The text says:

Code: [Select]
Warning: file_get_contents(/var/www/clients/client2157/web5417/web/lib/plugins/batchedit/images/file-document.svg): failed to open stream: No such file or directory in /var/www/clients/client2157/web5417/web/lib/plugins/batchedit/admin.php on line 833

Warning: file_get_contents(/var/www/clients/client2157/web5417/web/lib/plugins/batchedit/images/pencil.svg): failed to open stream: No such file or directory in /var/www/clients/client2157/web5417/web/lib/plugins/batchedit/admin.php on line 833

Warning: file_get_contents(/var/www/clients/client2157/web5417/web/lib/plugins/batchedit/images/arrow-down.svg): failed to open stream: No such file or directory in /var/www/clients/client2157/web5417/web/lib/plugins/batchedit/admin.php on

Thinking, oh, my person added the images in the directory, having taken them from the github download (trusting that this might be welcome), and now seems to work fine, in regard of layou.

The download on docuwiki misses those images. My person told it via the forum .

How ever, the resultpages misses now the amout of pages matched, and sum of matches, which is a useful controll and estimation of success point.