Page MenuHomePhabricator

dumpHtml.php causes PHP fatal errors (due to premature output) when read access is locked
Closed, DeclinedPublic

Description

Author: mr.primus

Description:
I used MW 1.5 beta 3, upgraded to 1.52 and later to 1.6devel, got still a
problem with dumphtml.

One problem to be solved was the case if wiki is secured by password, i got a
workaround from Tim Starling for this, which get me some lines further.

new entry in function setupGlobals
$wgGroupPermissions;

in advance of $this->setupDone = true; the following line
$wgGroupPermissions['*']['read'] = true;

Before i had to change localsettings.php, but got the same at the end:

maintenance>php dumphtml.php
Creating static HTML dump in directory static.
Starting from page_id 1 of 1929.
Using database localhost
Writing MediaWiki~Common.css
Writing MediaWiki~Standard.css

After a while there is an error in php.exe:

Die Anweisung in "0x100ab2cb" verweist auf Speicher in "0x0240f000" Der Vorgang
"read" konnte nicht auf dem Speicher durchgeführt werden.

dumpHTML.inc contains following..

'REPORTING_INTERVAL', 500

Destination directory

var $dest;

Show interlanguage links?

var $interwiki = true;

Depth of HTML directory tree

var $depth = 3;

Directory that commons images are copied into

var $sharedStaticPath;

Relative path to image directory

var $imageRel = 'upload';

Copy commons images instead of symlinking

var $forceCopy = false;

  1. Make links assuming the script path is in the same directory as
  2. the destination

var $alternateScriptPath = false;

Original values of various globals

var $oldArticlePath = false, $oldCopyrightIcon = false;

Has setupGlobals been called?

var $setupDone = false;

List of raw pages used in the current article

var $rawPages;

Error only seems to appear if $setupDone = false, but with set to true it doesnt
write MediaWiki~Common.css and MediaWiki~Standard.css.

The next problem seems to be that category tables aren't found because the table
prefix is not used and the table isn't found because of that.

There is no index.html in /static. The pages which were created only link to
Namespace:Article, but not to http://xxx/Namespace:Article.

Any help available?


Version: unspecified
Severity: normal

Details

Reference
bz4132

Event Timeline

bzimport raised the priority of this task from to Low.Nov 21 2014, 8:57 PM
bzimport set Reference to bz4132.
bzimport added a subscriber: Unknown Object (MLST).

mr.primus wrote:

One problem i forgot: the pages in html which were created are in UTF-8, but
browser show them in ISO-8859-1, so the umlautz are incorrect.

mr.primus wrote:

Additional information: I use PHP: 5.0.4 (apache2handler), MySQL: 4.1.12-nt.

translation of error message in php.exe in english:

The instruction in "0x100ab2cb" refers to memory in "0x0240f000". The "read"
process could not be done in memory.

Tim Starling suggested to try another version of php, like 4.4.1, to use a
debugger or linux instead.

But it still remains a bug. :-/

mr.primus wrote:

The bug appears on Windows 2000 and Windows XP too, just checked it.

mr.primus wrote:

In the meantime I upgraded to MW 1.5.8 on PHP: 5.1.1, MySQL: 5.0.18-nt, Apache:
2.2.0

If I write $wgGroupPermissions['*']['read'] = true; into Localsettings
everything seems to work fine. But if I dont do so, I get these errors, even
with the modification of dumpHTML.inc suggested by Tim Starling, which I stated
at start here:

Warning: Cannot modify header information - headers already sent by (output star
ted at \maintenance\dumpHTML.php:66) in \includes\OutputPage.php on line 355

Warning: Cannot modify header information - headers already sent by (output star
ted at \maintenance\dumpHTML.php:66) in \includes\OutputPage.php on line 359

Warning: Cannot modify header information - headers already sent by (output star
ted at \maintenance\dumpHTML.php:66) in \includes\OutputPage.php on line 387

Warning: Cannot modify header information - headers already sent by (output star
ted at \maintenance\dumpHTML.php:66) in \includes\OutputPage.php on line 388

Warning: Cannot modify header information - headers already sent by (output star
ted at \maintenance\dumpHTML.php:66) in \includes\OutputPage.php on line 390

Any help available?

mr.primus wrote:

Yes, still there. The move from core into an extension did not change it.

Used Windows XP Pro, MediaWiki 1.13.4, PHP 5.2.8 (apache2handler), MySQL 5.1.30-community.

Error message: Warning: Cannot modify header information - headers already sent by (output started at \extensions\DumpHTML\dumpHTML.inc:619) in \includes\WebResponse.php on line 10

Aklapper lowered the priority of this task from Low to Lowest.Apr 2 2015, 2:46 PM
Aklapper subscribed.

The DumpHTML has been unmaintained and broken for many years. It is being archived. Declining this task per T280185.