Importing entries and pages from PivotX to Bolt

[Harm10]

Can you post the xml part of this content after export? What tags are these line breaks?
I assume the content displays normally when seen on your PivotX site?

[jult]

A couple of example snippets from the export .xml;

Code: Select all

<content:encoded><![CDATA[Maarten op de Mixfreaks mailinglist schreef;    > Hou jij van klassiek? Misschien niet, maar je voorouders zullen je zeggen  > dat dat de enige muziek is die er bestaat en rock 'n' roll of weet ik veel  > welke andere muziekstroming niets!    Het gaat mij om iets heel anders, namelijk dat men bij de radio  en in de detailhandel nogal geneigd is om op safe te spelen.  Tot aan ongeveer 1995 was dat echt beduidend minder het geval  en dat heeft niks met mijn leeftijd te maken.    De platenzaken die voorheen nog wel eens een plaat/CD hadden liggen  waarvan bekend was dattie NIET bij voorbaat de hitparade in zou komen  durven dat niet meer(?). Mijn muzieksmaak is ècht niet die van een  old-timer of gek, en ik kan heel erg meekomen in veel nieuwe muziek  die wel te koop ligt, maar BINNEN die genres neemt men  geen enkel risico meer. Er liggen bijv. geen import-plaatjes meer  in main-stream winkels. Vroeger lagen ze zelfs nog in V&D.    Ander probleem vind ik dat de platenzaak waar ik dan nog wèl  kleine labels en import kan vinden weer op zo'n overdreven  manier

All spaced out parts used to be linebreaks here ^ (the lines starting with ">" for example).
Another example:

Code: Select all

Not too unexpected; Debian is too slow releasing updates, let alone backports to older OS versions, like debian 6 (squeeze) or older.  So with them old OS-es you're stuck with old openssl crap through apt-get update, and standard install from source will *not* update a previous apt-get installed openssl! Still want to update to 1.0.1g or later? Here's how it *does* work: Check the link for the version that says <a href="https://www.openssl.org/source/">[LATEST] over here</a>. Currently that means you need to do <code># wget https://www.openssl.org/source/openssl-1.0.1g.tar.gz # tar xfz openssl-*.tar.gz # cd openssl-* # ./config --prefix=/usr zlib-dynamic --openssldir=/etc/ssl shared enable-ec_nistp_64_gcc_128 </code> It then asks you to do make depend, which you should, and you follow with the usual: <code># make depend # make # make install</code> and then you can check the version that's running on your server with <code>openssl version</code> and if it says: <code>OpenSSL 1.0.1g 7 Apr 2014</code> you're safe and sound in Snowden-land. This also works for Ubuntu 8 (hardy) by the way.  Be sure to reboot your server after that, renew your server certificates, ask (or better yet FORCE) your users to change their passwords.  Of course, you've now gone to source for this package. And trying <code>apt-get remove openssl</code> will result in sheer dependency hell. Forget about letting apt manage openssl in any way or manner. dovecot, nginx, postfix etc. should all still work fine through apt with backports and such. They'll use the right openssl libraries now.]]></content:encoded>

This used to look like this (as HTML here, straight from the pivotx editor):

Code: Select all

<p>Not too unexpected; Debian is too slow releasing updates, let alone backports to older OS versions, like debian 6 (squeeze) or older.<br /> <br /> So with them old OS-es you're stuck with old openssl crap through apt-get update, and standard install from source will *not* update a previous apt-get installed openssl!<br /> Still want to update to 1.0.1g or later? Here's how it *does* work:<br /> Check the link for the version that says <a href="https://www.openssl.org/source/">[LATEST] over here</a>. Currently that means you need to do<br /> <code># wget https://www.openssl.org/source/openssl-1.0.1g.tar.gz<br /> # tar xfz openssl-*.tar.gz<br /> # cd openssl-*<br /> # ./config --prefix=/usr zlib-dynamic --openssldir=/etc/ssl shared enable-ec_nistp_64_gcc_128<br /> </code><br /> It then asks you to do make depend, which you should, and you follow with the usual:<br /> <code># make depend<br /> # make<br /> # make install</code><br /> and then you can check the version that's running on your server with <code>openssl version</code> and if it says: <code>OpenSSL 1.0.1g 7 Apr 2014</code><br /> you're safe and sound in Snowden-land. This also works for Ubuntu 8 (hardy) by the way.<br /> <br /> Be sure to reboot your server after that, renew your server certificates, ask (or better yet FORCE) your users to change their passwords.<br /> <br /> Of course, you've now gone to source for this package. And trying <code>apt-get remove openssl</code> will result in sheer dependency hell. Forget about letting apt manage openssl in any way or manner. dovecot, nginx, postfix etc. should all still work fine through apt with backports and such. They'll use the right openssl libraries now.</p>

It's for around 900 entries, so it makes sense for me to fix this. Of course I can do replace all x by z in the xml, but it looks like the needed characters to do that properly are already missing in the xml..

[Harm10]

Apparently these are not normal line breaks. So can you paste one example how it looks in html in the PivotX editor?
The content is normally constructed with the same function as is used to display the content on your web site.
The fact that there are blanks in the export xml suggests that these codes are something special.......
Perhaps you have stumbled on a bug in PivotX.

[jult]

That last bit is what you're asking for. HTML from pivotx, as it is when you click on [edit HTML] and as it shows on published posts.
Looks like normal linebreaks to me. With <br /> and so on.
As far as I can deduce the problem; The export extension replaces <br /> with spaces.

[Harm10]

I am sorry to say that I cannot reproduce your problem.
In my introduction is this

Code: Select all

<p>Page breaks:<br />This<br />and<br />that</p>

and in my export is this:

Code: Select all

<p>Page breaks:<br />This<br />and<br />that</p>

So exactly the same.
I am on PivotX 2.3.10.

[jult]

Are you using the latest hook release?
Have you tried other settings for text-processing inside pivotX regarding line-breaks?
Can you send me your db/ser_config.php file where this export goes as it should? Perhaps I can use that as a base, change hostname/folders in your config and then try again..

N.B. I'm using the latest pivotx as well, that can't be it. Maybe there is a way to first import all my entries into a default pivotx folder?
What would be the best procedure for that? Apparently following the one from the docs/wiki is not working for a good export.
Maybe I should NOT overwrite ser_config and use the default for that?

[jult]

I am sorry to say that I cannot reproduce your problem.
In my introduction is this

Code: Select all

<p>Page breaks:<br />This<br />and<br />that</p>

and in my export is this:

Code: Select all

<p>Page breaks:<br />This<br />and<br />that</p>

So exactly the same.
I am on PivotX 2.3.10.

By the way, that can't be true. I mean, my export has: <content:encoded><![CDATA[ and no <p> in it. What do you use to export with?

[Harm10]

By the way, that can't be true. I mean, the export has: <content:encoded><![CDATA[ and no <p> in it. What do you use to export with?

I only pasted the part that was relevant from my export. The structure you mention is just there.
A more important remark you made is that you do not have the <p> tag as well? The export should contain all html characters that are in the content unchanged.
As the PivotX editor sets <p> tags automatically I am quite interested to see what the html view in your editor looks like.
Using my ser_config.php doesn't seem like a good approach right now.

I am using the latest version for wxrexport that Hans put on sourceforge.

BTW are you using flat files or SQL in PivotX?

[jult]

The examples I showed above encompass all you ask for, all the entries are like this;
HTML shows as it should in pivotx editor, as I already mentioned and you can see in the example.
There is no HTML whatsoever in any of the exported entries' content, except for some <pre> and <code> tags. (And the <code> tag is weird, since that is not even HTML.)

Database is MySQL. All entries show as they should (with html) see here for a screengrab.

I bet you the exporter is at fault here. The only difference between your and my export is the config.

[Harm10]

I think the exporter is OK. I am searching for a potential problem just after generation.
What option do you use in your browser to get the created XML?
You just save it somewhere? Or catch it in some tool (I use Notepad++ for instance for that)?
BTW which browser do you use?

The screen shot you send shows a lot of html tags like <p> which should be in the generated xml.