Minor notes: Any revision of trunk should run. If you are making commits without testing, please use the bleeding-edge branch & than merge changes over periodically. includes/classes/ are all upstream libraries (though some are no longer maintained). It may be better to download latest versions and/or contribute patches upstream and/or migrate to a better-maintained modern library, rather than patching them in our source tree.
Minor notes: 1. Any revision of trunk should run. If you are making commits without testing, please use the bleeding-edge branch & than merge changes over periodically. 1. includes/classes/ are all upstream libraries (though some are no longer maintained). It may be better to download latest versions and/or contribute patches upstream and/or migrate to a better-maintained modern library, rather than patching them in our source tree.
Minor notes: Any revision of trunk should run. If you are making commits without testing, please use the bleeding-edge branch & than merge changes over periodically. includes/classes/ are all upstream libraries (though some are no longer maintained). It may be better to download latest versions and/or contribute patches upstream and/or migrate to a better-maintained modern library, rather than patching them in our source tree.
Works for me; I've added you as a developer. --Rick
Hi, all. First off: thank you for the renewed interest in the project. It is great to see after so many years. refbase is nearly 20-years old! There have been several points raised in this thread. I'll try to tackle them, but feel free to raise any questions that you might have. Please do feel free to fork this. Others have (and the git repo in the linked thread is not the only example we've had). That is part of the value of free/open source software. Much of the development team have different...
No idea, but you can edit initialize/ini.inc.php & set $databaseBaseURL to be a some static value that doesn't use preg_replace (e.g. $databaseBaseURL = 'http://yourdomain.com/refbase/';)
I pushed a fix. The commented code was a legacy implementation. Those newlines only need to be added for command-line clients (and that's what the global $client checks).
reformat return message
Crossref shows that's a preprint. It is very likely that we don't support that type yet. I know Zotero has built support for that type. It is likely that refbase could support this at some point, but developer time on it is razor thin at the moment.
Crossprint shows that's a preprint. It is very likely that we don't support that type yet. I know Zotero has built support for that type. It is likely that refbase could support this at some point, but developer time on it is razor thin at the moment.
Re. DOIs: please share examples if you encounter this again. The output is not consistent and is not well documented, so it is likely that a few edge cases won't work.
This patch purportedly works, but I haven't gotten my dev environment repaired to check.
Additional sanitization
Ugh; So I think this was fixed by Thibault back in June in this patch, but somehow it didn't get pushed to SVN yet.
The bleeding-edge branch is the most current version that is in SVN. You can look at error.php; if it doesn't say Modified: $Date: 2018-02-23 04:16:36 +0000 (Fri, 23 Feb 2018), then you're working from an older version.
I haven't used the template import that you are using before. I suspect that you can just edit the source of Template:Cite journal & replace it with the source from the earlier version that I provided manually & be in reasonable shape. Neither your 2/17 nor 2/21 versions are identical to the version that I provided & both seem to be invoking some of those templates that aren't working for you.
Given the results for "Example of the manual invocation of the template cite_journal", it appears the issue is with MediaWiki, not with the refbase extension (at least for cite_journal). You're getting similar errors in both. It appears the new version of Template:Cite journal has dependencies that may need Scribunto. Mediawiki's developers may have other guidance, but my advice is still to try older versions of the template. The diff doesn't show the template code. My older version does not call...
I'm glad it mostly works for you again & hope we can fix the last few of these. Many, though possibly not all, of the deprecated /e modifiers have been removed in the SVN version of refbase. Switching to that may remove those warnings. Refbase performs many string replacements to pretty up output. Some of those replacements won't work in the version you have installed on PHP7. I'm not sure how detrimental that will be, though.
I don't use PMID import routinely. I'm wondering if this depends on a function deprecated in more recent versions of PHP. You may try the bleeding-edge branch rather than 0.9.6 to test further. Can you please enable logging of warnings and errors and check your logs?
This capability predates Scribunto and does not depend upon it. I haven't tested w/ mediawiki 1.31, though this is on my list. Have you tested to see if manual invocations of the cite journal template on youre MediaWiki deployment work? It it doesn't, that needs to be fixed first. That being said: I haven't tested with the latest version of the Cite_journal template either. It works with an older version of cite-journal. I'll test on the new one. Similarly: Do you have the Cite extension loaded and...
I'm not sure what issues with the forum we're running into. There doesn't appear to be anything that I'm supposed to moderate/approve in this thread. I rec'd emails re. your 3 posts with time stamps matching th original post date. Did you delete them ~10-15 minutes after posting? The errors you're getting show up on the initial connection of the database. It may be hard for us to troubleshoot those connection issues. Can you try connecting with the same host, user, pass, and database name using some...
Even this install-script for php 7 does not work. It ends with "Error 1044 : Access denied for user 'xyzxyz'@'%' to database 'mysql'" on the main page. ('mysql' shall not be my refbase-database.) and an empty refbase-database. I have not tried that script & can't vouch for it. It initially connects to 'mysql' with your root credentials in order to create the database that refbase will use, the user that will work with that database, and set the proper permissions. You may not be providing the database...
Do you have the same php.ini files for 5x and 7x? My prod environment is 5.6; I need to get a 7x dev environment back up for a number of reasons, but I'm not sure when I'll get to this.
There is native support to import RIS, so this is not related to bibutils. Does it fail if you use the textbox to paste records in? Does it fail if you upload a file? I haven't tested extensively under php7, but it seems more likely this is a permissions/php setting issue.
VOTE: avidemux
Glad you figured it out. The default PHPSESSID cookie applies to the entire domain. Either using a subdomain or modifying refbase to use session_set_cookie_params() to limit to a folder may work. Not sure what our default behavior should be. While I know of a few installs with multiple folders, I'm not sure if I know of any where both folders are intended for "production" (most people just do this to test an upgrade or for active development).
It seems that your localization might not be working as expected. Have you compared your current initialize/db.inc.php to your old one? Does this happen when you're not logged in (so using the default locale)? Can you perform the mysql query: SELECT 'language' FROM 'users'; to see if that makes sense? There are almost certainly going to PHP warnings. Can you please check your logs (increasing the warning level if needed)? --Rick
I'm not aware of other attempts to install on synolog. There is no support for multiple files per record. Others have used the user file and url fields to sort of allow multiple files in the current structure. I believe at least one deployment has. There is a desire to add true multifile support, but new feature development has been slow to non-existent. I believe at least one deployment has added added the feature, but they haven't pushed it upstream. We can advise how these changes would be made...
You need to edit initialize/ini.inc.php on the server.
error message cleanup
Sanitize Go Back Referer
SourceForge "ate" a more detailed reply. For now: It may be worth seeing what the MediaWiki/Monobook skin does, as it keeps "main fields" seleted The "right" file is includes/include.inc.php (which is where the QuickSearch gets built). A work around to make it always default to "main fields" is to comment out the section that defaults it to the first search field. A proper implementation may be to add a token to "main field" queries or to search for signs that this is how the query started, rather...
SourceForge "ate" a more detailed reply. For now: It may be worth seeing what the MediaWiki/Monobook skin does, as it keeps "main fields" seleted The "right" file is includes/include.inc.php (which is where the QuickSearch gets built). A work around to make it always default to "main fields": --- includes/include.inc.php (revision 1420) +++ includes/include.inc.php (working copy) @@ -2054,10 +2054,10 @@ } $optionTags = buildSelectMenuOptions($dropDownFieldNameArray, "//", "\t\t\t\t\t\t\t", true);...
Thank you for reporting this. I am able to reproduce this. I've pushed a patch on svn: --- includes/include.inc.php (old) +++ includes/include.inc.php (working copy) @@ -4308,7 +4300,7 @@ if (empty($userID)) // Find all unique language entries in the 'languages' table that are enabled: // (language names should be unique anyhow, so the DISTINCT parameter wouldn't be really necessary) - $query = "SELECT DISTINCT language_name FROM $tableLanguages WHERE language_enabled = 'true' ORDER BY order_by";...
re-add inadvertently deleted fieldInfoArrays
fix language ordering
This functionality exists already: http://www.refbase.net/index.php/Configuring_refbase#How_can_I_setup_refbase_so_that_users_can_register_themselves.3F Be sure to edit $defaultUserPermissions in inc/ini.inc.php to match your requirements.
refbase does some of thes very well, some of these not at all, and most somewhere in-between. It is free to try, so I suggest you use one of our demo servers or install it and see how well it will meet your needs. organise (using custom topic labels) the papers I have read, refbase, like most referemce managers, does this. store the content of the papers either directly (pdf format) or via a direct link, Either mode of storeage is possible. make my own (searchable) comments for each paper, This is...
That's a warning, not an error. It should function, but we'd of course like to fix it. But: I can't seem to replicate; you appear to be using a customized and/or out-of-date version? The code still exists, but at line 5271 (not 5288). It'd be especially useful to know what behavior triggered this, as the searchReplaceText function gets called in a few places & it'd be useful to know if we can just change a replace string somewhere or if we will need to use preg_replace_callback instead.
You should be able to delete or comment those lines out with no ill effects. I've update the bleeding-edge branch for this, but have not tested rigorously or made sure that we don't need those attributes somewhere. Thanks for the report & sorry for the trouble. -Rick
remove attributes that don't seem to be present in mysqli-result::fetch_field_direct and don't seem to be used
Please use the SVN version.
The SVN version has had a patch contributed by Denis Potapov that replaces mysql...
Merge bleeding-edge to trunk
- added primitive export capabilities (based on...
Commented out help link, as help is neither in ...
Minor bugfixes: install.sql uses cite_key rathe...
Changed a German word to English
Initial import of FSF documentation, mostly cop...
Fixed a few words in the english locale which w...
Initial import.
* Version number to 0.8.0 in anticipation of mi...
* Added cross-site searches
Fiexed minor grammatical error, added Endnote c...
Fixed regex which inserts '\\N' between multipl...
Cleaned up comments.
* Additional WIN32 notes
* Fixed path to Mac Endnote directory
Initial checkin; Exports all Endnote types to a...
* Changed version to 0.8.0
Grammar & type popup bug.
Quick and dirty textfile processer
Fixed series export.
Page numbers must start with a number, have whi...
Fixed output of chapter/article editor.
Exports user-specific data
* Fixed typos
Fixed relatedItem bug
For the edit record form, the current type is a...
If there are multiple call numbers, only the on...
*Fixed handling of theses
Separates keywords:
REALLY fixed the editor export. Doesn't make d...
* Series information
Fixed linewrap on one line.
* Fixed typos, clarified comments, other cleanup
* Put is_uploaded_file() back in. Please test ...
Names with out ',' delims will be exported as i...
Call number is only exported if it has been def...
Fixed type drop-down bug.
Added pages to books & other sources
Script to move files from CVS locations for dep...
Added Andreas Czerniak's en-importer.sh
Initial implementation of single record ISI (we...
fixed xml2bib->xml2end
work-around for exec() bug on win32; all exec()...
Auto-location of bibutils.
cvsdeploy.inc now handles lit/install_utf8.sql
New .inc.php include files (install, transtab f...
aulast, aufirst support in OpenURL
Generalize OpenURL support
* More quote_smart() fixes
* more quote_smart() fixes
Fix typo in new extractAuthorsGivenName function
defensive quote_smart handling of every query
Added zip.inc.php
reuse stripSlashesIfMagicQuotes in quote_smart