This is from version 3.0.3 (I'm not using 3.1 due to this bug) MariaDB [itop]> select picture_mimetype,picture_filename, hex(substr(picture_data,1,4)),substr(picture_data,5,10) from person limit 10; +------------------+-------------------+-------------------------------+---------------------------+ | picture_mimetype | picture_filename | hex(substr(picture_data,1,4)) | substr(picture_data,5,10) | +------------------+-------------------+-------------------------------+---------------------------+...
When a new person is created there's no picture and when a person is updated it keeps the previous picture. This is the configuration of the Synhcro LDAP Person: * Name: Synchro LDAP Person * Description: Synchronization of persons from the LDAP directory * Status: Production * Target class: person * User: undefined * Contact to notify: undefined * Icon's hyperlink: (empty) * Application's hyperlink: (empty) * Data table: synchro_data_ldap_persons * Reconciliation policy: Use the attributes * Action...
user picture no longer synchronized from synchro_data_ldap_persons
done thank you.
CompileBranding shouldn't consider line breaks inside tags as relevant
I found the problem. After modifying the rrmdir function not to remove the temp directory (so I could look into it), I saw that the file was there but it turns out it was looking for a filename with line breaks in it! I think it's a bug, but the simple workaround is to define the tags in one line, e.g. instead of <main_logo _delta="define> my-extension/images/my-logo.png </main_logo> use <main_logo _delta="define>my-extension/images/my-logo.png</main_logo>
Ok, just tried and it doesn't work
No, I don't think so (or maybe I did, I'm not sure). I'll try.
Hello, just like Federico I cannot manage to use a path, only the embedded base64 version. The file is in extensions/my-extension/images/my-logo.png, it gets copied to ./env-production/my-extension/images/my-logo.png.
mmh, I'm puzzled by a problem invoking synchro_exec.php. If I specify both sources separated by comma it seems it's repeating the synchro for the first one twice, but if I specify an invalid source as the second one it correctly tells me it doesn't exist. I executed these in sequence, there was a pending update for source 5 and it wasn't executed with the first call. I looked at synchro_exec.php but I cannot see where the problem is. The solution is simple, I can call synchro_exec.php two times,...
Hello, just a suggestion, please add a link to synchro_exec.php in the section "Using MySQL statements to populate the synchro_data_xxxx table". I scratched my head for a while seeing that updating that table didn't change the synchronized data (I thought a trigger would do the job, then I saw no such trigger in the table definition). Eventually I found it, but it took a while.
Hello, just a suggestion, please add a link to synchro_exec.php in the section "Using MySQL statements to populate the synchro_data_xxxx table" here I scratched my head for a while seeing that updating that table didn't change the synchronized data (I thought a trigger would do the job, then I saw no such trigger in the table definition). Eventually I found it, but it took a while.
Hello, just a suggestion, pleas add a link to synchro_exec.php in the section "Using MySQL statements to populate the synchro_data_xxxx table" here I scratched my head for a while seeing that updating that table didn't change the synchronized data (I thought a trigger would do the job, then I saw no such trigger in the table definition). Eventually I found it, but it took a while.
OK, I got it: I have to use method 1 here https://www.itophub.io/wiki/page?id=2_6_0%3Aadvancedtopics%3Adata_synchro_overview (i.e. write the data directly in the synchro table then use synchro_exec.php
Oh, I now see that the synchro table (as well as the person table) doesn't have a picture field but picture_data,picture_mimetype,picture_filename. I guess I cannot put the "picture_data" in the csv file.
Oh, I know see that the synchro table (as well as the person table) doesn't have a picture field but picture_data,picture_mimetype,picture_filename. I guess I cannot put the "picture_data" in the csv file.
Hello, now that I have ldap-data-collector working (I had to modify it to use starttls BTW), I'd like to also import the picture from active directory. If I modify iTopPersonLDAPCollector.json to add a "picture" field, and do the same to params.distrib.xml, the csv file has a picture field (for now empty, I plan to put there later an url), but the upload fails with Error - Failed to import the data from '/net/storage01/mnt/pool1/usuarios/luca/itop/ldap-data-collector/data/iTopPersonLDAPCollector-1.csv'...
Hello, now that I have ldap-data-collector working (I had to modify it to use starttls BTW), I'd like to also import the picture from active directory. If I modify iTopPersonLDAPCollector.json to add a "picture" field, and do the same to params.distrib.xml, the csv file has a picture field (for now empty, I plan to put there later an url), but the upload fails with Error - Failed to import the data from '/net/storage01/mnt/pool1/usuarios/luca/itop/ldap-data-collector/data/iTopPersonLDAPCollector-1.csv'...
postgresql: add keepalives related parameters
Instead of we'll release a new stable version 7.2.6 That's what I meant. Thank you.
Regarding the second question, I modified the test program to generate a table without rowid and a table with an integer primary key and it both cases it worked.
Tested OK with the 7.3-testing branch (both with the test program and the "real" project). Befor closing a couple of questions/remarks will it be backported to the stable branch? an sqlite database could be defined without rowid, also a column defined as "INTEGER PRIMARY KEY" would be an alias for rowid, does the patch take into account those two cases?
7.1.4 to 7.2.4 regression in sqlite VARCHAR mapping (ftString in 7.1.4, ftMemo in 7.2.4)
No, the demo doesn't segfault, but adding the Undefined_Varchar_AsString_Length=40 property to the connection also seems to fix the segfault in the real program (probably caused by the mismatch between the field defined as tStringField and zeos interpreting it as tMemo). I don't understand why it was wrong before, but I'm closing this bug anyway. Thank you for the pointer to the forum, a comment in the release notes would also have been nice.
Zeos thinks that "SELECT rowid,* from table" with an sqlite database involves more than one table (regression)
7.1.4 to 7.2.4 regression in sqlite VARCHAR mapping (ftString in 7.1.4, ftMemo in 7.2.4)
Investigating the same problem I found a different solution: in the procedure TZPostgreSQLConnection.DeallocatePrepareStatements, in the call to CheckPostgreSQLError change the parameter lcExecute to lcUnprepStmt. This way it won't raise an error in case the connection is dropped. Since I'm not familiar with the code I cannot say if it has any adverse side effect.
VOTE: lazarus
Thank you!
Hello, instead of copying snap7.pas in every project, I made a simple lpk. Just drop...
compile fixes for vdr 2.3.1
Thank you. Don't forget to merge it to trunk.
CheckSQLiteError needs the database handle, not the statement handle.
CheckSQLiteError needs the database handle, not the statement handle.
RxSwitch doesn't show gliph at runtime