Bug #79217 [Opn]: Last character of a UTF-8 string gets messed up

Edit report at https://bugs.php.net/bug.php?id=79217&edit=1 ID: 79217 Updated by: cmb@php.net Reported by: steve dot lloyd at gmail dot com Summary: Last character of a UTF-8 string gets messed up Status: Open Type: Bug Package: ODBC related Operating System: Centos PHP Version: 7.3.14 Block user comment: N Private report: N New Comment: > The bug you referred to seemed to say this was patched but it is > still an issue in PHP 7.4. I rather think that some issues have been fixed, but likely not all. > The field I am querying is set to nvarchar(25). If I cast it as > nvarchar(50) it works. That's interesting, since echo strlen('因采购/仓库而被冻结'); // 28 echo strlen('gesp. für Arb.plan/Stückl'); // 27 So apparently either the driver or our binding (or maybe some configuration setting) is confused by UTF-8. I'm not sure how to debug this best. Is there something on CentOS like the Windows ODBC tracing? A trace of a minimal reproduce script could be very helpful. Previous Comments: ------------------------------------------------------------------------ [2020-02-04 16:09:26] steve dot lloyd at gmail dot com Interesing bit of information: The field I am querying is set to nvarchar(25). If I cast it as nvarchar(50) it works. test query: ---------- select cast(mtstb as nvarchar(50)) mtstb from sltschema.t141t Results: --------- Mtstb 因采购/仓库而被冻结 因任务清单/BOM而被冻结 MPN:BOM抬头冻结 gesp. für Besch./Lager gesp. für Arb.plan/Stückl ------------------------------------------------------------------------ [2020-02-04 13:09:26] steve dot lloyd at gmail dot com PDO has the same problem. test script: -------------------- try { $pdo = new PDO("odbc:{$datasource}",$username, $password); $query = "select * from sltschema.t141t"; $poo = $pdo->prepare($query); $poo->execute(); $row = $poo->fetch(PDO::FETCH_ASSOC); print_r($row); } catch (PDOException $e) { echo "Connection failed : ". $e->getMessage(); } Result: --------------- Array ( [MANDT] => 000 [SPRAS] => 1 [MMSTA] => 01 [MTSTB] => 因采购/仓库而被冻tc [ZMODIFIEDTS] => 20200131152705439 [ZMODIFIEDTS_TEST] => 2020-01-31 15:27:05.4380000a ) -------------------------- The bug you referred to seemed to say this was patched but it is still an issue in PHP 7.4. How do I patch/resolve this issue? ------------------------------------------------------------------------ [2020-02-04 08:56:37] cmb@php.net This seems to be related to bug #44278. Can you try basically the same script with PDO_ODBC instead? ------------------------------------------------------------------------ [2020-02-03 22:06:52] steve dot lloyd at gmail dot com Description: ------------ Last character of a UTF-8 string gets messed up. See stackoverflow URL in Test Script section. I have tried this with multiple PHP versions. I have concluded it must be an issue with PHP ODBC functions as it works fine in other languages (python, isql). Test script: --------------- https://stackoverflow.com/questions/59904479/last-character-issue-with-sap-hana-utf8-on-php Expected result: ---------------- 因采购/仓库而被冻结 因任务清单/BOM而被冻结 MPN:BOM抬头冻结 gesp. für Besch./Lager gesp. für Arb.plan/Stückl Actual result: -------------- 因采购/仓库而被冻ad 因任务清单/BOM而被ader MPN:BOM抬头冻结 gesp. für Besch./Lager gesp. für Arb.plan/Stüca ------------------------------------------------------------------------ -- Edit this bug report at https://bugs.php.net/bug.php?id=79217&edit=1